Leica SL Firmware v.2.0 answers our wishes

[steppenw0lf]

Advertisement (gone after registration)

I went to the Leica SL download area and tried to download the newest manual. But there is still the link to the version from last October.

Can you tell me where to find the updated manual ?

I think somebody mentioned somewhere that the manual has been upgraded - or am I wrong ?

 

Thanks.     Stephan

[J S H]

These parameters are complex enough that I could see timing issues arising with fast bursts.  The distortion and LCA corrections are put into 12 opcode parameters, that depend quite sensitively on the lens focal length, since the 24-90 has barrel distortion from 24 until around 50 mm and a smaller amount of pincushion distortion from there to 90 mm.  There are four parameters for each color, and they are not the same...  You won't find them in the EXIF, but in a section for manufacturer's profile that a tool like the Adobe SDK can extract.  It's called "Warp Rectilinear," and that's what it does.

 

So it might be valuable to have another user with the 24-90 verify this and report it if it is a regular occurence.  Of course card speed might be a factor.  And the cards with the fastest specs are not always the fastest in a Leica camera.

 

scott 

For what it's worth, the problem isn't tied to the 24-90 or firmware 2.0. I've seen it with M lenses and with both 1.2 & 2.0 FW (I don't recall seeing it with the original firmware, but it's possible that I never pushed the camera hard enough to cause problems during that short period before FW 1.2 was released). I believe that it takes quite a bit of processing power for the camera to implement the lens corrections, display them in real time in the VF and write them to the DNG as it's being recorded on the card. My suspicion is that the code may be buggy and if (when) the processor is overtaxed, you will start to see errors.

 

I see it most often when I use the interval drive mode of the camera and choose 1 sec intervals with say a 1/8 sec shutter speed. Theoretically, there should be plenty of time for the camera to do everything needed, but somehow it loses itself in the process and starts saving files without the lens corrections embedded. I don't think it has anything to do with the card, but more to do with the buffer, as it only manifests after the buffer fills up. Also, I am using Sandisk Extreme 512 GB UHS II cards (I tested with UHS III cards and found absolutely no difference in final throughput once the buffer was full, so I now use the larger capacity UHS II cards exclusively).

 

I suspect that if you were to set up a 450 shot interval sequence with these parameters, you would see the same problems towards the end of the sequence. Granted, that's not everyone's cup of tea and it would actually be hard to figure out, as you would really need to render the sequence as video footage (with either the slideshow module, or in After Effects) to most effectively discern whether and when the problem occurs.

[adamdewilde]

Along the lines of what you are seeing, I have discovered a similar problem. I suspect the two issues are connected...

 

There is definitely a glitch in how the SL manages the lens corrections and writes them to the DNG files. I am a cinematographer and I use the SL occasionally for timelapse sequences because it is my most weather resistant camera.

 

If the camera has plenty of time to do everything, it usually produces DNG files with the appropriate lens corrections embedded, which are then visible in LR and everything is as it should be. However, when I try to shoot the sequence as fast as the buffer will allow, I often see DNG files in the sequence without the lens corrections embedded. When this happens, the footage is unusable, as the corrections are quite noticeable, especially on the wide end. Within the confines of LR, there is no way to fix the problem, as the corrections are not being implemented on a level that the user can access. It's possible that some other RAW developer would allow you to disable the corrections, but it obviously can't create them out of thin air, so the problem doesn't have a fix on the user side and will need to be fixed on the firmware side. I have never noticed this behavior by looking through the viewfinder, but it would be hard to discern, unless you were looking very carefully. Either way, there is definitely an issue with the firmware, but it happens so randomly that it may be tough for Leica to diagnose and correct (or they may already know about the issue, but cannot fix it).

 

Attached are 2 successive images from a timelapse sequence that were taken 1 second apart with the 24-90 at 30MM. You can see that the 2nd image is a little wider, maybe 10%. It wouldn't be a huge issue for normal photography, but it is a deal breaker for timelapse cinematography. It goes without saying that my other "professional" cameras don't suffer from this problem. Also, for what it's worth, it happens with the M lenses as well, assuming that the camera knows what lens is attached and that there is a correction profile for the lens.

 

 

 

So much for no compromise

This is why I don't like digitally corrected lenses. This is why I don't believe the Vario is worth the price since it's digitally corrected. I could see if you could turn off corrections. BUT if you did that, the wide images would be unusable. Hence my annoyance with digital corrections. 

 

Maybe the primes will be better.

[jrp]

I don't know whether anyone is keeping a list of firmware glitches to be addressed in the next release, but the exposure compensation values that the SL writes into DNG files are misinterpreted by Photo Mechanic / Lightroom.  Instead of -1, I get +6552.6, for example, or +6553.1 for -1/2, or maybe it's the other way around...

[ramarren]

As far as I'm aware, LR does not interpret camera settings and tweak adjustments accordingly. It just displays them. 

 

I don't know what "+6552.6" would mean with respect to exposure compensation ... LR's EC runs -5.0 to +5.0. Those are probably color temperature numbers, which are mostly wacky anyway. 

[J S H]

So much for no compromise

This is why I don't like digitally corrected lenses. This is why I don't believe the Vario is worth the price since it's digitally corrected. I could see if you could turn off corrections. BUT if you did that, the wide images would be unusable. Hence my annoyance with digital corrections. 

 

Maybe the primes will be better.

It's really a shame. Truthfully, I would be fine with the camera writing the DNG's to the card with no lens corrections embedded, as long as I had control over everything in LR (I know that I could accomplish this with M lenses and a Novoflex adapter, but that's a step backwards). But if I use the 24-90, my hands are tied when the camera doesn't write the files correctly. I know that I could probably use a different RAW developer to go in and monkey with the embedded corrections, but that would trash my existing workflow which is optimized for high volume (100,000+ images per month). Instead, I am forced to use my Canon/Sony systems because they are completely reliable. It's pretty disappointing that the SL loses to a lowly A7S when it comes to reliability, but that's where we are.

 

Leica really needs to get their act together if they want to continue trumpeting the SL as a "professional" camera system.

[dgktkr]

Advertisement (gone after registration)

I don't know whether anyone is keeping a list of firmware glitches to be addressed in the next release, but the exposure compensation values that the SL writes into DNG files are misinterpreted by Photo Mechanic / Lightroom.  Instead of -1, I get +6552.6, for example, or +6553.1 for -1/2, or maybe it's the other way around...

 

 

I'm not able to reproduce your results.

 

Lightroom (6.5.1), Iridient Developer (3.1.0) and exiftool (10.0.7) all properly display the exposure bias (exposure compensation) recorded in the one SL (FW 2.0) DNG file that I looked at.

 

dgktkr

[jrp]

As far as I'm aware, LR does not interpret camera settings and tweak adjustments accordingly. It just displays them. 

 

I don't know what "+6552.6" would mean with respect to exposure compensation ... LR's EC runs -5.0 to +5.0. Those are probably color temperature numbers, which are mostly wacky anyway. 

 

 

Well it may just display them,in which case they are saved the camera in the incorrect format.  But I think that they are saved as a number in raw binary format -- the sign bit being set adds 65536 = 2^16, which I assume is the sign bit.  Anyway, surely Leica would check how Lightroom shows the value and either get Adobe to change, or correct their own firmware.

[jrp]

I'm not able to reproduce your results.

 

Lightroom (6.5.1), Iridient Developer (3.1.0) and exiftool (10.0.7) all properly display the exposure bias (exposure compensation) recorded in the one SL (FW 2.0) DNG file that I looked at.

 

dgktkr

 

Did you look at a file with -ve exposure compensation?

[dgktkr]

I don't know whether anyone is keeping a list of firmware glitches to be addressed in the next release, but the exposure compensation values that the SL writes into DNG files are misinterpreted by Photo Mechanic / Lightroom.  Instead of -1, I get +6552.6, for example, or +6553.1 for -1/2, or maybe it's the other way around...

 

 

I repeated a test of the EXIF information with negative exposure compensation. Instead of -2/3, which is set in the camera, exiftool, Lightroom and Iridient Developer show 6553, 6553 and 6553.0, respectively.

 

So it does indeed look like FW 2.0 is not reporting negative exposure compensation properly in DNG files.

 

dgktkr

[jrp]

Thanks.  So how can we get this fixed?

[ramarren]

DNGs and JPEGs ... sequence of five files set to JPEG+DNG.

 

Using exiftool -ExposureCompensation *.jpg *.dng > EXIF.txt

this is what I get: 

 

======== 160430-1020121.jpg

Exposure Compensation           : +6552

======== 160430-1020122.jpg

Exposure Compensation           : +6553

======== 160430-1020123.jpg

Exposure Compensation           : 0

======== 160430-1020124.jpg

Exposure Compensation           : +0.6

======== 160430-1020125.jpg

Exposure Compensation           : +1.3

======== 160430-1020121.dng

Exposure Compensation           : +6552

======== 160430-1020122.dng

Exposure Compensation           : +6553

======== 160430-1020123.dng

Exposure Compensation           : 0

======== 160430-1020124.dng

Exposure Compensation           : +0.6

======== 160430-1020125.dng

Exposure Compensation           : +1.3

 

Of course, it should be noted that these values are simply what's recorded in the EXIF data. They make no difference to how the image appears or is rendered by LR or any other image processing software. 

 

(The bug looks like a sign extension error to me... but that's for Leica to figure out.)

[jrp]

Thanks,  An uninspiring schoolboy error.  How do we get it fixed?

[ramarren]

I've reported it on the firmware beta test list ... Hopefully that will get someone's attention. Don't expect it will be fixed overnight, however. The fact that it's taken this long for someone to notice puts it in the class of " nice to have" fixes. 

[jrp]

Thanks. Equally it will not be hard to fix. 

 

Had my first SL lockup today. New firmware. Was selecting a non-6-bit M lens from the menu. Got to the point where you click the rear wheel / joy stick to select, when the camera locked and would do nothing, not even switch off. Remedy: remove battery. 

[kiubabar]

It still has bug after you attached the SF58. The back screen can't show the correct ISO value during the selection. Has anyone experience this situation ????

[FMB]

I  don't know if the thing has been already said but at least in my case every time I format my card again  the numbering increases ten thousand units. Is there some tip to solve this issue  in camera although in manual mode?

 

Francisco.

[thighslapper]

I  don't know if the thing has been already said but at least in my case every time I format my card again  the numbering increases ten thousand units. Is there some tip to solve this issue  in camera although in manual mode?

 

Francisco.

 No, you are stuck with it. I format my cards in the computer with SD Formatter specifically to avoid this. 

 

This is the sort of simple idiotic problem that drives me mad.

 

It was mentioned many months ago and I emailed Leica directly about it.

 

I cannot believe this and the incrementing issue (going from xx999 to 10000) would take more than a re-write of a few lines of code....... which is why I find it so exasperating. 

[FMB]

thighslapper,

 

Thank you and I'm very angry because of we are obliged to bear with idiotic problems for months or years...

 

Francisco

[Advertisement]

Hallo,

Look here: Leica L-Mount

[Ranma13]

2 firmware why DNG format or automatic noise reduction, a waste of the battery and the time, it should be allowed to do LR.

2.0固件为什么DNG格式还是自动降噪，浪费了电池和时间，应该让LR去做。

 

For those that are interested in a more accurate translation that's not Google translated (I speak Chinese):

 

"Why does the 2.0 firmware still do automatic noise reduction in DNG mode? It's a waste of battery life and time, and should be handled by Lightroom."

 

To play Devil's advocate, he does have a point. The forced automatic noise reduction means that certain types of shooting either become very difficult or impossible to do, most notably fireworks (especially towards the end when they start shooting them in rapid succession) and time lapses (though it's debatable whether it's smart to wear out the shutter on a $7000+ camera). Having owned the M240, Q, and SL, I understand that this is just part of Leica's ideology, but it would still be nice to have the option. Sometimes I really don't want to or can't afford to take a 2-minute long exposure, then sit around for another 2 minutes while I wait for the camera to detect hot pixels.