Off to Ballarat today to hear the
results of the tests I had done last week. Everything
OK; but the cough is still there. It seems to be completely out of the ordinary. And the
rash on my leg hasn't responded to any kind of treatment. I'll go back on Monday and have
(another) punch biopsy.
It would have been too much to expect that the new NBN tower would get erected without a fuss. Wendy McClelland is up on her hind legs
again, and has distributed another set of fliers (though not to me) warning of the
“microwave radiation tower”, and again naming the owners of the property. It's very low on
content, even less so than earlier ones; apart from the facts (the NBN will be giving
information sessions), the only statement of any relevance is:
...and have connections to it by pulsing microwave radiation out at the people which
radiates us all.
The mind boggles. I have no idea what she means. But then, she probably doesn't either.
Apart from that, it's clear that she is running out of material: most of the flyer is
identical with the flyer she sent out complaining about her misunderstanding of my tower
maps 8 months ago:
Still, she'll annoy us. Spent some time gathering information to update my documentation of
the issue, including collecting all the fliers that Wendy has distributed
over the years. I still need to rearrange the analyses that I have done of most of them.
Didn't get round to much work in the garden, but benefited from the current cool weather and
transplanted a couple of plants: the
small Alstroemeria stapricamil
“Camilla” that we bought a year
ago and planted in the middle of the garden. It had proved to be too small for the
location, and it's now next to
the Strelitzia reginae further
forward. In the process it lost a number of stems. I don't think that will be a problem:
Also moved the Strobilanthes
“Goldfussia” to where the Alstroemeria had been. It's looking decidedly less happy. Let's
hope that the new location will suit it better:
Started a routine backup of my photos today. I back up to an external disk, which should be
connected by eSATA, but I've had
trouble with that: hot plug doesn't seem to work The disks also have USB interfaces (of
course), but I've had trouble with that too. So until I sort out the eSATA hot plug issues,
I've been backing up to a system where I don't care so much if the system crashes.
Currently that's Yvonne's system, lagoon.
But today things didn't work as expected:
Dec 2 11:19:16 lagoon kernel: da0 at umass-sim0 bus 0 scbus0 target 0 lun 0
Dec 2 11:19:16 lagoon kernel: da0: <ST2000DL 003-9VT166 > Fixed Direct Access SCSI-2 device
Dec 2 11:19:16 lagoon kernel: da0: 40.000MB/s transfers
Dec 2 11:19:16 lagoon kernel: da0: 1907729MB (3907029168 512 byte sectors: 255H 63S/T 243201C)
Dec 2 11:19:16 lagoon kernel: ugen1.2: <Sunplus Technology Co.,Ltd.> at usbus1 (disconnected)
Dec 2 11:19:16 lagoon kernel: umass0: at uhub1, port 1, addr 2 (disconnected)
Dec 2 11:19:16 lagoon kernel: (da0:umass-sim0:0:0:0): AutoSense failed Dec 2 11:19:16 lagoon kernel: GEOM: da0: the secondary GPT table is corrupt or invalid. Dec 2 11:19:16 lagoon kernel: GEOM: da0: using the primary only -- recovery suggested.
Dec 2 11:19:16 lagoon kernel: (da0:umass-sim0:0:0:0): Synchronize cache failed, status == 0xa, scsi status == 0x0 Dec 2 11:19:16 lagoon kernel: (da0:umass-sim0:0:0:0): lost device Dec 2 11:19:16 lagoon kernel: (da0:umass-sim0:0:0:0): removing device entry
That's not one serious error: it's two. First, the GPT table is corrupt, suggesting
wide-ranging corruption on the whole (2 TB) disk. And secondly, the USB interface didn't
work. That raises several questions: What has gone wrong? What component has failed? How
do I recover? Can I trust the remaining data on the disk?
The easiest thing to check was the USB interface to lagoon. Took the disk
to teevee, the lounge room computer, and tried there:
Dec 2 11:23:44 teevee kernel: ugen1.2: <Sunplus Technology Co.,Ltd.> at usbus1
Dec 2 11:23:44 teevee kernel: umass0: <Bulk Only Interface> on usbus1
Dec 2 11:23:44 teevee kernel: umass0: SCSI over Bulk-Only; quirks = 0x0000
Dec 2 11:23:45 teevee kernel: umass0:3:0:-1: Attached to scbus3
Dec 2 11:24:59 teevee kernel: da0 at umass-sim0 bus 0 scbus3 target 0 lun 0
Dec 2 11:24:59 teevee kernel: da0: <ST2000DL 003-9VT166 > Fixed Direct Access SCSI-2 device
Dec 2 11:24:59 teevee kernel: da0: 40.000MB/s transfers
Dec 2 11:24:59 teevee kernel: da0: 1907729MB (3907029043 512 byte sectors: 255H 63S/T 243201C)
Dec 2 11:24:59 teevee kernel: GEOM: da0: corrupt or invalid GPT detected. Dec 2 11:24:59 teevee kernel: GEOM: da0: GPT rejected -- may not be recoverable.
At least this one didn't have any USB problems. But that just managed to show me more
problems with the GPT, and it didn't create the /dev/da0p1 that I needed to look at
the partition. gpart didn't recognize any GEOM: the disk was somehow dead in the
water. I was able to dump the initial sectors of the disk, and they seemed to contain valid
data, but I don't understand the format completely.
So was it maybe the USB interface? The enclosure has an eSATA interface too, but my eSATA
card was in defake, which I had lent to Chris Yeardley. But Chris was finished
with defake, so went to her place and picked it up, bringing the second backup disk
while I was at it. And when I put that in lagoon, things worked as expected:
Dec 2 12:57:17 lagoon kernel: ugen1.2: <ST> at usbus1
Dec 2 12:57:17 lagoon kernel: umass0: <ST ST2000DL003-9VT1, class 0/0, rev 2.00/3.00, addr 2> on usbus1
Dec 2 12:57:17 lagoon kernel: umass0: SCSI over Bulk-Only; quirks = 0x0000
Dec 2 12:57:18 lagoon kernel: umass0:0:0:-1: Attached to scbus0
Dec 2 12:57:18 lagoon kernel: da0 at umass-sim0 bus 0 scbus0 target 0 lun 0
Dec 2 12:57:18 lagoon kernel: da0: <ST ST2000DL003-9VT1 3.00> Fixed Direct Access SCSI-4 device
Dec 2 12:57:18 lagoon kernel: da0: 40.000MB/s transfers
Dec 2 12:57:18 lagoon kernel: da0: 1907729MB (3907029168 512 byte sectors: 255H 63S/T 243201C)
So: whatever it was, it had to do with the specific disk. The two disks are identical, but
the probe messages differ. Checked the geometry:
That should be identical for both disks, so if all else failed I could copy the first 34
sectors from one to the other. But first I wanted to check the eSATA interface, which meant
bringing up defake. That wasn't helped by the fact that I had done some undocumented
changes to the config when I gave it to Chris, including changing the root password to
something I now forget. Finally got it reconfigured, connected the disk via eSATA,
and—nothing. The system didn't see it. Tried the other disk:
Dec 2 13:43:44 defake kernel: ada0 at siisch0 bus 0 scbus0 target 0 lun 0
Dec 2 13:43:44 defake kernel: ada0: <ST2000DL003-9VT166 CC32> ATA-8 SATA 3.x device
Dec 2 13:43:44 defake kernel: ada0: 300.000MB/s transfers (SATA 2.x, UDMA6, PIO 8192bytes)
Dec 2 13:43:44 defake kernel: ada0: 1907729MB (3907029168 512 byte sectors: 16H 63S/T 16383C)
Hot plug works! Could it be that I only ever tried hot-plugging with the other disk? But I
wasn't completely free of problems. When I tried mounting it, I got the dreaded
Dec 2 13:44:35 defake kernel: WARNING: R/W mount of /photobackup denied. Filesystem is not clean - run fsck
Why is that? I've seen too many of them. Is it some issue with writes to the disk not
completing before umount? fsck worked fine with no problems, but it's still
disturbing.
In any case, it's now clear beyond any doubt that the original problem was in the disk or
its USB/eSATA adapter. Took it out of the case, put it into defake, rebooted, and:
Dec 2 14:17:13 defake kernel: ad10: 1907729MB <ST2000DL003 9VT166 CC32> at ata5-master UDMA100 SATA 1.5Gb/s
And the disk did not require fsck. So the problem was with the enclosure and
its minimal electronics. Left the disk in defake, and I'll leave it there until I
can get another one, probably next week.
I've been following the TV reception problems for some months now. I'm gradually coming to
the conclusion that there are several issues, few of which are related to my hardware (which
is what I originally thought). In particular, there are some recurrent programmes that
consistently fail, while others record well. Many recordings are fine most of the way, and
then suddenly fail catastrophically, and it seems to be at least somewhat related to date
and time. All this points to some kind of interference. At some point I'll move the
information to a database and do some analysis.
But things aren't that simple. If it's interference, it's strongly dependent on the
frequency. Last night I had:
Programme
Date
Start
End
Daisy chain
File
Number of
name
time
time
Channel
Tuner
position
size (GB)
recoding errors
Mickey Blue Eyes
1 December 2011
20:27:02.732
23:05:00.488
2080 (GEM)
2
1
12
0
In Her Shoes
1 December 2011
20:27:02.999
23:50:00.610
2006 (PRIME7)
1
3
5.7
11, died at start
Those two recordings started at (almost) exactly the same time, yet one had no problems, and
the other was uselessly mutilated. What causes that? I suppose it could be interference on
some specific frequency, but I would have expected interference to be less specific than
that.
Warren Toomey has written an article entitled “The strange birth
and long life of Unix”. It's good reading, and it gave me an insight that I didn't
have before. I know Warren has researched this stuff carefully, and even apart from that it
also has a ring of authenticity about it. An excellent addition to the collection of Unix
history.
My work in the middle east garden stopped for an unusual reason: I found it difficult to
tear out all the strawberries that were growing there. We had eaten some of the ones I
picked a couple of days ago, and they didn't
taste bad. So, as planned, finally got round to putting the old wash trough on the north
verandah, not helped by
a Jacky dragon who was in
the area and didn't want to move:
It later became clear that this wasn't a Jacky Dragon at all, but
a Blue-tongued lizard.
I know that they tend to freeze when threatened, but I had thought that if I pushed him with
the handle of a garden fork (in the photos), he would go away. Not so. We left him and
attended to another issue, what to do with the pot that had contained the mini-pond on the
east verandah. After finishing the real pond,
it was somewhat superfluous. After some discussion, put it in the Japanese Garden:
Then back to the trough. The dragon was gone, and with some difficulty we moved it to the
west end of the verandah, filled it with soil, and Yvonne planted the strawberries:
House photos again today. Most went smoothly enough, but I'm still playing around with the
lighting of the verandah panorama. Apart from the fact that 24 full-power flashes don't do
the flash unit much good, they don't always do the picture quality much good either. I end
up with overexposed areas like this, taken 2 weeks ago:
So last week I tried using TTL flash. That has the advantage that the illumination is more
even, but I take three images of each view, and I only want a flash on the first one. That
works with a full-power flash, since the flash unit then takes a couple of seconds to
recover. But with TTL, it only uses a fraction of the power, and it's quite happy to fire
on all three images. So I need to take one image, turn unit off, take the other two, turn
on again, reposition, repeat. It's not easy, and I made a couple of mistakes. But then
something strange happened: the camera forgot its sequences. This looks like something I
saw a couple of days ago. In this case, it missed one of the brackets:
Looking back at the sunset photos I processed last Wednesday, that proved to be a completely different issue: I had numbered the photos incorrectly.
But what made the camera drop a sequence here?
As a result, I only took flash photos of the lower row. I think that's a mistake. Here the
last three panoramas: two weeks ago, full flash; one week ago, TTL flash on all images
(except the zenith, for which I never use flash); and today, TTL flash on the lower row
only:
Today's wasn't the best; I do need to use flash for the upper row as well. But is
there more to it than that?
As if that wasn't enough, I had another subject: my office is in a filthy mess, and it's
time to tidy it up. First, though, some “before” photos—a panorama, of course. That's not
the easiest, due to the extreme perspective issues, and the resultant panorama didn't want
to fit at all. Tried an incremental approach, like I did in August, but didn't get beyond the lower row, and that with
enormous errors, particularly round the 3rd of the 5 monitors:
Cannelloni have been on our menu for decades.
The recipe we use comes from Time-Life's “Die Küche in Italien”, a German translation of a
book published in 1968, which I've had for about 35 years. It goes into some detail to
explain how to make the tubes out of pasta
fresca, but that's a lot of work, and nowadays you can buy “no-cook” tubes.
That's what we did today—San Remo cannelloni tubes. The result: terrible. You need much more liquid to cook the
tubes, but this particular one seems to be particularly bad. It wasn't as if the pasta was
just uncooked; it tasted like glue. Possibly it would have been better with even more
water, but there's no warning on the packet, and I'm not convinced. For the moment, at any
rate, it's a brand to avoid.
Power failure this morning at 0:41. Or
was there? My radio alarm clock in the bed lost power, but nothing else seems to have done
so. This is clearly not the 2 second “autorecloser” style power failure.
Despite the power failure, turned the radio on at 7:00 this morning and heard—noise. There
was almost no signal to be heard. It went on all morning until it finally recovered. And I
also had no less than three TV recordings that were completely useless—one, which should
have been about 5 GB, recorded only 4.9 MB of nothing recognizable. All the more reason to
believe interference or transmitter problems.
Since writing my weather station software, now a
couple of years ago, I've been continually monitoring the results for errors. So this one
caught my eye today:
On checking, though, it's correct. The highest temperature of the last 5 days was 23.9°, at
0:53 on 30 November 2011. I'm continually amazed how variable the day's
temperatures are.
Yvonne and Chris visited Lisa Graber today to have a horse
covered, and Yvonne came back with a number of plants: a number
of Buddleja davidii seedlings,
a Betula pendula birch tree—just
what we're trying to get rid of, except that this one has been trained as
a bonsai plant, and some
large Protea flowers:
More work on the middle of the eastern garden today. Planted a number of bulbs, I
think Hippeastrum, in what was once
bed number 2, and set to to remove the remaining strawberry plants, which proved to be
carrying a significant number of worm-eaten fruit. I must have collected 50 plants of
various sizes, and there are still a number to be done. Sent out a message on Freecycle, which their software showed (almost)
correctly in the preview window and then wrapped unappetizingly in the final post.
Coming into the office this morning, the first thing I noticed was the disk access light
on dereel: full intensity. The system was still running, but further investigation
showed that the disk subsystem had hung itself up again. The log files showed a similar
problem to the one I had last month:
Dec 4 21:24:38 dereel kernel: ahcich2: Timeout on slot 19
Dec 4 21:24:38 dereel kernel: ahcich2: is 00000000 cs 00080000 ss 00000000 rs 00080000 tfd 1d0 serr 00000000
Dec 4 21:25:37 dereel kernel: ahcich2: Timeout on slot 27
Dec 4 21:25:37 dereel kernel: ahcich2: is 00000000 cs 08000000 ss 00000000 rs 08000000 tfd 1d0 serr 00000000
Dec 4 21:27:35 dereel kernel: ahcich2: Timeout on slot 8
Dec 4 21:27:35 dereel kernel: ahcich2: is 00000000 cs 00000100 ss 00000000 rs 00000100 tfd 1d0 serr 00000000
Dec 4 21:28:41 dereel kernel: ahcich2: Timeout on slot 24
Dec 4 21:28:41 dereel kernel: ahcich2: is 00000000 cs 01000000 ss 00000000 rs 01000000 tfd 1d0 serr 00000000
Dec 4 21:29:40 dereel kernel: ahcich2: Timeout on slot 21
Dec 4 21:29:40 dereel kernel: ahcich2: is 00000000 cs 00200000 ss 00000000 rs 00200000 tfd 1d0 serr 00000000
Dec 4 21:32:03 dereel kernel: ahcich2: Timeout on slot 7
Dec 4 21:32:03 dereel kernel: ahcich2: is 00000000 cs 00000080 ss 00000000 rs 00000080 tfd 1d0 serr 00000000
Dec 4 21:32:51 dereel kernel: REDZONE: Buffer overflow detected. 4 bytes corrupted after 0xcb38f900 (128 bytes allocated).
Dec 4 21:32:51 dereel kernel: Allocation backtrace:
Dec 4 21:32:51 dereel kernel: #0 0xc0b4ad8a at redzone_setup+0x3a
Dec 4 21:32:51 dereel kernel: #1 0xc08cdd40 at malloc+0x100
Dec 4 21:32:51 dereel kernel: #2 0xc049e6d4 at camq_resize+0x34
Dec 4 21:32:51 dereel kernel: #3 0xc049e766 at cam_ccbq_resize+0x36
Dec 4 21:32:51 dereel kernel: #4 0xc04a0427 at xpt_dev_ccbq_resize+0x37
Dec 4 21:32:51 dereel kernel: #5 0xc04a059d at xpt_start_tags+0x6d
Dec 4 21:32:51 dereel kernel: #6 0xc04a7442 at probedone+0x822
Dec 4 21:32:51 dereel kernel: #7 0xc04a3f31 at camisr_runqueue+0x2e1
Dec 4 21:32:51 dereel kernel: #8 0xc04a408f at camisr+0x13f
Dec 4 21:32:51 dereel kernel: #9 0xc08b737b at intr_event_execute_handlers+0x13b
Dec 4 21:32:51 dereel kernel: #10 0xc08b8a3b at ithread_loop+0x6b
Dec 4 21:32:51 dereel kernel: #11 0xc08b3eba at fork_exit+0xca
Dec 4 21:32:51 dereel kernel: #12 0xc0c15f64 at fork_trampoline+0x8
Dec 4 21:32:51 dereel kernel: Free backtrace:
Dec 4 21:32:51 dereel kernel: #0 0xc0b4ad19 at redzone_check+0x179
Dec 4 21:32:51 dereel kernel: #1 0xc08cda88 at free+0x38
Dec 4 21:32:51 dereel kernel: #2 0xc049e712 at camq_resize+0x72
Dec 4 21:32:51 dereel kernel: #3 0xc049e766 at cam_ccbq_resize+0x36
Dec 4 21:32:51 dereel kernel: #4 0xc04a0427 at xpt_dev_ccbq_resize+0x37
Dec 4 21:32:51 dereel kernel: #5 0xc04a059d at xpt_start_tags+0x6d
Dec 4 21:32:51 dereel kernel: #6 0xc04a7442 at probedone+0x822
Dec 4 21:32:51 dereel kernel: #7 0xc04a3f31 at camisr_runqueue+0x2e1
Dec 4 21:32:51 dereel kernel: #8 0xc04a408f at camisr+0x13f
Dec 4 21:32:51 dereel kernel: #9 0xc08b737b at intr_event_execute_handlers+0x13b
Dec 4 21:32:51 dereel kernel: #10 0xc08b8a3b at ithread_loop+0x6b
Dec 4 21:32:51 dereel kernel: #11 0xc08b3eba at fork_exit+0xca
Dec 4 21:32:51 dereel kernel: #12 0xc0c15f64 at fork_trampoline+0x8
Once again it repeated itself with different slot numbers, along with essentially the same
backtrace:
Dec 4 21:33:22 dereel kernel: ahcich2: Timeout on slot 12
Dec 4 21:33:22 dereel kernel: ahcich2: is 00000000 cs 00001000 ss 00000000 rs 00001000 tfd 1d0 serr 00000000
Dec 4 21:33:54 dereel kernel: ahcich2: Timeout on slot 3
Dec 4 21:33:54 dereel kernel: ahcich2: is 00000000 cs 00000008 ss 00000000 rs 00000008 tfd 1d0 serr 00000000
Dec 4 21:34:51 dereel kernel: ahcich2: Timeout on slot 20
Dec 4 21:34:51 dereel kernel: ahcich2: is 00000000 cs 00100000 ss 00000000 rs 00100000 tfd 1d0 serr 00000000
So: what is it? Everything points to the disk controller rather than to the disk—amusingly
like my issues with USB disks a few days ago. I had a spare controller, so moved the disk to it:
Nov 22 15:50:31 dereel kernel: ada2 at ahcich2 bus 0 scbus2 target 0 lun 0
Nov 22 15:50:31 dereel kernel: ada2: <SAMSUNG HD103SJ 1AJ10001> ATA-8 SATA 2.x device
Nov 22 15:50:31 dereel kernel: ada2: 300.000MB/s transfers (SATA 2.x, UDMA6, PIO 8192bytes)
Nov 22 15:50:31 dereel kernel: ada2: Command Queueing enabled
Nov 22 15:50:31 dereel kernel: ada2: 953869MB (1953525168 512 byte sectors: 16H 63S/T 16383C)
...
Dec 5 08:31:01 dereel kernel: ada2 at ahcich3 bus 0 scbus3 target 0 lun 0
Dec 5 08:31:01 dereel kernel: ada2: <SAMSUNG HD103SJ 1AJ10001> ATA-8 SATA 2.x device
Dec 5 08:31:01 dereel kernel: ada2: 300.000MB/s transfers (SATA 2.x, UDMA6, PIO 8192bytes)
Dec 5 08:31:01 dereel kernel: ada2: Command Queueing enabled
Dec 5 08:31:01 dereel kernel: ada2: 953869MB (1953525168 512 byte sectors: 16H 63S/T 16383C)
Into town to have a biopsy done on the rash on my leg. In the end it wasn't a punch biopsy
after all—not surprisingly, since it's directly on my shin—but a scraping (I think that's
the term she used). And afterwards I had to wait 20 minutes with a heavy bandage round my
leg before I was allowed to walk, and I need to keep it bandaged for a few days, a far cry
from what normally happens when I cut myself.
While in town, dropped in at Formosa
Gardens to look for a Protea for
Yvonne. Found
a Protea cynaroides “King
White”, and also an Echium
candicans “Pride of Madeira”—the latter seems to be less a cultivar name than the common
name for the species.
Back home, and exceptionally managed to plant both of them within a couple of hours. Put
the Protea in the eastern (dry) bed, not far from
the Leucospermum
cordifolium. The branches are quite long already, and it's pretty windy there, so I was
concerned that they might break off. But it's too big for the plastic wind protection tents
that we have, so ended up cutting part of the ALDI toy greenhouse that David Yeardley gave me some months ago and wrapping it round
some bamboo stakes. The results are more interesting than pretty, but I suppose it'll do
the trick:
Madeline (I think) was the first person to come and pick up our giveaway plants. It's
amazing how many things we were able to give her, including stuff that wasn't on the list,
such as Viola tricolor
and Tropaeolum, both plants that
reproduce like fury. I hope she's happy with it all.
Came back in from the garden to find that we were off the network. No signal strength
issues, but no connectivity. I'm used to this now: more often than not it's not an issue
with the connection, but with this horribly flaky Huawei USB modem (there, USB again). So popped the modem and reconnected
it. ppp process redialled, established connection—and still not connectivity.
Stopped and restarted the ppp process, and it worked. Do we have software issues
here instead of (or as well as) hardware issues?
Off to the Dereel Hall this afternoon for
the information session about the
NBN fixed wireless tower. Wendy McClelland,
her husband and one supporter were standing outside, I think distributing pamphlets—they
didn't offer me one—and exposing themselves to strong electromagnetic radiation, one of the
few sources that is, indeed, proven to cause cancer: the sun.
I had expected the session to be some kind of presentation, but in fact it was much less
formal: a series of posters talking about different aspects of the project, not a “session”
at all. Met Scott Weston for the first time—he came in just after me—and spent some time
talking to him and Peter . Also met
Gabriel from Ericsson and an engineer from NBN, both
of whom were able to give me some details about the technical side of things.
As Peter had indicated, the connections will
be LTE,
specifically TDD,
at 2300 MHz. The speed quoted was 12 Mb/s downstream, but there was some disagreement about
whether this was the minimum or the maximum. Peter said it was the minimum, but Gabriel
gave more specific speeds: downlink maximum 12 Mb/s, guaranteed 500 kb/s. Uplink 1 Mb/s
maximum, 150 kb/s guaranteed. The NBN engineer wasn't so
sure; I suspect this might be an area that isn't cast in concrete. Certainly 12 Mb/s is
nothing special nowadays: Telstra do a maximum
of 22.6 Mb/s with their 3G technology, and I've heard of up to 80 Mb/s with LTE. And even
my current HSPA
connection has uplink speeds that can exceed 1 Mb/s.
There's other stuff, though, that makes the difference. The tower will be connected to the
backbone network via a microwave link of (probably) 200 Mb/s, and there's a Point of Interconnect (PoI)
in Ballarat. The bandwidth of this link
is enough to ensure no contention. But that would be reached with only 17 stations
downloading at 12 Mb/s, another indication that the 12 Mb/s is more likely to be a maximum.
My understanding from the NBN engineer is that they're not
expecting there to be enough contention to lower the speed below the maximum except in
exceptional circumstances.
More interesting is the latency: 20 ms to the PoI, about a quarter of what I have with HSPA.
That would probably be enough to make a switch worthwhile. But then there's the question of
cost: they mentioned a wholesale price of $25, which without further qualification is not
very meaningful. But the statement that it would be the same price as an equivalent fibre
connection is very meaningful. Internode's
pricing for a 12/1 Mb/s link is $50 for 30 GB traffic, or $70 for 300 GB. Those are
both more than the $40 I'm paying now, but only a little, and I only get 9 GB traffic.
One of the advantages of “fixed wireless” is that the NBN can plan the availability better.
In general they plan for a coverage range of 3 to 5 km, which is fixed before construction
begins, so they're in a better position to guarantee bandwidth.
The equipment is a fascia mount antenna, which also contains the modem, and an internal box
that looks like a 4 port Ethernet switch, which the NBN call (but don't document) a Network Termination Device
(NTD), rather than the more common “Network Termination Unit” or NTU:
The NTD is the small box at the bottom, on top of the antenna mount. The NBN engineer tells me that it's more than a switch. The connection
from the external to the internal unit is standard
Cat 5e Ethernet, but the
wireless link is layer 2, and the box converts the layer 2 to a layer 3 interface. I don't
understand that. Switches are layer 2, so how can any conversion take place? But as
the NBN engineer explained, it does mean that you can have
multiple connections to different ISPs. Presumably the real purpose of the box is to
provide some form of authentication. But that's an issue we haven't discussed at all yet.
The tower will look pretty much like a mobile phone tower: three antennas pointing at
(nominally) 120° from each other, each with a maximum power of 20 W. As one of the posters
indicated, the closest you can get to the antennas is 430 m, at which distance the EME
represents 0.0041% of the ARPANSA
exposure limits:
It would have been nicer if they had specified it in μW/m² and compared it with other
sources, but if I read the recommendations correctly, 0.0041% corresponds to 41 μW/m².
That's at the closest point, on an uninhabited stretch of road. Round here, still not far
away, it would be closer to 2.5 μW/m².
It appears that I read the recommendations incorrectly. At 2300 MHz, the ARPANSA
limits are 10 W/m², so the values above would be 410 μW/m² and 25 μW/m², for all
the difference that makes.
And the location of the tower? They considered a total of four places, including the Dereel
Hall area and the north-east end of our property, and also a place in Swanson Road close to
the airstrip:
It's interesting that they note one of the reasons for rejecting our site as the lack of
power: there's a power cable that goes straight past the site. But being too far west is
certainly an issue.
The fact that our site was mentioned made me wonder whether Wendy had inside information
when she published her slanderous claims about me in March. But no,
the NBN engineer told me that they didn't start looking until
1 June 2011. They want to be finished by June 2012, but that will
depend on how many spanners Wendy wants to throw into the works. Hopefully they'll deal
with them quickly.
So: is it a worthwhile offering? Doubtless. Is it optimal? Definitely not. Give me fibre
any day. But it's not likely that we'll get that.
Three more people came along today to pick up plants. I hadn't expected them to be so
popular. One even took nearly all
the Watsonia and Chasmanthe floribunda corms.
She'll have fun with them—there must be enough for a couple of hectares, but no, she wanted
them all.
It seems to be a perfectly normal snail, and it was on a water lily leaf. How did it get
there? When I first saw it, it was nearly completely in the water, though it clearly didn't
like that much. But it must have swum there in the first place.
Our goldfish have been hiding almost since we put them in the pond, but they're gradually
showing themselves again. The small ones have grown by almost 30%. I wonder what they're
eating—insects on the surface, maybe.
Call at 8 am this morning from Prue Bentley of ABC local radio regarding Wendy McClelland
and the topic of the NBN info session yesterday. She wanted me to participate in a news programme on the subject, but
unfortunately I didn't get to the phone on time, so she followed up with an email: “I tried
your home number but it wasn't working... do you have another phone contact?”. I replied,
but didn't hear back from her—clearly it was too late.
The ABC did publish an
article on the web site, though. It's interesting that they don't mention “DATA”, only “A couple who live at Dereel, south of
Ballarat, is considering taking legal action... Wendy and Stuart McLelland say they are
worried radiation from the National Broadband Network (NBN) wi-fi towers could be
carcinogenic.” But more interestingly, where are the others? Are they alone after all? In
passing, the factual accuracy is interesting. They misspelt “McClelland”, and they refer to
the tower as “wi-fi”. But then, what can
you expect when you invent silly terms like “Wi-Fi”?
Later got a phone call from Pia Akerman of “The Australian”, who didn't want to talk about the merits, just what I thought about
the alternative of fibre or wireless. There's no question that fibre is better, of course,
and that's what I told her, also explaining that there wasn't a hope of getting fibre in the
next 10 years. I also sent her a link to my writeup of yesterday's events. But what she wrote gave the impression that I agree with Wendy. Of course, if she were
to achieve a fibre connection for Dereel,
I'd be very grateful. But I don't see the slightest hope of that. Another case of creative
reporting.
This article cleared one thing, though: it includes a photo of Wendy with 8 people in the
background. So there are a couple more:
It's interesting to note that Wendy is holding an NBN information kit in her left arm.
What's the significance of that? And “retired IT worker” for me? I tend to call myself a
computer industry has-been, but “IT worker” sounds like a euphemism, just as people use the
term “sex worker” to mean “prostitute”.
Finally got around to putting down weed mat round some of the more deserving plants, notably
the Araucaria bidwillii and
the Podocarpus elatus
(Illawarra plum). Also finally planted
the Photinia × fraseri robusta that we
bought three months ago. I would have planted more, but
the temperatures shot up again, and we had a top temperature of 35.8°, compared to just
18.5° three days ago. Wouldn't it be nice to have
mid-range temperatures?
Partial loss of power this evening at 21:10, enough to set the UPSs screaming, but not
enough to affect even the bedside alarm clock. There seem to be a lot of these lately.
What's causing them? And what do you call them? I had always called
them brownouts, but it
seems that that's a more prolonged condition, and also intentional.
Another hot day today, a maximum of 36.1°. I had intended to continue my work in the
garden, but in the end only did some work in the greenhouse, which was also in dire need:
repotted the Chile poblano and attended
to the “Giant tree tomato”, which really needs much more in the way of support. Also pruned
the passion fruit, which is developing too many side shoots. Planted the prunings; there's
a good chance they'll strike.
Probably the main objection that Wendy McClelland has to wireless communications is that
they cause cancer. Nothing we can say can prove otherwise. One of the problems, of course,
is that electromagnetic radiation, in sufficient dosage, really does cause cancer. How
much? As I've noted previously, the presentation of the data doesn't make it easy to
compare. I established that the maximum radiation from the NBN tower would be about 41 μW/m². But what's the level of solar
radiation? Discovered a new word,
“Insolation”, along with some typical
values: about 1 kW/m² in bright sunshine. There are more specific pages at Aussie RV products, which shows an
average of 5.1 kWh/m² per day in Melbourne, and the Bureau of Meteorology, which shows a map of Victoria showing a current insolation of
about 33 MJ/m² per day. The site is broken and won't return average values for a year, so
the two are not completely comparable. Aren't units wonderful things? I'm reminded
of Andy Tanenbaum: “The
nice thing about standards is that you have so many to choose from”.
In any case, divide Joules per day by the
number of seconds in a day (86,400) to
get Watts, so 33 MJ in a day is an average
of 382 W, spread over day and night. So the average insolation in Melbourne is just shy of
10 million times as much as the highest level of radiation from the tower. Put another way,
to get the same amount of radiation from the tower as Wendy got standing out in the sun for
an hour, she would have to be exposed to the tower for about 2,800 years. And
even that assumes that the effect is cumulative.
To Geelong in the afternoon to a regular
periodontal checkup. Looks like I'm going to need some more serious work on my gums.
I had planned to do other things, like visiting Bunnings, but it was too hot, and in addition the car was misfiring—looks like I may
have problems with the ignition circuit, probably related to this silly theft alarm which
has given me problems a couple of times in the past. So put that off until tomorrow. On
the way home, finally got some rain, and things cooled down. At home, in the 2 hours from
16:08 to 18:08, the temperature dropped from 35.1° to 19.1°—an average of 0.13° per minute.
While in Geelong, dropped in at MSY to replace
my external disk enclosure. Yes, it's under warranty, so they'll send it off, and at some
time I'll get a replacement. But they didn't
have anyeSATA
enclosures, so I left with nothing. I wonder if it's worth going there any more.
I've had this funny cough since mid-September, and I've
been to the doctor about it a couple of times. The interesting thing was that it started
just after a previous visit to the doctor, and I thought I might have caught an infection in
the waiting room.
There had been a change of doctor since then, and I told her the background, including my
suspicion that it had something to do with my previous visit. She gave me a broadband
antibiotic—$2.60 per pill!—and asked me to come back if it was still there. It was.
The next time round she
considered asthma. I had had asthma
as a child, and that wasn't completely impossible. So
ended up buying lots of inhalers. Did they work? Hard to say, since the change isn't
immediate. But in the meantime I had decided that no, they didn't. And it's not clear why
I ended up buying two inhalers to treat acute attacks of asthma when the last attack was
nearly 50 years ago.
So I did a bit of thinking. I had started new medication just before the cough
started, Coversyl. Checked the
side effects and
found:
Rare and mild, usually at the start of treatment.
Cough
...
So when I went into town today, I asked about the side effects of Coversyl. Immediate
response: “cough”.
So I had been on a wild goose chase this last month, and I had bought medication costing
nearly $100 in total, all barking up the wrong trees. Not happy. She prescribed me an
alternate product, Micardis, without
mentioning any side effects. We'll see.
The real reason for the visit, though, was the skin lesion on my shin. That proved to
be lichenoid keratosis, and
it seems that she has already removed most of it. Otherwise there's the option of
cryosurgery. But is it correct? This time it was the laboratory, of course, but reading a
more detailed
description, it doesn't seem to match the symptoms.
It's becoming more and more clear that I don't know the identity of the bulbs and corms that
I collected. There are at
least Watsonia
and Chasmanthe floribunda,
but the more I look at it, the more I get convinced that there's a third kind there too. So
I've borrowed lots of books from the libraries. So far, they're no help. Very few of them
are designed to help identify plants—you buy them at the
“nursery”, right? Found one with the
promising title “Bulbs”,
by Roger Phillips and Martyn Rix, which did publish photos of live plants including
the corm/rhizome/whatever, but they seem to have run out of steam in the middle. In
particular, there's not a single photo that matches the corm on the left in this photo:
But then, it doesn't mention Chasmanthe either. Part of the problem may be that the book
was published in 1981 in the UK, so long ago that the publisher doesn't want to know. Maybe things have changed significantly in that
time. Still, I have about 12 other books to look at, though a number are clearly
uninteresting. Wouldn't it be nice to have a single web resource for this sort of thing?
There don't seem to be as many flowers this year, but this one is much bigger than last
year. Somehow some of the rhizomes have died back; I wonder if that's because of the
competition it had until a couple of weeks ago.
Still more fun with the exposure for my verandah photos. This was the brightest day since I
started experimenting with the flash output. Last week I
used normal TTL flash for the first image of each group of 3, but that proved not to be as
bright as I had wanted (the intention was to have an overexposed image), so today I
overexposed by 3 EV. Apart from finally draining a set
of Nickel-Zinc batteries,
that proved not only to be too much, but it also created problems with the zenith shot,
where I don't use flash (since the roof is reflective). The results were this kind of
difference, showing the zenith shot and a couple of random upper layer shots to which the
zenith had to attach:
Over ten years ago I found our cat
Lilac with a baby rabbit in her basket. That time
round I didn't have the presence of mind to take photos, so when it happened again today, I
first took photos and then investigated:
In the first image, Lilac is meowing, not hissing or anything else. Part of the chest of
the rabbit was wet, but it's not sure what Lilac had intended: lick or bite? It seems that
she has conflicting instincts. In any case, the rabbit was alive, but scared to death, and
in the end I took it outside to let it escape. Arguably that was the wrong action: rabbits
are pests, after all.
When I came back an hour later, it was still there, dead, so I let Lilac come and pick it
up. She ate the front half, and Nemo was a grateful
recipient for the back half.
After discharging the
(Nickel-Zinc) batteries in my
flash unit, had a chance to see how they recharged. I haven't measured the charging time,
but I guessed it to be about 5 hours in the “fast” charger, quite a difference from the 2½
maximum stated on
the Wikipedia page.
Confusingly, that refers to a document published by the
maker of the charger:
Fault conditions:
Stop Charge [sic] if any of the following conditions occur:
Total charge time exceeds 2.5 hrs.
Temp. of the cell rises by more than 15°C
Temp. of the cell exceeds 40 °C
Voltage is less than 1.6V
It's not quite clear how you charge a battery with a voltage less than 1.6V, but that's
probably one of these typical inaccuracies in this kind of document. Another document, also
by the maker of my charger, shows a charger with a slightly different appearance boasting
“Enjoy a quick charge in little as 1.5 hours*”. That * leads nowhere, but in the
small print it says “Charge two AA or AAA NiZn batteries in 1 to 1.5 hours, or charge 4 AA
batteries in about 3 hours.“.
That's still a lot less than the 5 hours I estimated. Could I be that wrong? A closer look
at the charger says no:
So what's going on? Except for the colour, and the completely unfounded claim “fast”, this
unit looks pretty much identical to the one in the web page, even down to
the detail that it will charge 2 AAA or 4 AA batteries (in the middle two slots):
It's difficult to be sure from the advertising photo, but it looks as if the moulding for
the white charger is the same as for the black one, including the two arcs at the top and
the cutouts immediately above and below the batteries. Conceivably the right-hand LED is
green, which would make sense—mine keeps the LEDs on until the batteries are charged, and
then turns them off. But why is the black one (5 hours) marked “fast” when the white one
(ostensibly 3 hours) isn't?
Stupidly, I didn't measure the voltage before charging, but after charging the voltages were
pretty consistent: two with 1.861V (including the one that I had previously marked as having
a low voltage), and one each with 1.860V and 1.859V. Possibly the voltage is also dependent
on external considerations such as contact impedance in the charger.
There are a number of things in the garden that I've been neglecting, and today I looked at
some of them. One was the plants that I have been trying to raise from seed for some months
now—that's fairly simple, since most of them either didn't sprout, or they died. Found
three Chile poblano seedlings which
looked relatively healthy, and planted them.
In the process, looked at some particularly unhappy looking tomato seedlings in the
greenhouse. Some of them are beyond hope, but others, despite being in too small a pot,
might still make it. So off to plant them in the veggie garden, another area that I have
been neglecting for too long. It's somehow discouraging to see the plants I have planted so
carefully (OK, not so carefully) being outdone by chance seedlings:
That also makes it clear that it's high time to weed the patch. But there are other urgent
things to do to, including pruning the ornamental vine on the verandah, which is now
covering most of the beams and joists, and which will hopefully cover the entire area by the
end of the season.
The tomatoes aren't the only area where chance seedlings outstrip my plantings. Two months
ago I planted some Lobelia seeds, and
they're only barely visible (about 2 mm across, in an egg carton), while a grass bush has a
chance infestation of happy looking flowers that must have self-seeded from last year's
flowers:
Discussing my lack of success with seeds yesterday, Peter Jeremy commented: “You shouldn't
get discouraged until the veges growing out of your compost heap outdo the ones you planted
and nurtured”. But that doesn't help. Here's my compost heap and some of the plants that,
I think, won't make it:
Another plant that I think won't make it is
the Ficus benjamina. I had
already noted that there was more sun in the area than I had expected, but it took very
unkindly to the situation. Here two weeks
ago and today:
A surprising number of new plants and flowers have cropped up over the past week. One is
a grass that we planted a long time ago, and which we had never expected to flower:
Elsewhere, the Stachys
byzantina are gradually coming into flower. I'm still not sure that they're completely
in bloom, but they look much nicer than the one in Wikipedia (last image):
For once the weather was neither too hot, too cold nor too windy. Spent some time weeding
the north bed—which I've done before—and extracted enormous quantities of weeds from a very
small area. Clearly I need to mulch as soon as I've removed the weeds.
And I had plenty of space in the east bed that I needed to mulch, including beds where I
wanted to plant the remains of the seedlings I got at
the beginning of last month. Put in several barrowloads of mulch and planted the
remaining seedlings:
Bought a strange device on eBay today, a combined
flash card and SATA disk docking
station. I'm not convinced that it will work well, but it wasn't expensive, so we'll see.
But what got me was the quick shipping:
Date: Sun, 11 Dec 2011 18:46:15 -0700 (GMT-07:00)
From: eBay <ebay@ebay.com>
To: groggyhimself@lemis.com
Hi groggyyourself,
We are writing to inform you that we have shipped the
item 110786777659 to you! Normally the shippment to worldwide is used to
take 8 to 25 business days,because it is a Cross-border
transactions.
After adjusting time zones, two things are clear: this message was sent less than 80 minutes
after I purchased the item, and at 9:46 on 12 December in Hong Kong, where the seller does
business. Yes, it's possible that they really did send it that early, but given the
delivery time, you'd expect them to do that sort of thing in the evening.
There have been other similar cases, notably with the Nickel-Zinc batteries that I bought a couple of months ago, also from Hong Kong. On
that occasion I received a similar message telling me that the item had been shipped on
27 September 2011 (US time), really 28 September 2011,
but the post stamp showed that they had been shipped on 10 October 2011.
“Fwaggle”, or maybe Jamie (surname unknown) on IRC came up with the explanation: this
message is automatically generated when the mailing label is printed. That would also
explain the incorrect time zone and the early morning in the correct time zone. It seems
that eBay can't think of a better text (“being processed” comes to mind), and the seller
can't do much about it.
TV reception is still very flaky. I'm becoming more and more convinced that it's
interference. Today I found a recording floundering round 900 MB after an hour of
“recording”. Clearly it was toast, but it was worth trying recording on other tuners.
Tried recording the same programme on another tuner, and a different programme at the same
time on the third. The results:
Programme
Date
Start
End
Daisy chain
File
Number of
name
time
time
Channel
Tuner
position
size (GB)
recoding errors
All I Want
12 December 2011
11:57:05.212
14:30:00.028
2006 (PRIME)
1
3
1.1
22, died at 51%
Test recording
12 December 2011
13:57:35.363
14:32:00.274
2006 (PRIME)
2
1
0.3
42, died at start
Test recording
12 December 2011
13:58:32.614
14:33:00.804
2203 (SBS HD)
3
2
2.4
0
This shows two things: firstly, it's nothing to do with the cabling to the tuners, since the
good recording was in daisy chain position 2. And clearly the problem is related to the
channel—either transmission problems or interference at specific frequencies. I'd call them
up, except that I'd just end up talking to script readers.
So now I have a total of 13 books from two different libraries about bulbs and other like
plants. I've already established that one of them doesn't show any corm that looks even
remotely like this one:
What about the others? The plants I know
are Watsonia
and Chasmanthe floribunda.
I finally found the answer: the third kind
is Crocosmia. But the books didn't
help much, beyond showing some flower images that I can correlate with my older photos.
Here's an overview of the coverage. The numbers are the number of books that fulfil the
requirement:
Image for
Other
Image of
Plant
Mentioned
identification
image
Tuber
Chasmanthe
2
2
0
0
Crocosmia
12
8
4
0
Watsonia
10
4
7
1
That's really not much use. Most of the images of the Watsonias don't show enough of the
plant to help distinguish it from Crocosmia, and I'm sure I've confused them in the past.
What's wrong here? For this kind of information, books are obsolete. Photos are expensive
to print (“A picture is worth 1000 words, but a good photo takes a million bytes”). Books
get out of date easily (the oldest one in my collection was published in 1967, and even the
newest is 5 years old), and they're constrained by size. Many are also geographically
constrained: only one of the books appears to refer to the Southern Hemisphere, and so
indications like “flowers from September to October” are meaninglessly ambiguous.
Clearly we need a web database, and Wikipedia is doing a good job in that direction. But somehow it seems to be tied up in copyright
issues; that's particularly evident in the photos. Hopefully the future will bring less,
not more copyright restrictions.
Didn't have much to do today, so did little of it. Spent a little time mulching, pruning
and tying up creepers, even planting
a Jasminum polyanthum. Also
transplanted a couple of
volunteer begonias, one from the pot with
the Mandevilla and the other from a
too-dry part of the succulent bed north of the verandah; put them both in the bed round
the Ginkgo, where there's a third plant.
I wonder where they came from.
About the most energetic activity was mowing the lawn, which Yvonne did. And then she ran out of petrol, filled the tank, and the thing wouldn't
start again. On the face of it you'd think that was flooding, but the behaviour wasn't
typical: after a while, it started again, ran, and then stopped, which suggested fuel
starvation to me. Played around looking at the fuel and air filters, both OK. Does the
system need bleeding? No mention of it in the manual. So left things at that; we'll see in
the morning.
Once again TV reception is terrible, at
least on some channels. Over the months I've eliminated a number of causes, including poor
cables—I think—and cross-talk between the tuners. And more and more it seems to relate to
specific channels and specific times of the day.
But at the moment it seems to be happening all the time on some channels. At the very
least, I need to find a way to monitor what's going on. One starting point is the
program femon that I looked at a few months
ago. It's like tzap, but
it works when MythTV is running. The output is
a little hard to interpret, but I can change that—if I could find the sources. Did a
bit of looking and found
So it's part of szap, whatever that may be. Went looking on Google and relatively high up on the hit list came a message
from Jürgen Lock, directly above the Schweizer
Zuchtgenossenschaft für Arabische Pferde web site, something that addresses other
interests. Jürgen was on IRC, so I asked him. He reminded me that szap is the satellite equivalent of
(terrestial) tzap, and pointed me to another
mail message with pointers to the source, accessible via Mercurial:
hg clone http://linuxtv.org/hg/dvb-apps
cd dvb-apps/util/szap
gmake 'CFLAGS+=-I/usr/local/include -DO_LARGEFILE=0'
That didn't build under FreeBSD, but I didn't
need it to. It didn't build under Linux in its entirety either, but I didn't need most of
the stuff. What I did need was lib/libdvbapi and, of course, utils/femon.
That worked and showed slight changes since the version I have. A bit of playing around
with the source enabled me to add a few options (reporting interval and output format) that
helped me get something more like what I'm looking for.
For once, hacking the source was simple. What's more interesting is to know what I'm
looking for. The code of the program is split into a couple of gratuitous functions that I
gradually removed again: it just required more and more parameters, and ended up being a
little silly. But basically it calls a function dvbfe_get_info (), presumably
in libdvbapi, which returns info in a struct dvbfe_info. Took a look in
that and found little further of interest except frequency, which would have been
really useful, except that it was always the same, and didn't match any valid frequency
(round 134 MHz). So for the moment I'm looking at things like this, for a good channel
(SBS, at least for the moment) and a bad one
(PRIME):
2011-12-15 10:40:51 Adapter 0: status SCVYL signal 193, S/N 202, noise -10
2011-12-15 10:41:51 Adapter 0: status SCVYL signal 194, S/N 199, noise -6
2011-12-15 10:42:51 Adapter 0: status SCVYL signal 195, S/N 203, noise -9
2011-12-15 10:43:51 Adapter 0: status SCVYL signal 194, S/N 200, noise -7
2011-12-15 10:44:51 Adapter 0: status SCVYL signal 194, S/N 200, noise -6
2011-12-15 10:45:51 Adapter 0: status SCVYL signal 195, S/N 204, noise -10
2011-12-15 10:41:19 Adapter 1: status SCVYL signal 180, S/N 159, noise 20, 488277 block errors, 2214 uncorrectable errors
2011-12-15 10:42:19 Adapter 1: status SCVYL signal 182, S/N 164, noise 17, 30353 block errors, 1 uncorrectable errors
2011-12-15 10:43:19 Adapter 1: status SCVYL signal 185, S/N 169, noise 15, 258438 block errors, 28 uncorrectable errors
2011-12-15 10:44:19 Adapter 1: status SCVYL signal 169, S/N 151, noise 18, 6496 block errors, 2435 uncorrectable errors
2011-12-15 10:45:19 Adapter 1: status SCVYL signal 188, S/N 176, noise 12, 23614 block errors, 27 uncorrectable errors
By contrast, the old version of femon reports, for adapter 1,
status SCVYL | signal 67% | snr 61% | ber 102521 | unc 7078 | FE_HAS_LOCK
status SCVYL | signal 67% | snr 60% | ber 102521 | unc 3085 | FE_HAS_LOCK
status SCVYL | signal 72% | snr 66% | ber 102521 | unc 27 | FE_HAS_LOCK
status SCVYL | signal 67% | snr 60% | ber 303808 | unc 27 | FE_HAS_LOCK
Is that an improvement? I don't know. Certainly measuring signal
and signal-to-noise ratio
in percentages doesn't make much sense to me. Bit it certainly makes it very clear which
tuner is receiving a good signal and which is receiving a bad one.
The real issue is: what do these numbers mean? My intention here is to distinguish between
poor signal (fault of the transmitter or propagation) and noise (potentially interference).
If I can believe the numbers, it would seem that both applies. And I still need to find a
way of correlating the output with a frequency, which may require going through the MythTV
logs and creating separate information about which frequency the tuner is tuned to at a
specific time. It certainly makes a difference.
2011-12-15 16:15:16.087 TVRec(1): ASK_RECORDING 1 0 0 0
2011-12-15 16:15:16.159 TVRec(1): Changing from None to RecordingOnly
2011-12-15 16:15:16.194 TVRec(1): HW Tuner: 1->1
2011-12-15 16:15:16.370 Started recording: Sea Princesses "The Missing Princess": channel 2022 on cardid 1, sourceid 2
2011-12-16 00:01:30.735 TVRec(1): ASK_RECORDING 1 29 0 0
2011-12-16 00:02:03.090 TVRec(1): Changing from None to RecordingOnly
2011-12-16 00:02:03.124 TVRec(1): HW Tuner: 1->1
2011-12-16 00:02:03.257 Started recording: Carnage: channel 2032 on cardid 1, sourceid 2
It seems that the tuners remain tuned to the old channel (in this case ABC 2) until changed. The
corresponding femon output after tuning to channel 2032 (SBS 2) was immediate:
2011-12-16 00:00:46 Adapter 0: status SCVYL signal 146, S/N 194, noise -49, 22660 block errors, 1288 uncorrectable errors
2011-12-16 00:01:46 Adapter 0: status SCVYL signal 151, S/N 198, noise -48, 357 block errors, 1294 uncorrectable errors
2011-12-16 00:02:46 Adapter 0: status SCVYL signal 194, S/N 205, noise -12
2011-12-16 00:03:46 Adapter 0: status SCVYL signal 192, S/N 206, noise -14
But this also suggests that the noise was lower before. I'm beginning to wonder whether the
signal-to-noise ratio has any meaning, or whether it's just an absolute signal quality
indicator.
Again didn't do much in the garden. Established that the lawn mower started without any
problem, so presumably, despite all indications to the contrary, it really was flooded. But
that's Yvonne's job, and she had other fish to fry:
We've been considering a remote controlled “electric collar” for Nemo for some time. The idea is to give the dog a mild electric
shock if it is disobedient. Traditional dog trainers are, of course, horrified about the
idea of giving the dog an electric shock, and it took us some time to accept the idea. From
a purely training point of view, of course, most trainers accept that animals must be
punished for disobedience, though rewards for good behaviour are much better. But in
general punishment requires the proximity of the animal. Horse trainers can sing a song
about that one.
So any device that can give the animal the impression that it can't get away from you that
easily can be a good training help. We've used water pistols on puppies and kittens for
years, and it hasn't broken them. On the other hand, we once had
a Borzoi bitch—a real bitch!—who
continually used her speed to get away from punishment. One day we chased her with a horse,
and she was so shocked that she left home and had to be picked up a couple of kilometres
away. So caution is required.
But after Nemo nearly got himself killed chasing a kangaroo last month, it was clear that the electric collar might be the lesser
of two evils. And any training tool can be abused. You shouldn't use whips on dogs, but
Yvonne typically goes walking with
a stockwhip over her shoulder (shown in
use courtesy of Rob Pike in the third
image):
The purpose of the whip isn't to hit the dog—that would severely injure it—but to make a
noise to which it pays attention.
The collar arrived today, and it took me 10 minutes to work out how to use the thing,
greatly hindered by the instructions. It seems that every time it powers up, it
needs to be synchronized with the remote control (wait for beep from collar, press function
button on remote control, wait for second beep from collar). And the collar powers down
(only) after a certain time without movement. That means you can't leave it on the dog, or
the batteries will drain quickly, and every time you put it on the dog, you have to resync.
Still, the idea of the vibrator and the sound signal is good. An intelligent dog will react
to that, and will probably never need more than one or two shocks to remind it of what would
come next if it disobeys. All that is in the future, though: the instructions recommend
that the dog wear it for a month to get used to it before using it at all. I think we can
restrict that to a week: Nemo is used to having a harness on when he goes for a walk, and
this doesn't make much difference.
So I've identified that I have at least three kinds of similar looking bulbs or corms. The
only ones I'm reasonably sure about are
the Watsonias, which have bulb-like
corms and flower in the summer. They have alternate flowers:
On the face of it, not much similarity. But I'm not convinced. The Watsonias start off
with shoots which look almost identical to those of the Crocosmias. Today I took some more
photos:
On the face of it, the first two (same plant) are Watsonias, and the other is a Crocosmia.
But I can't see a clear distinction. The Watsonia is flowering, the “Crocosmia” is not.
I wouldn't be sure that they're not the same kind.
What I have established: we have two different flowers that bloom in the spring (tra
la), and at least one of them has the flattened corms. But there's also this one, which at
one point I had considered to be
a Chasmanthe floribunda.
But apart from the time of blooming (July to September), it looks almost the same:
About the biggest difference is the height of the flowers, way above the leaves. Is this
what happens when it's finished? Looking at the photos, it looks as if this is the same
“plant” (clump of corms) as the second photo above:
On the face of it, these flowers look almost like little onions. Maybe it's something else
again that I have forgotten; certainly there's only one place that looks like that, (now) to
the south-east of the pond.
There are yet other flowers of interest at the moment.
The Strelitzia reginae seem to
have somewhat varied flower forms. Last year I had a double flower, one pointing in each
direction:
House photos again today, and once again problems. For one thing, it was windier than I
would have liked, making merging HDR images difficult—how I wish people would come
out with digital sensors with a higher dynamic range. I also managed to take most of the
photos with the focussing rail set off by a couple of centimetres, which in fact didn't make
as much difference as I had feared. In addition, discovered what looks like a firmware bug
in my Olympus E-30.
I'm taking the verandah photos in groups of three, and only the first is supposed to have
flash. So I take one image, turn the flash off, and take the other two. I get it right in
the majority of cases, but it seems that when turn the flash off at the wrong time, the
camera firmware hangs. I think it happened last week, and I solved the problem by powering
off the camera. Today it hung twice, and each time it recovered when I turned the flash
back on.
Last week I set the flash to +3 EV, which proved too much. This time I set it to +1EV, and
also used flash to take the zenith shot (twice, in case one reflected), and in fact that
worked out quite well, within lighting limitations. But it's becoming increasingly clear
that a simple flash for one of the images isn't enough. I'm still getting some areas
overexposed, and the lighting is so different that Hugin has difficulty identifying the
relationships. So did I. Here a dark corner, the column close to the camera, and the
zenith. All three should join:
Apart from that, I managed to take the upper layer of images at too low an angle, so the
zenith shot wasn't enough to cover the top of the sphere. But by the time I found that out,
I couldn't be bothered any more. Next week will be better, I hope.
Nemo had been told to wait for a signal before eating,
and I had wanted to take another photo of him finally eating the food. But the flash
unit—again!—gave me problems. It fired at full strength, and of course by the time it had
recycled, Nemo had eaten it.
What causes that? It happens now and again. My best bet is that the contacts between
camera and flash aren't as good as they should be; they're based on decades-old technology
and are probably not well-suited to high-impedance digital circuits. I suppose I should try
keeping them cleaner, but it'll take a while to see if it makes any difference. Somehow
flashes are more trouble than all other things put together.
In Australia we've had to put up for years with incompetent legislators interfering in the
network infrastructure. Twelve years ago they brought out the Broadcasting Services Amendment
Act (BSA), designed to stop filth on the Internet, or some similarly vague idea. It
was passed, implemented and forgotten. And somehow it seems appropriate that the Government
web site with the text of the act should be overloaded on a Sunday morning:
Today I can't even easily find a clear reference to what it was intended to do. I know that
when I got some spam that was (marginally) obscene, I contacted the ACMA and reported it, only to be told that the BSA did
not apply to email.
But now we have a new government, and
a new minister for “Broadband,
Communications and the Digital Economy”, whatever they think that means. And he, too, wants
to stop filth. He'll censor the Internet! Everybody in the industry tells him that it's impossible, but he knows better, and it's
Going To Work. Well, at least it's going to take a lot of work.
From time to time, people in more enlightened (or is that backward) countries point to our
legislation as an indication of our repressive regime—less often of downright stupidity.
And of course, China's—relatively effective—censorship is completely unacceptable in a
Western society, and Hillary Clinton called for an end of censorship, apparently likening it “to the rise of communist Europe,
warning that a new "information curtain" threatened to descend on the world unless action to
protect internet freedoms was taken”.
That's easy to say when other countries are doing it. But there are at least two legitimate
issues here: firstly, the Internet does need some kind of regulation, like any other
part of society, despite what the freedom fighters say. For example, few people would
disagree that something should be done to stop spam, and the abuse of the Internet for
distributing child pornography is particularly ugly. Secondly, though, and far more
important, those who seek to regulate it must understand it. And, worldwide, they don't.
Now the problem has hit (the United States of) America. That's probably a good thing. A
large proportion of the world's most influential techies are either American or
Americocentric. And now the US Congress has shown that they're by no means lagging behind
the idiots in other countries, and have brought out a bill designed to break
the Domain Name System. With
the best of intentions, of course—aren't they always? It's called
the Stop Online Piracy Act
(SOPA), in itself a good idea—you'd think. But it's censorship. How does that match
Clinton's statement? And what does it have to do with the DNS? Well, that's how they're
going to implement it: intercept DNS replies, it seems.
And while the House of Peers withholds
Its legislative hand,
And noble statesmen do not itch
To interfere with matters which
They do not understand,
As bright will shine Great Britain's rays
As in King George's glorious days!
It's amazing how modern that operetta still appears. Other songs presage changes that
happened in the British government over a century later. Wouldn't it be nice if legislators
could understand it?
Another day with little to show for itself. Spent the morning finishing off my house
photos, and in principle should have done some work in the garden in the afternoon. But
once again it was so windy that I couldn't be bothered. Pulled out some more of the flat
corms, which I still haven't identified to my satisfaction. It seems that a lot of them
didn't flower, though they're clearly past flowering now. Maybe they are
all Chasmanthe floribunda
after all. I should take them in and show them to the Friends and ask.
I have divided my diary entries into 10 different categories, regarding various things I do.
Many blog systems have an order of magnitude more, but I think 10 are enough; in general it
makes more sense for people to display all and skip stuff that doesn't interest them. But
the titles of these categories awake certain expectations: the term “computer” increasingly
leaves out large areas of digital technology. I also have “photography” and “multimedia”,
and frequently the topics overlap.
That's OK, since I can choose as many topics as I like, but maybe the tag “computer” is
misleading for many topics. So I've renamed the topic to “technology”, really short for
“digital technology”, which would be untidily long. That's retrospective, and I'm not
convinced it's the right choice, so it may change again. But it's clearly the right choice
for things like the nitty-gritty about digital cameras and digital TV reception, so if I
change it again, it probably won't be back to “computers”.
Into town this morning to see the doctor; a routine checkup for my blood pressure.
Surprise: the appointment was tomorrow. But they measured my blood pressure anyway, and
found it a little low. So the doctor saw me anyway and took me off the medication for a
while. Time to maintain better records of my blood pressure.
DxO Labs have come out with a new version of their
DxO Optics "Pro" package, which I tried some months ago, and
which I had ultimately found too slow for serious use. The new version, they claim, is up
to four times faster (“a remarkable speed”).
But, again, it only runs on Microsoft and Apple, and the Apple version always lags the
Microsoft version. So tried once again to borrow the computer that Chris Yeardley lent me
last time. Surprise: I already have it, and I'm running my Internet gateway on it. But
this is just for a test—I must really get the 64 bit version of FreeBSD running on dereel so that I can run
Microsoft in a VirtualBox with sufficient
memory—so she offered to lend me her main laptop, with 4 GB memory, until the end of the
week.
Problem: it runs “Windows” 7 (a number that I don't
understand; shouldn't it be at least 8?). First thing, of course, was to connect it to my
network. I even know how to do that under “Windows” XP. It involves going to the control
panel and selecting various non-obvious menu items. I suppose it's the same here, but now
there are different non-obvious menu items:
I don't run a DHCP server in my network, so
I needed to set the IP address manually. “Internet options”? Of course not. That's a
browser function! Instead, I finally found it under “View network status and tasks” →
“Change adapter settings”, conveniently hidden in a side bar. After that, things looked
much the same as under XP.
Next was to mount Samba shares. What are my
options?
Clearly that's sharing. But that tab led me to a concept called variously “homegroup” and
“HomeGroup”. Much querying and I found a pretty picture showing two systems on the network:
the laptop and cvr2, my MythTV box, a
“MythTV AV Media Server“. That was certainly unexpected, and to my surprise I was able to
connect to it and view videos with no more than a few mouse clicks. Score one for Microsoft
(or was that MythTV?).
But I couldn't find my Samba servers, and there was nothing to tell me how to do it. Found
an article on the Samba web
site, which worked, but it wasn't until I worked my way through that I discovered it didn't
do what I wanted to do.
Then Callum Gibson suggested I should try a different tack. Start “Windows Explorer”
and... Wait a while, how do I start “Windows Explorer”? No idea. It's not there in the
list of programs:
It seems that the canonical way is to press (Broken window key)-E. What's
wrong with this picture? Programs have had names since the mid-1950s at the latest, and
menus haven't changed that. They've just claimed to make it easier to find programs. At
least “Windows” 7 fails here: I had a couple of other people trying to help on IRC, with
comments like:
<Darius> they hid the start menu shortcut for explorer pretty well in win7
Finally I got it started and dragged a “short cut” (apparently an emasculated graphical
symlink) to the root window, then started it. I had to select not networking, but “My
Computer”—isn't that obvious? No. Microsoft have changed the terminology, or maybe it's
because the computer really belongs to Chris Yeardley. Anyway, the menu item is now called
“Computer”, another gratuitous confusion. Then I was able to select “Map Network Drive” the
way I've been used to it. And it didn't work. Ran the “troubleshooter“ with a predictable
outcome: “Troubleshooting couldn't identify the problem“. Finally discovered that I
couldn't ping either. The problem was in my DNS settings. I needed to set the equivalent
of these two lines from /etc/resolv.conf:
domain lemis.com
nameserver 192.109.197.135
For that I had to go through a hierarchy of 9 menus and set things in two different places,
and I had forgotten to set the domain name (sorry, “Append these DNS suffixes“). After
that, things worked OK. Maybe that was the only problem.
So why was it so much trouble? Mainly because of the terminology and the location of the
individual settings. I don't know if the current arrangement is the result of history or
careful deliberation, but the evidence suggests the former. Then Microsoft has gone and
changed names and terminologies, and introduced “homegroups” where you'd expect to
find SMB shares, and then
hid the SMB share elsewhere, so it would have been reasonable to assume that “HomeGroups”,
too, were an obfuscatory renaming of SMB shares.
In any case, it finally worked, after only about an hour, and though I can't use remote
desktop (why would anybody want that in a “Home Premium” system?), I was able to install
TightVNC to access the system remotely even
more slowly than I already do with the Apple. So, at least for the experiment, things
work. But how can people live with this stuff?
So finally I got Microsoft to the point where I could install DxO Optics "Pro", another
200 MB of package. It downloaded relatively quickly, and once again I was left scratching
my head as to how to use it. But it's user friendlier now: it now allows you to select
files instead of “projects”, like all other programs in the Microsoft space by clicking on
silly icons. For the test I had chosen the photos I had taken on 10 December 2011, all 371 of them. And of course it had to display icons for all of
them, which took about 5 minutes.
Finally I was able to tell it to process the images (convert them
to JPEG). Estimated time: 5 hours! It
would only have been 3 with Olympus
Viewer 2. Let it run, and to my surprise it was finished in 70 minutes, roughly the 4
fold increase in speed that they had claimed. So maybe there is some use in the thing after
all. Next I need to compare the output from the two programs and also see how fast Olympus
Viewer runs on this platform.
Not surprisingly, I didn't get much else done today. Did a bit more tidying up in the
greenhouse, which is getting emptier and emptier. Also some wide-ranging planning. The
weeds are gaining the upper hand again, and I think I'm going to have to remove the plants
from some areas, spray and mulch, and then replace the plants. That should work for plants
like Gazanias.
Spent some time this morning considering how to compare the results of DxO Optics "Pro" with
other image processing methods. Came up with a relatively mechanical system where I take
two trees with files with the same name and generate 5 crops from each: top left, top right,
centre, bottom left and bottom right. Then I created a web page with each image, using the
established technique of mouseover image switch to show a direct comparison of each image.
The first attempt was simply a single line per image (old and new), so it's not immediately
clear what part of the original image it represents. In addition, I made the mistake of
comparing a big directory—371 images—so I had a total of 3710 small images to compare, a
total of 174 MB of them, quite a bit for a single web page. I'll have to try something a
little more sophisticated. Even so, though, it showed some interesting details. Here an
extract from that page. In each case, the image on the left was processed by DxO and
changes with mouseover to the version processed by Olympus
Viewer 2:
The original version of this entry showed two images, one changeable (the one kept
below) and one static like the changed version. That proved to cause significant
problems with the RSS versions.
The underexposure here is deliberate. This crop (extreme top left) shows much better shadow
detail in the Olympus version. This may be unfair, since I turned exposure compensation off
on DxO, but I didn't think it was on with Olympus either. Like all the images, it shows
significantly different distortion correction. I'll need to experiment further to decide
which corrects distortion better—possibly Hugin can be the judge of that.
Here there is significantly worse chromatic aberration with Olympus. But that's manual;
conceivably I could get just as good results with a little tuning.
Here again the exposure is not as good with DxO, and the chromatic aberration is better.
In summary, though, I'm surprised by the good quality of all the images. These are the
extreme corners of the image at natural size (the pixels are rendered one to one). At the
100 dpi resolution of my screen, the entire image would be about 100 x 75 cm in size. But
I'll have to do more work to make it easier to compare the images.
It was rather stupid that I went into town yesterday for today's doctor's appointment,
because I had to go to the optometrist today anyway. Took the opportunity to get a
haircut—it must have been four months since the last time.
Then on to the Friends of the Ballarat Botanical
Gardens with some of my flat corms for identification. I don't know whether I should
be happy or sad that nobody could positively identify them. Bruce Holland at first thought
that they might be Watsonia, a
surprising idea from an expert, considering that my Watsonias had completely different
corms. But when I told him the flowers were orange, he said they couldn't be Watsonia:
Watsonias were blue, again not what I knew. Yvonne C wasn't sure either, but decided that
they were some kind of invasive species, and didn't want any of them. She's certainly
correct that they're prolific, but I haven't seen any evidence of invasiveness.
Back inside and talked to Bruce again. He grabbed a big book—which didn't show the
corms—and came up with a picture
of Chasmanthe floribunda.
So yes, it seems that at least these ones are Chasmanthe. And the summer flowering ones?
And this one in the garden?
In any case, checked further in the Big Book and found Bruce's blue Watsonia, a different
species from the ones I have. At least that clears the doubts about his expertise.
Did a bit of talking to Mike Sorrell and Maree (surname forgotten). Mike gave me Yet
Another Buddleja “for Christmas”, one
that apparently has very dark blue blooms when they first open, and Maree convinced me to
buy a Campanula cochlearifolia“Bavaria Blue”, a blue flowering ground cover and the sort of thing we need in the garden.
To Specsavers after that for an eye
test, once I found them—they were nowhere near where Google Maps put them. They're certainly a busier place then
Kevin Paisley “fashion
eyewear”, and they had some interesting machines to check my eyes even before I was sent in
to the optometrist, including an eye pressure measurement done with a puff of air, and a
retinal photo.
The eye test itself was nothing special; slight deterioration, as I would expect, and a
choice of new glasses. Finally I think I'm going to have to give up with glass lenses;
they're almost impossible to find, and they seem to significantly limit my choice of
multifocal styles. So I'll try plastic—again—and maybe this time it'll be OK.
Date: 19 Dec 2011 10:22:48 -0000
From: saopun Moderator <saopun-owner@yahoogroups.com>
To: groggyhimself@lemis.com
Subject: Welcome to the saopun group
Message-ID: <1324290168.511.9229.w8@yahoogroups.com>
I've added you to my saopun group at Yahoo! Groups, a free,
easy-to-use service. Yahoo! Groups makes it easy to send and receive
group messages, coordinate events, share photos and files, and more.
Description of the group:
------------------------------------------------------------------------
ebey4qk1012u04ge8u1
Complete your Yahoo! Groups account:
...
The description is enough to know that it's spam of some kind. The correct thing to do with
this should be to send it to abuse@yahoogroups.com, and they should deal with it.
But I've already established in the past that Yahoo!, like other quality sites such as Google, want you to fill out web forms to complain about what they have allowed
happen to you. So off again to try it out. But this time I didn't find any way to send the
offending mail, just a really clever (and suboptimal) tip to tell me how to unsubscribe.
Clearly they don't care.
In fact, to unsubscribe all you need to do is to reply to the message. It contains a
header:
Reply-To: saopun-unsubscribe@yahoogroups.com
And of course that results in a “Do you
really want?” message from Yahoo! But it's really objectionable that these sites seem
to care less and less about the fact that people are being inconvenienced by actions
stemming from their sites.
Yvonne returned from the morning walk this morning without
Nemo: he had done another runner into the lagoon, and
returned later with proof that there's still water in the lagoon.
Clearly he hasn't learnt from his experience in
the Wimmera last month. Time to put the electric collar into action—if he
will pay attention to it. Tried it out, and it wouldn't register with the remote control:
it was full of water. That's a clever construction. Took the collar apart to dry it out;
hopefully it will still work when it's dry again.
Spent much of the day playing around with methods to compare software, and finally came up
with a set of pages showing comparisons between the four corners and the centres of photos
processed in two different ways. It was a lot of work, and it's not perfect, but it shows a
surprising number of things. Firstly, the DxO “HDR” function makes so little difference that it's barely recognizable. I spent about 10
minutes trying to work out why my mouseover function no longer worked, until I realized that
the images were effectively identical (though with careful examination minor differences
were apparent).
Took a number of photos with different lenses. In each of the following examples, the DxO
version is with the cursor away from the image, and the mouseover version was converted with
Olympus Viewer 2. In each case I used the standard settings. This one was taken with the Zuiko Digital ED
70-300mm F4.0-5.6 telephoto at 70 mm. Here the “thumbnail”, which shows a
significant difference in the framing between the two programs:
There's noticeable chromatic aberration at the top left corner, and DxO seems to do nothing
to correct it. Olympus doesn't have automatic correction for chromatic aberration, and I
had thought that would be a great advantage of DxO, but this suggests that it's not much
use:
Where DxO does seem to be better is with the issue of sharpness. Here the leaves of
the Acacia in the centre and bottom right
look considerably sharper:
But is that enough? Paying $100 for good software isn't much, but is this good enough? On
the whole, I prefer the gradation of the Olympus Viewer. More investigation needed,
including comparisons with Saturday's photos.
Callum Gibson is the only person I know who always reads
the RSS version of my diary. He had problems
with my report on photo stuff yesterday. It proves that the JavaScript tricks I'm doing
with the photos are incompatible with RSS, so I had to remove them and replace them with a
link to the diary.
I've already noted the strange shape of
some Strelitzia reginae
flowers. Last year I had a clear double flower, one half pointing at 180° to the other:
This year I don't have that, but there's another flower which seems to be a triple decker.
The flowers should look like the first one, but the other seems to have three flowers on top
of each other. The blue petal pairs (apparently a “nectary”) are present not once, but
three times, though the older ones are dying off:
Lately we've been having “almost” power
failures. This morning at 1:27 I heard the UPSs beep, then stop, then beep again. My
bedside alarm clock failed. But that was all; nothing else in the house had problems.
Power failure or not?
Once again we have lots of plants to plant, and gradually we're getting up enough enthusiasm
to plant the area round the pond. We still have one kind of grass that I originally thought
might be appropriate for the marginal area of the pond, but then decided that it wasn't. But I didn't know what it is, so I'm not
sure how I decided. Made a compromise: split the pot into two parts and planted one in the
marginal area and one outside. We'll keep an eye on the former and remove it if it starts
showing signs of drowning.
Then Yvonne wanted to transplant the Phormium “Jester” that
we bought two years ago. It had completely
filled out the pot, and Yvonne now wants it in the garden. Took half an hour to dig it out
of the pot, and then we were no longer sure where we wanted it, so left it half out of the
pot to decide.
Chris Yeardley had some concerns about web site security related to Google Chrome. They were clearly unfounded—but how
many times do I say things like that only to be proven wrong? Clearly it's worth trying out
to be completely sure. Since I still have her laptop, tried installing Chrome on it. The
download page I got was amazing:
Callum Gibson still wasn't happy with my copout in serving
the RSS of my diary. It seems that there
were two issues: relative URLs (a problem I've had to deal with again and again) and
the onmouseover functionality. If I solved the first issue, I could at least
display the “before” image, though experiments showed that onmouseover still broke
the rendition in newsfox. Ended up writing a function to do the switch in HTML and just show
the former image in RSS. I suppose Callum's happy now.
Nemo's electric collar was dried out today, and it
works again. Went out for a walk to try it out. Not a revelation. The range seems to be
about 10 metres when you're lucky, and you can't rely on it to work further away than 3
metres—less than the length of the leash. Nemo noticed the vibrator when it was set higher
than about 30%, and looked at us with interest. Certainly nothing to stop him in his
tracks. If the shock works at all, he did a good job of ignoring it.
There are two issues here, of course: is this particular device a good example of its kind,
and does an electric collar make any sense for dog training? So far I'm tending to answer
both questions with “no”.
Notification slip in the letter box today: a package had arrived for me. I've been waiting
for a rotator for my camera, so went in to the post office to pick it up. And of course it
was the other package, the eSATA docking
station that I ordered at the same time. Sure, I want that too, but it wasn't nearly as
important.
Unpacked the device and discovered it wasn't exactly what was advertised. Yes, it has two
disk slots, but one's for PATA and
the other's for eSATA:
The instruction booklet proved to be only to describe how to use the supplied software (for
Microsoft only, of course) and the on-device firmware for “hardware cloning”, copying a disk
in an unspecified direction. Given that most PATA disks are much smaller in capacity than
eSATA, it's not clear what us it is anyway.
Discovered I didn't have any free eSATA disks, so with some physical difficulty managed to
insert a PATA disk, which, as far as I can tell, didn't power up. Tried with eSATA
interface in defake, and got no response, possibly because the eSATA interface only
works for the SATA disk (not documented, of course). Gave that up as a bad idea and
connected to dereel via USB (Powercor has put paid to defake's USB subsystem) and inserted a handy 8 GB
SDHC card instead. Success! Well, not complete failure:
Dec 23 13:27:43 dereel kernel: ugen5.5: <vendor 0x1a40> at usbus5
Dec 23 13:27:43 dereel kernel: uhub8: <vendor 0x1a40 USB 2.0 Hub, class 9/0, rev 2.00/1.11, addr 5> on usbus5
Dec 23 13:27:44 dereel kernel: uhub8: 4 ports with 4 removable, self powered
Dec 23 13:27:45 dereel root: Unknown USB device: vendor 0x048d product 0x1336 bus uhub8
Dec 23 13:27:45 dereel kernel: ugen5.6: <Generic> at usbus5
Dec 23 13:27:45 dereel kernel: umass1: <Generic Mass Storage Device, class 0/0, rev 2.00/1.00, addr 6> on usbus5
Dec 23 13:27:45 dereel kernel: umass1: SCSI over Bulk-Only; quirks = 0x0000
Dec 23 13:27:46 dereel kernel: umass1:6:1:-1: Attached to scbus6
Dec 23 13:27:46 dereel kernel: da1 at umass-sim1 bus 1 scbus6 target 0 lun 0
Dec 23 13:27:46 dereel kernel: da1: <Generic Storage Device 0.00> Removable Direct Access SCSI-2 device
Dec 23 13:27:46 dereel kernel: da1: 40.000MB/s transfers
Dec 23 13:27:46 dereel kernel: da1: 7580MB (15523840 512 byte sectors: 255H 63S/T 966C)
That's an improvement on the current USB card reader I have, which FreeBSD doesn't recognize. The problem was that the I
couldn't read the card. I thought it was OK, but when I tried it later in boskoop
(the Apple) with the old card reader, all I got was:
So: the card seems to be scrambled. Did it happen before being put in the docking station
or later? To be investigated. In the meantime tried a disk, which didn't get recognized.
Ran a camcontrol rescan on it, and the system gradually ground to a halt and had to
be rebooted. I really shouldn't experiment with USB on dereel. But so far, the
device hasn't exactly crowned itself in glory.
While in Ballarat, picked up two DVDs
from the Central Highlands Library.
Nothing special. Then in the afternoon, the truck from the Geelong Regional Library came along with
another DVD for me. And this one, like so many before it, was unreadably damaged. From
memory, I've never had a damaged DVD from the Central Highlands library, but about half of
the DVDs from the Geelong library are unusable. Can I get them to fix it?
The vine growing on the verandah is growing like fury, and spent some time pruning it. I've
decided to train some shoots between the rafters, pointing east or west where they can get
held in place by the corrugations until they're firm enough to stand by themselves. Cut off
a surprising amount of foliage. This will probably be the last year I need to do any
training.
House photo day today, and also the last day for the free trial of DxO Optics "Pro".
Together they kept me going all day.
More firmware problems taking the verandah photos: there are clearly serious synchronization
problems between my Olympus E-30 and the Mecablitz 58 AF-1 O digital flash unit, including hanging the camera when the flash
is turned off and not firing when it's turned on again, although it was charged. The latter
problem went away when I power cycled the flash unit, but I ended up taking a total of 36
sets of images where I only needed 21.
Processed all the photos with DxO, first with the same settings as I used earlier this week. DxO didn't show itself from its best
side. At startup it seems to hang. How many mouse clicks do you need to get these
Microsoft-space programs to react? Sometimes it's a single click, sometimes it's a double
click, and frequently the programs seem to ignore the mouse altogether. In days gone by
they often showed an hourglass to say “Yes, I heard you, but I'm too lame to do much else
now”. That way I at least know not to click any more, but that no longer seems to be
modern. As it was, I ended up with some very delayed messages telling me “There can only be
one”.
That wasn't the only problem. DxO go to some detail in their relatively voluminous
documentation to stress the importance of
the EXIF data:
DxO Optics Pro must always be the first program to edit your images. It is important to
remember that EXIF data should not be modified or altered in any way, as this data is
essential to insure[sic] the selection of the correct optical correction module
and, therefore, the proper automatic optical corrections.
But how do you tell it to save the EXIF information in the
output JPEG image? I don't know. I
selected something that suggested it would give me EXIF output, but it only had the very
barest of information:
File Name : PC250764.jpg
Directory : .
File Size : 2.9 MB
File Modification Date/Time : 2011:12:24 13:30:01+11:00
File Permissions : rwxr--r--
File Type : JPEG
MIME Type : image/jpeg
JFIF Version : 1.02
Resolution Unit : inches
X Resolution : 314
Y Resolution : 314
XMP Toolkit : Image::ExifTool 8.50
Author : Greg Lehey
Image Width : 4032
Image Height : 3024
Encoding Process : Baseline DCT, Huffman coding
Bits Per Sample : 8
Color Components : 3
Y Cb Cr Sub Sampling : YCbCr4:4:4 (1 1)
Image Size : 4032x3024
And that's all. No camera information, no exposure details, just silly things like
“Resolution”. And without this information, Hugin is dead in the water. Yes, I know the problem, and I have worked out a
solution long ago. But why should I need it in a
supposedly high-quality program?
The program does cater for having the contents of the directory moved while it's
working; it has a “refresh” option which can be invoked by pressing F5, and apart
from taking an eternity to create pretty icons of the photos, it seems to work. Most of the
time, anyway. I had 187 images to process, but for some reason “Select all” chose only 55
of them—not once, but twice. A couple of times it just crashed, and it certainly loses
count when these things happen:
Many of the problems probably stem from the Microsoft attitude that the program should be in
charge instead of just being a tool; I continually have to work around that. But in the end
the conversions were OK,
After processing the photos normally, tried a couple of likely candidates with the DxO, and
then selected some for processing with
the HDR “preset”. How do I know whether it's
been set or not? For this image or for all of them? I still don't know, but the results
suggest that it has been set. There's clearly a lot to read up on here.
And the results? Parts of them were excellent. Here are four views with normal processing;
moving the mouse over the image will change it to the DxO-HDR version:
My impression is that the first two images are better with the full HDR processing, but the
other two actually look better with the DxO processing of a single image. The first two
images are also some of the most complicated. The first is, of course, the famous verandah
panorama where I use flash to improve the gradation, and for the second one I took 5 images
per position, using ones exposed at -1 EV, +1 EV and +3 EV. But that's the exception;
normally I take 3 images and use the ones at +0 EV and +2 EV.
Even in these examples, though, DxO has clear advantages. Since each position is a single
image, there's no ghosting, and in the second image the lack of halo round the tree at the
back on the right is obvious. So there is some advantage in the software after all, and I
bought it. It looks like it'll be a steep learning curve, though.
Christmas eve this evening, and Chris Yeardley was there, as just about every Saturday.
We're planning a real Christmas dinner on Tuesday, when
our daughter Yana and also Nele Koemle and her mother Magda are coming, so tonight we just
had a glazed ham. My mother made one decades ago, but I no longer had the recipe for the
glaze, so ended up making up my own version based on recollections and a recipe
in Stephanie Alexander's
The Cook's
Companion. Didn't taste bad, but somehow something was missing. I'm sure that the
following had nothing to do with the food:
I returned Chris Yeardley's laptop to her yesterday, as promised. But that wasn't the best
of ideas: now I have purchased DxO Optics "Pro" and
don't have any machine to run it on, and that on the day of the month where I have the most
photos to process. Over to Chris' place to borrow another machine, this one running
Microsoft “Windows” XP. Setting up was amazingly easy compared to “Windows” 7, and I was
going to praise it for ease of use until I discovered that this was a machine I had already
borrowed and configured 8 months ago and been
through similar pain then.
Still, the result was that I got up and running pretty quickly, though it's clear that this
machine is nowhere near as fast as Chris' laptop. On the laptop it took me about 11 seconds
per image; on this machine it took 70. That's quite an amazing difference considering that
this machine is not that old and has a 2.8 GHz processor. It brings home to me how
little the clock frequency means nowadays. But it also gives the lie to DxO's claimed
fourfold speed increase: in April I converted 72 images in “over an hour”. With this version it would take 84 minutes, possibly
even longer than with the old version. It's still on the machine; it might be fun to try it
out for comparison if it accepts the same activation key.
Last Sunday of the month today, and time for the monthly photos of the flowers in the
garden. In fact, I'm planning to move them gradually to the middle of the month, so until
then I'll go in steps of 4 weeks.
Once again we have a number of flowers blooming for the first time.
The Gladioli are just coming into
flower, as are the Kniphofia:
A number of the plants that we bought at Lambley
Nursery are now coming in to season: the
the Asphodeline liburnica
that we bought three months ago and
the Eryngium bourgatii. The Asphodeline
is particularly spindly, and the flowers only open when they want to. It's pretty, but
something that needs to develop to be really pretty. The Eryngium, on the other hand, is
quite spectacular:
And the Pyrethrum that we bought
at the end of August is growing like fury,
outpacing other daisy-like flowers, and apparently killing the flies fool enough to land on
the flowers. Judging by the flowers, it's really Tanacetum cinerariifolium.
The Agapanthus are flowering for the
first time since planting, as is
the Campanula
cochleariifolia that I bought only on
Tuesday:
The Santolina
chamaecyparissus is now flowering, as are the
volunteer Begonias that popped up in the
middle succulent bed and which I have now transplanted under
the Ginkgo:
Last year both kinds of Clematis
flowered at roughly the same time. This year the darker blue “vagabond” has already
flowered, and now the lighter blue “pearl d'azure” is taking over:
The Hebes are also going from
strength to strength. Like the Clematis, they're flowering in succession. The violet ones
(first image) are coming to the end of their season, and two others are just coming into
flower. There's still a pure white one which is on its way.
My red Hibiscus
rosa-sinensis has now recovered from last year's chill and is flowering happily again.
The flowers are up to 20 cm across. And this year the
Korean Hibiscus syriacus has
started flowering early—previously I haven't seen it flower until March.
The Mirabilis jalapa are another
flower that is particularly difficult to catch with the flowers open. The books say that
they flower in the evening, but clearly the flowers can't read.
For some reason, the Lonicera don't have
much to show for themselves yet, though they're not looking unhappy. Some have flowered a
little, but right now there's almost nothing. I wonder if it has been too warm lately.
I processed all the photos above with DxO Optics "Pro" and the
“realistic” profile for the “Single-shot HDR”. It's a good compromise. In some cases, I
think it would have been better without, but it reminds me of the first film I took in
an Asahi Pentax
“Spotmatic”, over 45 years ago. I didn't even need to print the negatives;
compared to the results I got with my SV (despite exposure meter),
everything just looked so evenly exposed that I was convinced. These results were similar.
Went to some lengths to compare them in more detail, including writing another script to
generate web pages, but it's clear that I need to do more than that. Things have to be easy
to handle the 170 photos I took today. So for today, at any rate, I've just left the “HDR”
versions. Next I need to learn more details of DxO.
In the afternoon, sitting at the computer, a sudden crackling noise came from the monitor in
front of me (:0.2), sounding something like a short circuit or a spark. It was
immediately followed by a loud thunderclap from outside to the west (towards Chris
Yeardley's place). She had been outside the garden and saw the lightning strike somewhere
close to us. No harm done, but the crackling noise from the area of the monitor suggests
strong electrical fields just before the strike.
It had one effect: it clearly punched a hole in the sky, and all the water ran out. In what
seemed to be only a minute, we had 2.8 mm of rain. From the IRC log, it can't have been
much longer:
* gr0ogle gets hit by lightning. [15:12]
<fenix> hey, how close was that?? [15:14]
<gr0ogle> I don't know.
<gr0ogle> Did it seem to come from my direction?
<fenix> sure did
<gr0ogle> I heard a sudden "snap" about 300 ms before the lightning.
<gr0ogle> From my monitor. [15:15]
<gr0ogle> Nothing seems to be damaged, though.
<fenix> oops
<fenix> I'm surprised the power hasn't blipped
<gr0ogle> For the rest of you: fenix lives about 1 km direct line from me, so the flash
must have been in that range.
<fenix> I was outside to close the car windows and was just looking across [15:16]
<gr0ogle> Certainly made a big hole in the sky.
<fwaggle> freaky :O
<gr0ogle> Ah, did you see it?
<fenix> :-)
<fenix> yep
<gr0ogle> How close to us or to you do you reckon? [15:17]
<gr0ogle> It's that red blob on the inner circle at 270°.
<fenix> not sure, but the thunderclap was almost instantanous
<gr0ogle> Then maybe closer to you.
<fenix> Let's just say my first thought was if it hit @ your place
<fenix> naw, it came down well behind the trees in your direction [15:18]
<fenix> I just hope we won't get hail now
<gr0ogle> We just did. [15:19]
* gr0ogle goes outside to count the animals.
<fenix> was it that bad??
<gr0ogle> No obvious damage. [15:21]
<gr0ogle> 2.8 mm rain in about one minute. [15:22]
fenix is Chris, of course.
It seems, though, that we got away lightly if this report is anything to go by. But we did get a marked drop in the temperature,
in itself welcome:
While bumbling around in a maze of twisty little menus, all different, on braindeath,
Chris Yeardley's Microsoft XP machine, found a selection “Is my copy of Windows[sic]
pirated?” or some such—I can't find it again to check. For the fun of it, selected it and
got this page: “Server Error in
'/howtotell' Application.”. The accompanying text relates to remote access, but I was
accessing it locally. I wonder if it ever works.
That wasn't the only problem. At some point I got this:
Spent a lot of time playing with DxO Optics "Pro".
Gradually I'm getting used to the idea of waiting 5 or 10 seconds to see if my mouse click
had any effect or not. Why can't they give some visual feedback?
Didn't really come up with any real new discoveries. Time to RTFM, all 141 pages of it.
Our daughter Yana came home for a few days in the evening. She
had all her hair shaved off in March, and it's still growing back, blonde for some reason.
There's nothing like having guests coming to make you tidy things up. We still had a lot of
junk on the verandah, including the remains of the Phormium “Jester”, so
set to planting that. Split the plant into three parts, two of which we planted:
Things always fail at the most inappropriate time. Round midday I started work on the
turkey stuffing, and couldn't find the
parsley. Tried to call Yvonne, who was at the Yeardley's,
and the phone started ringing immediately. Normally there's a one or two second delay while
the VoIP network does its thing. Despite
other failings, Telstra provides an
apparently immediate connection. So clearly I was being connected via Telstra. Why?
Closer investigation showed a very hot, non-responsive
ATA. Power cycling
didn't help: it was dead. Not the end of the world: I have two of them from the days when I
worked (homephone.lemis.com and officephone.lemis.com). officephone
was dead so I just put homephone in its place. All done bar the configuration.
But the configuration was the fun. The last time I used homephone was in the bad old
satellite days, and it was set up to connect to some long since departed network. Even the
IP address was incorrect (hint, courtesy of Internode:
fire up the machine, connect a phone, hit **** and you should get
a Dalek talking at you.
Hit 110# and it will read you the IP address).
But then came the setup for MyNetFone. I
had my personal details (user number and password), but
what SIP server?
What else might I need to change? Spent 20 minutes searching the MyNetFone site for the
details, downloading all sorts of documents aimed at new users. Nothing to tell me the name
of the server. Finally found a sample config that was really quite well done, and I was able to set the thing up in
a couple of minutes without making a mistake. But why is this information so difficult to
find?
Nele Koemle and her mother Magda Delva along for dinner this evening. All went well, and in
contrast to last year it was warm enough to sit outside on the verandah:
Yana left
for Adelaide
via Melbourne early this morning, and
we're alone again. It's nice having visitors, but I find myself getting more less and less
sociable, and it's fun alone.
Yana has training as a photographer, but for some reason she
doesn't use her Canon 30D much any more; instead, she has a Canon IXUS 130, a close relative of Yvonne's IXY 200F, also known as IXUS 105.
Even when we bought the latter camera, it was clear that the main reason for that choice was
the small size and the image stabilization. Other qualities came a way behind.
And, indeed, that became very apparent when processing yesterday's photos. A large number
of the photos were uselessly out of focus, like these two:
This is the original out-of-camera image without any postprocessing.
There's significant flare and chromatic aberration at the corners. There's no obvious
reason for that. Are the lenses really that bad? As far as I can tell, the lens is the
same as the one in Yvonne's camera, so went and took roughly comparable photos with her
camera and my own, not helped by difficulties deciding on the point from which Yana took her
photo. Here the results and details. First, the images, which don't line up at all well.
The first is with the IXUS 120 (or IXY 220), the second with the IXY 200F (or IXUS 105), and
the third with the Olympus E-30. In all these sequences, mouseover gives a comparison with the next in
line:
This suggests to me that there's something wrong with Yana's camera. Yvonne's isn't
spectacularly good, but it's a lot better than Yana's. And the E-30 is much better than
either.
All these cameras have roughly the same sensor resolution, and the aperture and focal length
were the same for both Canons (5 mm, f/2.8). Possibly a smaller aperture would have
improved the image quality, but the wide aperture seems intentional, since the sensor was
set at 18°/80 ISO, and the shutter speeds were 1/500s and 1/400s.
OK, a DSLR is a different class of camera. But it took me quite a while to realize that,
and I still don't know why Yana is happy with the quality of her images. Things aren't as
obvious as they would seem.
It's only been just over 6 weeks since we first put water in the pond. At that time I also added a couple of pinches
of duckweed. It's happy. Here the pond
seen from the verandah on 14 November 2011 and today:
It meant that we could remove the shade we put in for the fishes, but also that we have an
incredible quantity of duckweed to harvest. I've put it on the garden beds; it should be
high in nitrogen and phosphorus. I can see us doing that every couple of days from now on.
When we planted the micropond nearly 3 years
ago, one of the plants we had great hopes for was
the water lily. It bloomed once
two years ago and then not again. We had
decided that it wasn't getting enough sun, and it seems we were right: since planting, the
number of leaves has increased from 6 to 14 (including losing some of the old leaves), and
we already have the first flower:
It's been years since I realized that one of my biggest problems with my “house photos” was the limited dynamic range. While
looking through my diary recently I found this
entry with an example of the problem.
Fortunately, I have the raw images for this panorama. How does DxO Optics "Pro" handle
the issue? Well, it seems. Here's what I produced 3 years ago and a reprocessing today:
The eastern area of the garden, in front of the verandah, is the oldest part of the garden
that we planted. I later separated the northern part with a row
of Buddleja globosa, which
proved to be an excellent decision. To the north of that we had originally planted daisy
bushes and Tropaeolum, neither of
which—surprisingly—did well. Then I planted
some Gazanias, which did do well, but
they're surprisingly bad at keeping grass down.
I've had a couple of attempts to remove the grass, but it's hard to tell. In addition, we
had planned to put some shrubs to replace the dead daisy bushes. It's an uphill battle, and
I've finally decided to remove everything, kill off the weeds, and replant. The result is
more spectacular than this before/after sequence shows. The gazanias don't seem to be in
bloom very often when I take my house photos, so it's not clear that the entire periphery
was bright yellow—when the sun is shining brightly, not just because of the grass:
Sprayed. Now wait. How long? I want to mulch before I replant, and so I need to wait at
least until the weeds are dying back. In this warm weather it could be only a few days.
Suddenly, managing a PC starts to lot a lot like administering a UNIX box: “It's easy!
Just click here, then you have to turn off the printer to use the network (select here,
pull down this menu, and click on “Disable” and “Apply”), then pull down this menu, then
select the selector, type in your hostname here, then click here, here and double-click
here (dismiss that dialog box, it always gives that, I don't know why...), then pop up
here, select that menu, enable the network, then go over there to start up the TCP/IP
application, then—Whoops! We forgot to set the network mask; no problem, just go back to
the third menu selection and change the mask. Drat, that disabled the network, just fix
that (click, drag, click)... Great, now start up the TCP/IP application again (click), and
now you can use telnet! See, easy!”
Somehow it sums up all the pain of using overly complicated menu systems such as Microsoft
and Apple produce.
I was recently somewhat horrified to read a blog entry by Eric Allman
describing his and Kirk's computer environment. Yes, they both still use FreeBSD, up to a point. But they do most of their work
on Apples, and Kirk uses PC-BSD, which has a
Microsoft-like GUI, rather than straight FreeBSD.
What? The quotation above was signed by Eric and Kirk. Have they abandoned their
principles? I wrote them a message, and got a reply from Kirk that gave me cause to think.
The big issue is ease of installation and updating, which is the reason for PC-BSD rather
than FreeBSD. He doesn't use much of the graphical interfaces.
When it comes to installing and maintaining FreeBSD, I
wrote the book. So clearly I shouldn't have any difficulty, right? But then here I
am borrowing a Microsoft box from Chris Yeardley to run my Microsoft-based photography
software. I could run it on my FreeBSD machine with VirtualBox—if enough memory were available. It
is, too, under FreeBSD-amd64 (the 64 bit version), but I'm still running the 32 bit i386
version. I've been planning to upgrade for months now, but somehow it's just that little
bit too hard.
What's the FreeBSD community doing about it? Not much, and, it seems, not in the right
direction. There's currently a discussion going on on the freebsd-current mailing list about removing the sysinstall program from the upcoming version of
FreeBSD. What replacement? None.
Apple and Microsoft both have good update systems for their own software. I have my doubts
about the ease of updating third-party applications, but clearly there's enough incentive
even for people who don't like the eye candy. And maybe it's a limitation of free software
projects that people tend to develop for themselves, and for them it's not the problem that
it is even for experienced users of the software.
I didn't have much to do today, the wind was still and the weather was sunny—ideal weather
for some more weed spraying, in particular to the extreme east of the garden, where the
weeds have taken over. After yesterday's “success” with
the gazanias, it sounded like a good
idea. But looking at it more closely, it looks as if it's going to be quite a lot of work,
and the weather was warm.
So instead did very little. A number of plants are looking the worse for the heat, and
investigation showed that those that had drippers needed them to be cleaned. Did that, and
added drippers for some in the Japanese
garden—the Asphodeline liburnica and
some plant that I can no longer identify, but which I think is not a weed. Closer
examination showed that it had a large number of insects, mainly some kind of beetle, so
presumably water isn't the only issue.
Today I changed the way I take the house photos to take advantage of the increased dynamic
range of DxO Optics
"Pro". The big difference is that, for most images, anyway, I no longer need to
create a blended HDR image from two or three camera images. That means no ghosting in the
leaves. Here an excerpt from two weeks ago and today:
As I saw last week, that doesn't always work, and so I hedged my bets. In particular with
the verandah panorama I took a number of different views, not helped by the flash
misbehaving on one of the photos. Here the flash component of three consecutive images:
That was enough to confuse Hugin,
and I had to set the control points manually. Somehow the image didn't come out too badly,
and it's still better than the version without flash (2nd image):
Another issue is: should I help HDR by allowing the camera to select the exposure for each
component image? In the past, the answer has been a clear no, but if the HDR can help,
maybe I should reconsider the exposure. And sometimes it seems to work, like these photos
from the north-west corner of the house. The first one is with manual exposure, the second
with automatic:
That looks good enough at first sight. In particular, the foreground in the middle and the
shadows on the right look better. But there's a tell-tale issue in the sky, which is too
bright at the right. That effect is clearer looking at this pair (conventional composite
HDR, then DxO with automatically exposed images):
The DxO version shows a light patch above the shade cloth. This was more pronounced before
enblend got at it. But the HDR
version has its issues too, notably the halos round the tree line, and with some playing
around I was able to improve things:
But what would it have looked like with fixed exposure for all images? Last week it looked
worse. There's still room for playing around here.
And the ghosting? The lack of ghosting should make it easier to find control points and to
get a better fit. And it does. But today I tried it despite more wind than I would
normally have taken, and the result was a number of cases where I had to help Hugin
find its control points. But it worked.
The warm weather continues—today we had a high of 40.0°—and the rain we have had last year
has so far been missing. The effects are very clear in these photos, coincidentally taken
on the first and last day of the year:
We have a fair number of echidnas around
the house, but they're very difficult to photograph: they can disappear very quickly. Last
week I got one, but you have to look hard. The echidna is facing me, with the beak at the
bottom of the photos.
After dinner I was looking outside into the garden and saw an echidna sitting quietly just
outside. The pointed beak was unmistakable, so I went and got my camera, mounted the long
telephoto lens, and returned. Still there:
And no, I hadn't had that much to drink. Chris Yeardley, there for dinner, confirmed that
the light made it look like an echidna. But it's high time my new glasses are ready.
Do you have a comment about something I have written? This is a diary, not a
“blog”, and there is
deliberately no provision for directly adding comments. It's also not a vehicle
for third-party content. But I welcome feedback and try to
reply to all messages I receive. See the diary overview for more details. If you do
send me a message relating to something I have written, please indicate
whether you'd prefer me not to mention your name. Otherwise
I'll assume that it's OK to do so.
