Jump to content

The Leica Forum uses cookies. Read the privacy statement for more info. To remove this message, please click the button to the right:    OK, understood.

Photo
- - - - -

Leica M (240) Image Unique ID


  • Please log in to reply
42 replies to this topic

#1 biogon

biogon

    Neuer Benutzer

  • Members
  • Pip
  • 7 posts

Posted 08 March 2013 - 15:56

Advertisement (Gone after free registration)
Hello everybody!

Apparently Leica in the M has changed the way of numbering shots in the "Image Unique ID" field of the exif data, the M8/M9 Hexadecimal numbering is no longer being used. Now we have there what appears to be a random decimal number, not correlated with the number of shots taken by that camera.

Does anyone know whether that number is really random (not completely because it should anyway remain unique, I surmise), or a hidden code exists allowing to tell how many shots the camera has taken?

Thank you and all the best.

Edited by biogon, 08 March 2013 - 16:26.


#2 BerndReini

BerndReini

    Erfahrener Benutzer

  • Members
  • 1,351 posts

Posted 09 March 2013 - 22:47

I don't really know the answer to your question as far as shutter actuations, but you could actually do the same thing with the M9 with a little workaround: you just had to renumber a picture on the SD card in your computer, then continue shooting with the M9 and it would follow in consecutive numbers for file naming after this last renumbered picture.

#3 thighslapper

thighslapper

    Grumpy Old Fart

  • Premium Member
  • 3,825 posts
  • LocationUK - Shakespeares County

Posted 10 March 2013 - 17:13

Biogon......

You are quite correct ..... and the behaviour is not quite random.....

I seems to reset every time the camera is switched off (or possibly in standby too)..... and then produces unique id's with rising numbers but not in any sensible sequence starting at some random point ranging from the low thousands to several millions......

eg... (pic then unique id)

170 2274331
171 2314332
172 2388281
173 2435851
174 2537540
175 2601900 etc

next day:

207 95847
208 219539
209 404396
210 564961 etc

today:

233 6280
234 119385

camera off

235 37013
236 67552
237 143641
238 226094 etc.....

there is another number in the exif - unique raw id ......

but that has a basic form of 30303030303030x3x3x3x3x3x.....

where the x bits are in fact the image unique id (ie all the alternate 3's stripped out)

...... so you are correct .... without a degree in cryptography it looks impossible to tell how many actuations a camera has had........

.... and there is some sense in that ...... I returned my 'as new' ex demo M9 when I discovered it had in fact over 5000 actuations recorded .... which I argued made it 'second hand' ....... and got a new M9 for my trouble......

very interesting......:confused:

#4 biogon

biogon

    Neuer Benutzer

  • Members
  • Pip
  • 7 posts

Posted 10 March 2013 - 17:31

very interesting......:confused:


Yes, indeed...

Thank you so much thighslapper! Your message is a very good starting point toward understanding.

#5 pico

pico

    Sponsoring Member

  • Members
  • 10,110 posts
  • Location51°39'16.05" N 5°42'14.47" E

Posted 10 March 2013 - 18:03

Oh, thighslapper, now you have done it - teased a decoder. :)

The numbers past the sequence # cannot be pseudo-random for a number of reasons besides the fact that they are tied to the system hardware. For one, they increase with each new sequence # from power-on. Someone with more time and brains that I have will figure it out.

Edited by pico, 10 March 2013 - 18:06.


#6 thighslapper

thighslapper

    Grumpy Old Fart

  • Premium Member
  • 3,825 posts
  • LocationUK - Shakespeares County

Posted 10 March 2013 - 22:44

A BIG clue........

I fiddled about with this camera when I got it and reset the numbering to be 400Leica as the start folder and numbering to start from 4000001.... (so the numbers are unique and don't double up on previous M's)

This file has NO EXIF unique RAW id

The unique file id however is 00000000000000000000000000002cd1

no other alphanumeric coding features in any subsequent exif file id info.....

this one presumably seeds the whole of the subsequent numbering process....
  • pico, biogon and mabbo said thank you to this

#7 biogon

biogon

    Neuer Benutzer

  • Members
  • Pip
  • 7 posts

Posted 11 March 2013 - 08:53

2cd1 hexadecimal in a M8/M9 body would correspond to 11473 shots. Is that number a reasonable estimate of the pictures you have already taken with your M 240 thighslapper?

What happens if you restore the default folder/numbering? The camera would resume generating semi-random decimal IDs, or will retain the hexadecimal numbering system?

#8 wlaidlaw

wlaidlaw

    Erfahrener Benutzer

  • Premium Member
  • 9,814 posts
  • City / Ort:Tourtour, France and Sussex UK

Posted 11 March 2013 - 09:41

Could a date be being incorporated into the unique ID or alternatively, could this be linked to the GPS function, which will become available when the multifunction grips start to arrive? I was thinking that the unique ID = shutter actuation number + hexadecimal date (with start point ???)

Wilson

#9 thighslapper

thighslapper

    Grumpy Old Fart

  • Premium Member
  • 3,825 posts
  • LocationUK - Shakespeares County

Posted 11 March 2013 - 10:22

apologies folks ....... 2cd1 is picture no. 1 when I reset everything and set up my own folder with my own numbering. It was an M9 image I copied to the card and renumbered to seed the numbering sequence ......... and in fact it did this as per the behaviour of the M9

The original date on the first few pictures was 2008 (I forgot to set it) .... and I again I reset the time a day later (I was 12hrs adrift as I hadn't set it to 24hr clock) and it seems to make no difference to the pattern of numbering.....

I am not going to restore the default numbering to investigate further ........ there are a number of glitches in the way the system works ..... the main one being it works independently for each camera profile .... and when I did this before it resulted in me having to re-enter everything for each profile and then take a load of blank pictures to bring me up to my original numbering ..... you would have to fiddle with this camera to understand how this works .... but basically the file folder and file numbering is not a global setting....

We need someone who is a mathematician to explain how you can set up an apparently random numbering system that will always produce unique numbers ....... date and time seems a bit flaky as you could reset the cameras time/date ..... unless it registers elapsed time since 'birth'.
Presumably there is some loook up table that provides a multplier or offset that generates the final number........

Although the numbers seem to rise when in continuous use the jumps are pretty random .......

I will post some screenshots of the full exif data later and see if someone spots something.....

Edited by thighslapper, 11 March 2013 - 10:53.

  • wlaidlaw and biogon said thank you to this

#10 thighslapper

thighslapper

    Grumpy Old Fart

  • Premium Member
  • 3,825 posts
  • LocationUK - Shakespeares County

Posted 11 March 2013 - 15:07

exif data :

Attached Files



#11 biogon

biogon

    Neuer Benutzer

  • Members
  • Pip
  • 7 posts

Posted 11 March 2013 - 17:44

My question arose from looking at the exif data of some of the dng pictures posted by Ming Thein (Flickr: mingthein's Photostream), I do not own a M 240 (yet! ;)). In case they would help, I am listing their image numbers and corresponding Image Unique IDs below:

10 79393
22 754035
26 270199
27 344698
37 106609
41 123032
58 1014633
79 298026
85 313374
88 4718
95 517834
108 617410
111 194886
217 488822
249 424922
253 94525
278 523946
280 108239
282 45770
295 428897
305 276703
325 245075
334 179411
346 185423
362 421215
364 543819
373 260589
422 335311
429 316729
437 423110
438 491976
444 879230
457 531080
498 1845810
548 318712
563 319833
584 76126
733 117764
1273 114473
1336 506172
1348 761738
1362 869199
1364 90199
2115 78763
2230 164980
2250 154542
2423 458802
2505 455409
2532 74702
2887 926509
3101 259258
3108 218801
3143 2154800
3155 59695
3458 2564511
3492 4822417
3517 6475792
3532 90915
3539 559744
3641 2029507
3792 396759
3828 811918
4436 182029
4443 643539
4497 638831
4609 69941
4662 343922
4666 128843
4678 858528

#12 wlaidlaw

wlaidlaw

    Erfahrener Benutzer

  • Premium Member
  • 9,814 posts
  • City / Ort:Tourtour, France and Sussex UK

Posted 11 March 2013 - 17:58

My question arose from looking at the exif data of some of the dng pictures posted by Ming Thein (Flickr: mingthein's Photostream), I do not own a M 240 (yet! ;)). In case they would help, I am listing their image numbers and corresponding Image Unique IDs below:

10 79393
22 754035
26 270199
27 344698
37 106609
41 123032
58 1014633
79 298026
85 313374
88 4718
95 517834
108 617410
111 194886
217 488822
249 424922
253 94525
278 523946
280 108239
282 45770
295 428897
305 276703
325 245075
334 179411
346 185423
362 421215
364 543819
373 260589
422 335311
429 316729
437 423110
438 491976
444 879230
457 531080
498 1845810
548 318712
563 319833
584 76126
733 117764
1273 114473
1336 506172
1348 761738
1362 869199
1364 90199
2115 78763
2230 164980
2250 154542
2423 458802
2505 455409
2532 74702
2887 926509
3101 259258
3108 218801
3143 2154800
3155 59695
3458 2564511
3492 4822417
3517 6475792
3532 90915
3539 559744
3641 2029507
3792 396759
3828 811918
4436 182029
4443 643539
4497 638831
4609 69941
4662 343922
4666 128843
4678 858528


Where is Alan Turing when you need him?

Wilson
  • hoppyman, Fgcm, Sergius and 1 other said thank you

#13 pico

pico

    Sponsoring Member

  • Members
  • 10,110 posts
  • Location51°39'16.05" N 5°42'14.47" E

Posted 11 March 2013 - 18:12

Could a date be being incorporated into the unique ID or alternatively, could this be linked to the GPS function, which will become available when the multifunction grips start to arrive? I was thinking that the unique ID = shutter actuation number + hexadecimal date (with start point ???)

Wilson


It seems unlikely that there's a date in there. I'm quite familiar with a few date coding methods, for example those that store a date as seconds since Jan 1, (whatever year). It is not HEX, either.

thighslapper: We need someone who is a mathematician to explain how you can set up an apparently random numbering system that will always produce unique numbers


That's 'random without replacement' and it can be done, and part of the sequence will reveal the technique.

....... date and time seems a bit flaky as you could reset the cameras time/date ..... unless it registers elapsed time since 'birth'.


hundredths of a second (or clock ticks) since manufacture is a cool idea, however some of the numbers are too small to be date/time.

Biogon - thanks for going through the trouble of gathering the new numbers, but it would help a great deal if they were perfectly sequential.

Edited by pico, 11 March 2013 - 18:18.


#14 biogon

biogon

    Neuer Benutzer

  • Members
  • Pip
  • 7 posts

Posted 11 March 2013 - 18:48

Biogon - thanks for going through the trouble of gathering the new numbers, but it would help a great deal if they were perfectly sequential.

I agree. But as I said I do not own a M 240, and those numbers are all I could get a few days ago from the Ming Thein's site.

Any other M 240 owner willing to contribute to the M240 cryptanalysis?
  • pico said thank you to this

#15 wlaidlaw

wlaidlaw

    Erfahrener Benutzer

  • Premium Member
  • 9,814 posts
  • City / Ort:Tourtour, France and Sussex UK

Posted 11 March 2013 - 19:30

I agree. But as I said I do not own a M 240, and those numbers are all I could get a few days ago from the Ming Thein's site.

Any other M 240 owner willing to contribute to the M240 cryptanalysis?


What do you want?

Wilson

#16 thighslapper

thighslapper

    Grumpy Old Fart

  • Premium Member
  • 3,825 posts
  • LocationUK - Shakespeares County

Posted 11 March 2013 - 20:03

Perhaps someone could ask the nice Leica people in Solms .........

There seems to be a determined effort at obscurity here ..... so I doubt they will reveal all.... but they might divulge the rationale behind why they have done this..... :rolleyes:

#17 biogon

biogon

    Neuer Benutzer

  • Members
  • Pip
  • 7 posts

Posted 11 March 2013 - 20:18

What do you want?

What Pico is saying he needs: a pretty long list of perfectly sequential image numbers from M 240 with corresponding Image Unique IDs...

#18 biogon

biogon

    Neuer Benutzer

  • Members
  • Pip
  • 7 posts

Posted 16 July 2013 - 06:22

Biogon - thanks for going through the trouble of gathering the new numbers, but it would help a great deal if they were perfectly sequential.


Hello, I resurrect this thread because I finally found a list of sequential Image Unique ID from M Typ 240 pictures (only No. 5 is missing because it corresponds to a video file).

https://dl.dropboxus.../M240_IUIDs.txt

I look forward that they may help in deciphering the mystery...
  • pico and dwbell said thank you to this

#19 jperkins

jperkins

    Benutzer

  • Members
  • PipPipPip
  • 95 posts
  • City / Ort:South Cambridgeshire, UK.

Posted 22 August 2014 - 00:15

and the behaviour is not quite random.....

I seems to reset every time the camera is switched off (or possibly in standby too)..... and then produces unique id's with rising numbers but not in any sensible sequence starting at some random point ranging from the low thousands to several millions......


I think you're correct. Sampling some of my images I think the number is time from camera switch on, incremented at a rate around 1500/sec. I haven't ascertained yet if the count continues whilst it sleeps, but switching off seems to reset it.

I know its nice to speculate that there's a clever scheme here to encode the shutter count, but come on - we're talking Leica software engineering. All they need for the ImageUniqueID exif tag is a number with a reasonable chance of not being duplicated - why not just sample an internal processor tick count at some point in the image taking process? They haven't even bothered to store the number in non-volatile memory to carry on when next switched on, which given the size of the field would have pretty much guaranteed uniqueness.

I'm sorry but Leica doesn't have a track record in software excellence, so my money's on the under-engineered, get away with the minimum possible solution ;)

Jonathan

#20 CheshireCat

CheshireCat

    Frabjous Benutzer

  • Members
  • 1,538 posts
  • LocationWonderland Sunny California

Posted 22 August 2014 - 01:38

I'm sorry but Leica doesn't have a track record in software excellence, so my money's on the under-engineered, get away with the minimum possible solution ;)


+1

A proper algorithm would use:

- Camera serial number (unique globally - all Leica cameras).
- Shutter count (unique locally - your camera).

And given the fact I see in this thread a "663" number reported by biogon, it is extremely unlikely that Leica "engineers" are doing anything like that.

Edited by CheshireCat, 22 August 2014 - 01:45.

  • jperkins said thank you to this




0 user(s) are reading this topic