Leica Q: Streaks in Images

[dkaye]

Advertisement (gone after registration)

Is anyone else seeing these streaks? Normally I don't see them, but sometimes they appear. Are they a strange type of lens flare from the night street lamps (outside of this crop)? I haven't been able to reliably reproduce this, but it has happened more than once.

 

 

Image Notes: This is a cropped JPEG made from the RAW file. The exposure is cranked up about 1.5 stops from the original, which is certainly underexposed. But the streams are visible even without this exposure adjustment. You can see the uncropped JPEG here: https://www.dropbox.com/s/7wq6dg0ms6hklae/L1050035-c.jpg?dl=0

 

 

[Advertisement]

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

[thighslapper]

Fairly typical artefacts that appear when you try to excessively recover shadow detail, particularly if you use high ISO's ...... that's the horizontal ones ...... the two white vertical bands on the R look more like internal reflections caused by bright oblique lighting just off to the right .... do you have a UV filter on the camera ? 

[CharlesL]

On the full shot, the streaks are curved in one direction at the top left on the sky, but curved in the other direction on the pavement at the lower middle left. Could they be something like the distortions sometimes seen in desert air?

Or are the curvatures a result of LR applying distortion corrections for the Leica Q?
 

[jaapv]

That is simply the distortion correction for the lens.

[dkaye]

Fairly typical artefacts that appear when you try to excessively recover shadow detail, particularly if you use high ISO's ...... that's the horizontal ones ...... the two white vertical bands on the R look more like internal reflections caused by bright oblique lighting just off to the right .... do you have a UV filter on the camera ? 

 

I've owned more than 20 digital cameras and I've never seen artifacts like these when recovering shadows. And 1.5 stops is not much "recovery". No UV filter used. ISO 6400, which is admittedly high for this camera. High noise, sure. But that still doesn't explain the streaking. 

[CharlesL]

If you can see the streaks on in-camera review, next time they appear you can photograph the same scene at a series of ISOs for a possible clue to diagnosis.

 

[Nuno Cruces]

Advertisement (gone after registration)

This appears to be the banding that DPReview speaks of, when pushing the Raw files: http://dpreview.com/previews/leica-q-typ116/6

[CharlesL]

This appears to be the banding that DPReview speaks of, when pushing the Raw files: http://dpreview.com/previews/leica-q-typ116/6

 

They talk about banding, but with a cursory look and downloading one DNG, I did not see an example for comparison with what Doug showed us.

[dkaye]

It's definitely not what I think of as "banding", which to me accentuates the continuous tonal or color transitions in an image. This is alternating stripes of light and dark. Also, It has nothing to do with "pushing" the RAW file. It's quite obvious in the unaltered RAW file as well. 

[viramati]

Hi any chance of you putting the original DNG in dropbox as it will have a lot more info in it

[dkaye]

Here's the original RAW/DNG, straight off the SD card: https://www.dropbox.com/s/7c06j3ngu79rva8/L1050035.dng?dl=0

[jip]

It's banding, it's especially visible in the shadows.

 

It's not perfectly straight/horizontal because the Q does in camera lens distortion correction, even to the DNG files. 

 

I'm not saying it looks nice, but it is 'normal' to be able to see banding on high ISO RAW files without pushing... I sometimes see it on the M (Typ 240) and S (Typ 006) files as well. 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[CharlesL]

...the Q does in camera lens distortion correction, even to the DNG files. 

 

 

Does the Leica Q do in camera lens distortion correction, even to the DNG files? I doubt it. How do we explain this: compare a Leica Q DNG opened in Raw Therapee, without a customized camera profile, to the camera JPG or to the JPG embedded in the DNG file. You will see differences related to distortion and vignetting. You will also notice that area at the edges of the image is present in the DNG but missing in the JPG, presumably removed during lens correction applied to make the JPG but not applied to the DNG.

 

I don't have Lightroom. I presume the edition of LR for the Q that Leica makes available for download has a camera profile or preset, or it reads parameters written into the manufacturer data or Exif of the DNG. But the DNG image data proper does not appear to have lens corrections applied.

 

What is the source of the "Lens Profile Information" advisory that you screen captured? If LR, the explanation offered here would seem to apply.

 

The streaks are indeed visible in Doug's DNG image. Further investigation would go faster if someone discovers a way to reproduce it, perhaps a certain interior setup, and compares shots at different ISOs and exposures.

 

[dkaye]

I don't know what, if anything, the Q does internally to tweak RAW/DNG files, but Lightroom definitely applies a "profile" to DNG files during import. 

 

 

If you look at a DNG from the Q using a RAW file viewer that doesn't apply the profile (eg, RawDigger), it's quite clear. I believe the profile is added to the DNG file by the Q. I also believe micro-four-thirds cameras do the same thing. The profiles may actually be stored in the lenses. As I recall, versions of LR prior to the most-recent one weren't capable of applying the profile and hence displayed the un-profiled image. I do not believe there's any way in LR to disable this application of the profile.

[CharlesL]

The Exif data revealed by exiftool includes these three entries:
    Opcode List 1                   : (Binary data 28 bytes)
    Opcode List 3                   : (Binary data 184 bytes)
    Default User Crop               : 0.1 0.1 0.9 0.9
If the default user crop is twenty percent off width and height, that is a lot! Perhaps this crop is related to statements asserting that the lens is approximately a 24mm focal length, but you get a 28mm crop.

The two opcode lists could be instructions for correcting distortion, vignetting, and perhaps chromatic aberration used by programs in the know such as LR.

If someone knows how to strip these entries from a DNG file while leaving it, with internal position pointers, in usable form, LR might very well behave like Raw Therapee upon opening the file.

 

[dkaye]

I *believe* that's the crop to simulate a 35mm lens. I usually have the camera set to this mode. I can still see and use the full 28mm frame. When I open the RAW file in Lightroom, it shows me the 35mm crop. But I can use the Crop Tool to open it up to the full 24MP 6000x4000px image.

[CharlesL]

I *believe* that's the crop to simulate a 35mm lens. I usually have the camera set to this mode.

 

That must be correct. I looked at another sample DNG from the Web, and the numbers were 0 0 1 1, in other words, no User Crop.

Also, there are a lot of other Manufacturer Notes in the file in binary format. The supplied LR is almost surely aware of much of it, while other raw developer programs are not.

 

It's almost certain that the streaks are in the raw pixel data proper. The big question is how often they appear. After all, this photo in low light with a mix of light sources at different color temperatures is a challenge for almost any camera.