Leica M (240) Image Unique ID

[bephoto]

Advertisement (gone after registration)

Isn't it obvious, its

 

Sn = Serial Number

Sh= Shots

Un = Unique Number

 

((π x Sn) ^ Sh)/√1914^π = Un

[lct]

Is 1914 your birthdate?

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Guest JonathanP]

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.

...

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 wasn't expecting Leica's dodgy code to produce a duplicate ImageUniqueId quite so quickly, but today I load some pictures into Lightroom and one of my images shows a preview of a picture that I took on holiday 2 months ago! Examining the raw file in RawDigger I find that both have a "unique" image ID of 25046. This really confuses Lightroom's cache.

 

Leica really need to sort out their software engineering - this sort of mistake (actually incompetence is a much better word) is not acceptable in a camera at this price level.

 

To create a unique number is a trivial task for this application - I have written algorithms to do this in embedded software more than once in my career. Its not difficult, trust me.

 

Oh well, I guess I'm going to have to write another Lightroom plugin this time to randomise the ImageUniqueId. Has anyone found a reliable way for us non-beta testers to report software bugs to Leica?

 

Jonathan

 

[Lightrooms cache may be being confused by the RawDataUniqueId which Leica have generated using an ASCII encoding of the lower part of the ImageUniqueId. Both need to be actually unique.]

Edited September 6, 2014 by jperkins

[pico]

I wasn't expecting Leica's dodgy code to produce a duplicate ImageUniqueId quite so quickly, but today I load some pictures into Lightroom and one of my images shows a preview of a picture that I took on holiday 2 months ago! Examining the raw file in RawDigger I find that both have a "unique" image ID of 25046. This really confuses Lightroom's cache.

 

Leica really need to sort out their software engineering - this sort of mistake (actually incompetence is a much better word) is not acceptable in a camera at this price level.

 

To create a unique number is a trivial task for this application - I have written algorithms to do this in embedded software more than once in my career. Its not difficult, trust me.

 

 

Been there myself. During one of my software contract jobs I was nicknamed, "date time man." You and I know how trivial it is to time-stamp anything, even (as I had to) work with a sixteen bit word and rapid sequences. (RT-11 DEC 11/70).

 

I fail to understand why Leica wants to obfuscate image IDs. It makes no sense at all.

Edited September 6, 2014 by pico

[CheshireCat]

I fail to understand why Leica wants to obfuscate image IDs. It makes no sense at all.

 

That looks like some dead code no one completed because of some release deadline

 

In any case, a sequential numbering won't work. A unique image ID should be unique "in the world" (i.e. almost 0 chance of duplication). There are a lot of ways to do that.

 

Leica software "engineering" is hopeless. They should open source the firmware as soon as a new camera is released. The release firmware is nothing else than a reasonable beta (let's admit it: flash memory gives companies the excuse to release products with beta software)... then it takes just a few passionate people to make it great - for free.

[wlaidlaw]

As Leica got unique image ID correct on the M8 and M9, you would have thought they could have just ported that bit of code over to the M240. It is also messing up image sorting in Bridge - very sloppy. When I get back from Colorado at the end of this week, I will report the repeating image ID problem to Leica.

 

Wilson

[CheshireCat]

Advertisement (gone after registration)

you would have thought they could have just ported that bit of code over to the M240.

 

10 New boss comes in and says "This code is crap ! Who's the idiot who did it ? Let's redo everything."

20 GOTO 10

[pop]

Goto statement considered harmful.

[CheshireCat]

Goto statement considered harmful.

 

Sorry. Revised code:

 

for ( ; ; ) {

New boss comes in;

New boss says "This code is crap ! Who's the idiot who did it ? Let's redo everything.";

}

[pop]

Sorry. That's much better:

 

for ( ; ; ) {

New boss comes in;

New boss says "This code is crap ! Who's the idiot who did it ? Let's redo everything.";

}

You missed the case of the programmer promoted to boss.

[CheshireCat]

You missed the case of the programmer promoted to boss.

 

Allright:

 

for ( ; ; ) {

New boss comes in;

New boss says "This code is crap ! Who's the idiot who did it ? Let's redo everything.";

if (boss == programmer) add(useless_nerdy_feature);

else remove(useful_feature);

}

[Guest JonathanP]

When I get back from Colorado at the end of this week, I will report the repeating image ID problem to Leica.

 

Thanks Wilson. They need to correct both the ImageUniqueId and the RawDataUniqueId, and remove the ASCII encoding of the latter which is halving the number of effective digits.

 

Jonathan

Edited September 7, 2014 by jperkins

[semi-ambivalent]

I fail to understand why Leica wants to obfuscate image IDs. It makes no sense at all.

 

I think they want to obfuscate shutter wear/age/actuations.

 

s-a

[pico]

Thanks Wilson. They need to correct both the ImageUniqueId and the RawDataUniqueId

 

FWIW, the firmware does not use those strings, but we know what you mean.

 

I'm retired and have no access to what Softune REALOS/FR requires.

[Guest JonathanP]

FWIW, the firmware does not use those strings

 

Oh quite likely but:

 

<pedantic-software-engineer>

ImageUniqueID is the field name from EXIF Specification v2.2

RawDataUniqueID is the field name from DNG Specification v1.4

</pedantic-software-engineer>

 

they can call them fred and jim in the code for all I care*

 

*actually I do care because stupid names = code maintenance nightmare - I'm sure you've had your fair share of inherited mess as well

[CheshireCat]

*actually I do care because stupid names = code maintenance nightmare

 

Worst case I have ever seen was a piece of commercial software, where all the variables where named after p*rn stars and related items.

[k-hawinkler]

As Leica got unique image ID correct on the M8 and M9, you would have thought they could have just ported that bit of code over to the M240. It is also messing up image sorting in Bridge - very sloppy. When I get back from Colorado at the end of this week, I will report the repeating image ID problem to Leica.

 

Wilson

 

 

Hi Wilson,

 

Have you reported this problem to Leica yet?

This is certainly a fascinating subject.

TIA.

[wlaidlaw]

K-H,

 

I got someone to talk to them about it a PK. There was apparently, not a flicker of interest in doing anything about sorting the seemingly random and repeating unique image ID situation. Leica I am afraid, is increasingly beginning to look like a typical company co-run by Blackstone, where all that matters is the short term bottom line and s*d everything else. I would not be at all surprised to hear they have borrowed hugely to pay out a “special dividend,” as that is the way these private equity sharks seem to go about their business.

 

Wilson

[Guest JonathanP]

There was apparently, not a flicker of interest in doing anything about sorting the seemingly random and repeating unique image ID situation.

 

Thanks for trying anyway Wilson!

 

As a workaround, I have released a new version of my M240 green shadows plugin that now sets both the ImageUniqueID and RawDataUniqueID in the output file to a MD5 derived digest of the raw data - this will, to all intents, guarantee uniqueness. The free plugin can be downloaded from here. If anyone does get stuck with 2 files with the same ID like I did, at least it provides a quick and simple workaround.

 

Interestingly I found that Lightroom uses the RawDataUniqueID to index its preview cache, but Bridge uses the ImageUniqueID, so both fields need to be fixed (in case Leica is listening).

 

Hope it helps.

 

Jonathan

[Advertisement]

Hallo,

Look here: Leica M

[wlaidlaw]

Thanks for trying anyway Wilson!

 

As a workaround, I have released a new version of my M240 green shadows plugin that now sets both the ImageUniqueID and RawDataUniqueID in the output file to a MD5 derived digest of the raw data - this will, to all intents, guarantee uniqueness. The free plugin can be downloaded from here. If anyone does get stuck with 2 files with the same ID like I did, at least it provides a quick and simple workaround.

 

Interestingly I found that Lightroom uses the RawDataUniqueID to index its preview cache, but Bridge uses the ImageUniqueID, so both fields need to be fixed (in case Leica is listening).

 

Hope it helps.

 

Jonathan

 

Jonathan,

 

Many thanks for having a go a sorting this. Sadly I very rarely use LR, using C1 for my RAW converter. However I was using Bridge but until Leica sorts this problem, I can’t really use it. I can only think that Leica was trying to prevent people seeing how many shutter activations cameras had had before they received them and then querying it if the number was high (over say 200).

 

Wilson