Greg
animals brewing
food and drink gardening
general health
history language
music multimedia
opinion photography
politics Stones Road house
technology
Greg's diary
recent entries
Translate this page
Select day in August 2017:
Su Mo Tu We Th Fr Sa
1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31
Select month:
2016 Sep Oct Nov Dec
2017 Jan Feb Mar Apr
2017 May Jun Jul Aug
Today's diary entry
Diary index
About this diary
Greg's home page
Greg's photos
Network link stats
Greg's other links
Copyright information
    
Groogle

Sunday, 20 August 2017 Dereel Images for 20 August 2017
Top of page
next day
last day

Understanding GPS locations
Topic: technology, opinion Link here

As of yesterday, I wasn't much closer to establishing the cause of my GPS location inaccuracies. One photo was interesting. Here the complete photo, then crops:


This should be GPS-log-1-4.jpeg.  Is it missing?
Image title: GPS log 1 4          Dimensions:          3257 x 3674, 2387 kB Display location on map
Make a single page with this image Hide this image
Make this image a thumbnail Make thumbnails of all images on this page
Make this image small again Display small version of all images on this page
All images taken on Saturday, 19 August 2017, thumbnails          All images taken on Saturday, 19 August 2017, small
Diary entry for Saturday, 19 August 2017 Complete exposure details

   
This should be GPS-log-1-3.jpeg.  Is it missing?
Image title: GPS log 1 3
Display location on map
Complete exposure details
Dimensions: 411 x 656, 177 kB
Dimensions of original: 786 x 1255, 309 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Saturday, 19 August 2017:
thumbnails    small images    diary entry
 
This should be GPS-log-1-2.jpeg.  Is it missing?
Image title: GPS log 1 2
Display location on map
Complete exposure details
Dimensions: 414 x 651, 141 kB
Dimensions of original: 1904 x 2993, 878 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Saturday, 19 August 2017:
thumbnails    small images    diary entry

That was taken with my Olympus OM-D E-M1 Mark II, which normally records local time. But exiftool's -geotag function adds information. Here the tags, bold from exiftool:

Modify Date                     : 2017:08:19 14:38:59
Date/Time Original              : 2017:08:19 14:38:59
Create Date                     : 2017:08:19 14:38:59
Date Time UTC                   : 2017:08:19 04:38:59
GPS Date Stamp                  : 2017:08:19
GPS Date/Time                   : 2017:08:19 04:38:59Z

Looking at the photo, though, it seems that the two GPS receivers disagree. The difference in longitude (0.018') represents a distance of 2.6 m, about 10 times the width of the two devices.

But the tablet is showing 3 seconds in the future, 14:39:02 instead of 14:38:59! How did that happen? The time display at top left still shows 14:38, so this must be some bug. But where? The time in my camera's Exif data comes from two different sources: the first is the camera clock, synchronized at some time in the last few days, and the second is from the GPS log, which came from the phone. Both agree to the second. Everything points to an issue with the tablet.

Looking at the logs for the time shows something else. After trimming, here are the last three trackpoints:

<trkpt lat="-37.80020538" lon="143.75118283">
  <ele>347.79998779296875</ele>
  <time>2017-08-19T04:37:00.998Z</time>

<trkpt lat="-37.80019051" lon="143.75098896">
  <time>2017-08-19T04:38:02.000Z</time>

<trkpt lat="-37.80019512" lon="143.75101123">
  <time>2017-08-19T04:39:03.000Z</time>

In passing, it's interesting to look at the elevation that I left in the first entry: 347.79998779296875 m. That's not far off the mark, but what a ridiculous precision: to the nearest 10 femtometres, less than the diameter of the nucleus of a gold atom! The general accuracy of GPS altitude measurements is in the order of 10 m. Where do people get these ideas?

More important, though, are the readings themselves. They're 60 seconds apart! What about the phone?

<trkpt lat="-37.8002935" lon="143.7513125">
  <time>2017-08-19T04:35:29.007Z</time>

<trkpt lat="-37.800191645510495" lon="143.751177219674">
  <time>2017-08-19T04:36:35.000Z

<trkpt lat="-37.80014030635357" lon="143.75106004066765">
  <time>2017-08-19T04:37:36.000Z

<trkpt lat="-37.80014030635357" lon="143.75106004066765">
  <time>2017-08-19T04:38:37.000Z</time>

<trkpt lat="-37.8002935" lon="143.7513124">
  <time>2017-08-19T04:39:36.358Z</time>

Strangely, though the app is the same, the output format is different. In this case, the obvious difference is the precision of the coordinates: instead of “only” 9 digits after the decimal point (a precision in the order of 1 mm), it's 14 digits (a precision in the order of 10 nm) in most cases. I've already voiced an opinion on the precision, but why is it different between two devices running the same software? And why does the first set only have 7 digits? The result is surprising. Here the first two trackpoints in more detail:

<trkpt lat="-37.8002942" lon="143.7513122">
  <time>2017-08-19T04:31:05.390Z</time>
  <src>network</src>
</trkpt>
<trkpt lat="-37.80015740543604" lon="143.7512298580259">
  <ele>351.79998779296875</ele>
  <time>2017-08-19T04:32:11.000Z</time>
  <course>48.0</course>
  <speed>0.8757</speed>
  <geoidheight>-6.4</geoidheight>
  <src>gps</src>
  <sat>6</sat>
  <hdop>1.6</hdop>
  <vdop>2.3</vdop>
  <pdop>2.9</pdop>
</trkpt>

The first one states “<src>network</src>”. What network? How can my network give locations accurate to 10 mm? There's a lot more to understand here.

In any case, here too we have readings only once a minute. No wonder things are inaccurate. Looking at the app itself, I found:

 
This should be Screenshot_2017-08-20-10-24-52.png.  Is it missing?
Image title: Screenshot_2017 08 20 10 24 52
Complete exposure details
Dimensions: 402 x 670, 136 kB
Dimensions of original: 480 x 800, 77 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Sunday, 20 August 2017:
thumbnails    small images    diary entry

This should be Screenshot_2017-08-20-10-24-52-detail-1.png.  Is it missing?
Image title: Screenshot_2017 08 20 10 24 52 detail 1          Dimensions:          914 x 295, 45 kB
Make a single page with this image Hide this image
Make this image a thumbnail Make thumbnails of all images on this page
Display small version of all images on this page
All images taken on Sunday, 20 August 2017, thumbnails          All images taken on Sunday, 20 August 2017, small
Diary entry for Sunday, 20 August 2017 Complete exposure details

 

Not surprisingly, the interval was set to one minute. Next time I'll try it with the interval set to 0, though I take the warning seriously: even under current circumstances the GPS receiver drastically shortens battery life.

And then there's the second setting:

 
This should be Screenshot_2017-08-20-10-24-52-detail-2.png.  Is it missing?
Image title: Screenshot_2017 08 20 10 24 52 detail 2
Complete exposure details
Dimensions: 658 x 411, 39 kB
Dimensions of original: 415 x 259, 18 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Sunday, 20 August 2017:
thumbnails    small images    diary entry

What does that mean? Again something to try out.


Video and focus tracking
Topic: photography, opinion Link here

Yesterday evening Chris Bahlo and Yvonne asked me to take some video of them on horseback today. Not a thing that I'm very experienced in, but how hard can it be? Put the camera on a tripod with a ballhead, point, shoot.

The horses were coming to me, requiring constant focus correction. That's one of the things at which the Olympus OM-D E-M1 Mark II is supposed to be very good. But somehow it just didn't work for me. I had this problem before, though I chose not to show the videos: the “tracking” of continuous autofocus got lost and went off looking at unrelated things. So today I tried continuous autofocus without tracking.

The results were underwhelming. In every single case I lost focus somewhere in the middle, like here:

Why? Is it maybe because I have only one focus point active? That works well for stills (for me, anyway), but maybe it's counterproductive for video.

The other issue is the tripod. It really doesn't help much. I think I should try hand-held next time. One way or another, I'm going to need some image stabilization software, of which much is available.


FORTRAN history
Topic: technology, history, opinion Link here

There was a discussion recently on the Unix Heritage Society which diverged towards FORTRAN, discussing why it was such a horrible language. I had input on that:

I think the arithmetic IF was put into FORTRAN because it was easy to implement with the CAS instruction. It doesn't make much sense from a mathematical point of view.

... my guess is that the authors of FORTRAN were looking for the cheapest solution, not the fastest one. For decades to come, the fastest solution was assembler.

More input, including document pointers from Paul McJones, who maintains an interesting list of historical documents on FORTRAN. One was a number of documents related to John Backus' paper on “The history of FORTRAN I, II and III” from 1978. A very interesting read, and it completely disproved my claim that they weren't too interested in performance. It was the other way round: they wanted to write an optimizing compiler, and the language definition was secondary. I suppose that explains a lot about FORTRAN.

And arithmetic IF? Another document was the Specifications for the IBM Mathematical FORmula TRANSlating System of November 1954. It's interesting because the language really didn't change much between then and the first release some time 2½ years later. One thing that did change, though, was the arithmetic IF: it originally only had two labels.

And the use of the CAS instruction? It stands for “Compare Accumulator and Store” (“Store” a noun meaning “storage”). But that's not the semantics of IF: first you evaluate an expression, and presumably the result is in the accumulator. Then, depending on that result, you decide where to go. No need for “store”, and the IBM 704 had plenty of conditional branch (“transfer”) instructions. So basically both of my claims are wrong. My best guess now is that CAS may have been a godfather to arithmetic IF, but almost certainly not a parent.


Monday, 21 August 2017 Dereel → Ballarat → Dereel Images for 21 August 2017
Top of page
previous day

Fruitful mandarins
Topic: food and drink, gardening Link here

Yvonne bought some Mandarin oranges recently, of a kind that is no longer common. Here's what was left from only two of them:

 
This should be Mandarin-pips.jpeg.  Is it missing?
Image title: Mandarin pips
Complete exposure details
Dimensions: 526 x 514, 235 kB
Dimensions of original: 3873 x 3783, 3898 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Monday, 21 August 2017:
thumbnails    small images    diary entry

It brings home how far plant propagation has come: pipless fruit can't exist in nature.


Furniture and digital watches
Topic: general, opinion Link here

We've been looking for replacements for our aged lounge-room suite for years now, and we still haven't found anything we both like. Recently Yvonne found something in a Harvey Norman catalogue that she liked: a double recliner armchair joined together with a “table” area in the middle, and very wide. Off today to take a look. Yvonne likes it. I find it comfortable, but not what we're looking for. In particular, the width doesn't really add anything. And there's no other matching furniture.

While in town, also looked for a new battery for Yvonne's watch. $10, and afterwards it looked as if the battery was on its last legs: the display showed considerable shadows in the segments that should have been “off”. Back to the bookshop where she had it replaced, and was told that the watch had passed its “use-by” date. OK, it's 15 years old, but it worked normally up to that point, and the mode of failure doesn't appear convincing. But it's not easy to disprove the claim. In the end, Yvonne bought the cheapest watch that she could find, particularly ugly and apparently with only a 12 hour display, as a stop-gap. Now to look for a real watch. But the days of digital watches are long over, and there's almost no choice.


More GPS fun
Topic: technology, photography, opinion Link here

The trip into Ballarat gave us the opportunity to get a more detailed GPS log, this time with log delay set to 0. The result?

<trkpt lat="-37.560112" lon="143.8601387"><time>2017-08-21T05:38:25.142Z</time><src>network</src></trkpt>
<trkpt lat="-37.5601136" lon="143.8601253"><time>2017-08-21T05:38:29.171Z</time><src>network</src></trkpt>
<trkpt lat="-37.5601124" lon="143.8601356"><time>2017-08-21T05:38:33.167Z</time><src>network</src></trkpt>
<trkpt lat="-37.5601019" lon="143.8601346"><time>2017-08-21T05:38:37.087Z</time><src>network</src></trkpt>
<trkpt lat="-37.5601095" lon="143.8601355"><time>2017-08-21T05:38:57.655Z</time><src>network</src></trkpt>
<trkpt lat="-37.5601092" lon="143.8601309"><time>2017-08-21T05:39:01.564Z</time><src>network</src></trkpt>
<trkpt lat="-37.5600957" lon="143.8601412"><time>2017-08-21T05:39:05.469Z</time><src>network</src></trkpt>

Again, src is “network”. What does that mean? The coordingates show that we were outside The Good Guys, which is correct. So what network does it mean? And why is there no speed information in the log file?

The other thing to note is that the log entries are roughly every 4 seconds (with a noticeable gap in the middle). Is this due to “network”, or is this the maximum that the receiver can deliver? Still more puzzling.

Back home, and before I had discovered these details, I went around the house and took some photos with the phone from the places where I do my house photos. The difference in quality isn't surprising:


This should be house-from-entrance.jpeg.  Is it missing?
Image title: house from entrance          Dimensions:          958 x 282, 151 kB Display location on map
Make a single page with this image Hide this image
Make this image a thumbnail Make thumbnails of all images on this page
Display small version of all images on this page
All images taken on Saturday, 19 August 2017, thumbnails          All images taken on Saturday, 19 August 2017, small
Diary entry for Saturday, 19 August 2017 Complete exposure details

   
This should be house-from-entrance.jpeg.  Is it missing?
Image title: house from entrance
Display location on map
Complete exposure details
Dimensions: 600 x 450, 150 kB
Dimensions of original: 3264 x 2448, 1288 kB
Display this image:
thumbnail    hidden   alone on page
Display all images on this page as:
thumbnails    this size
Show for Monday, 21 August 2017:
thumbnails    small images    diary entry

Conveniently, the file names of the photos include date and time. This one was 20170821_175106.jpg. But all the photos had the same location and the same incorrect timestamp (20170821, 17:50:10, up to 5 minutes off the real time. Looking at the date information, I find:

=== grog@eureka (/dev/pts/36) ~/Photos/20170821/orig/JPEG 342 -> exiftool 20170821_175106.jpg | grep -i time
File Modification Date/Time     : 2017:08:21 18:34:28+10:00
File Access Date/Time           : 2017:08:22 10:21:43+10:00
File Inode Change Date/Time     : 2017:08:21 18:46:48+10:00
Exposure Time                   : 1/17
Date/Time Original              : 2017:08:21 17:51:06
GPS Time Stamp                  : 07:50:10
GPS Date/Time                   : 2017:08:21 07:50:10Z

So the first issue is that it seems I'm using the wrong time; “Date/Time Original“ seems to be the one to go by. Why the discrepancy in the GPS position and time stamp? The log shows updates every 4 or so seconds, as before, and exiftool -geotag happily sets the locations.

Somehow this is all a can of worms.


This page contains (roughly) yesterday's and today's entries. I have a horror of reverse chronological documents, so all my diary entries are chronological. This page normally contains the last two days, but if I fall behind it may contain more. You can find older entries in the archive. Note that I often update a diary entry a day or two after I write it.     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. 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.


Greg's home page This month Greg's photos Greg's links

RSS 2.0 Valid XHTML 1.0!