Speculation, EOS R5 camera, Style setting changes CR3 file

[johnrmoyer]

it seems to me that the "Fine Detail" style in the EOS R5 produces a different CR3 file than does the "Standard" style.

An internet friend asked about an image with a bright spot in it. I downloaded the CR3 file and used darktable to capture a screen shot of the bright spot with no demosaic applied. The style in the camera had been set to "Fine Detail". When I first got my EOS R5 I tried fine detail, but it seemed to me that the CR3 files were more difficult to edit until I switched to using the "Standard" style in the camera.

I speculate that there was some fine particle floating in the air in the foreground, maybe glitter or dust. To emphasize this fine detail, the camera brightened 3 red pixels and dimmed surrounding green pixels with a result being a tiny magenta blob in the image that is emphasized with increased contrast.

Assuming for the moment that my speculation is correct. Does anyone know whether saving a DPRAW file would enable one to undo the "Fine Detail" style when selecting another style in DPP. From my experience, that seems impossible if the camera saved RAW or CRAW and I have been too lazy to experiment with DPRAW since I stopped using the "Fine Detail" style after I had difficulty with it.

screen capture from darktable

[Peter · Accepted solution]

Thanks! I will try myself later this week with my R6.

[Peter]

My guess is still, as I wrote at Dpreview, laser. Pink dots in clusters are what I mostly have seen. In this case I see five clusters of bad pixels.

it seems to me that the "Fine Detail" style in the EOS R5 produces a different CR3 file than does the "Standard" style.

I don't think so, but I have never checked. I always shoot in Standard or Neutral.

What if you take two pictures at ISO 100 and CRAW with Fine Details in camera and two pictures at ISO 100 and CRAW and Standard in camera and then open the four in darktable? Choose LMMSE as demosaicing method and increase brightness with the exposure module. Now remove the crop at the left side and the top, see https://discuss.pixls.us/t/how-to-learn-image-dimensions-in-darktable/32147/27

Inspect the compression artifacts in the optical black area. Any difference?

[johnrmoyer]

---

@Peter wrote:

My guess is still, as I wrote at Dpreview, laser. Pink dots in clusters are what I mostly have seen. In this case I see five clusters of bad pixels.

it seems to me that the "Fine Detail" style in the EOS R5 produces a different CR3 file than does the "Standard" style.

I don't think so, but I have never checked. I always shoot in Standard or Neutral.

What if you take two pictures at ISO 100 and CRAW with Fine Details in camera and two pictures at ISO 100 and CRAW and Standard in camera and then open the four in darktable? Choose LMMSE as demosaicing method and increase brightness with the exposure module. Now remove the crop at the left side and the top, see https://discuss.pixls.us/t/how-to-learn-image-dimensions-in-darktable/32147/27

Inspect the compression artifacts in the optical black area. Any difference?

---

You may be correct. The difference I seem to see may be subjective. I attempted an experiment this morning with a ceramic coffee cup under led light. My tripod is old and I did not use a remote trigger, so the image crops do not align exactly. I did not expect the light to vary, but the autofocus point did vary. Also, the camera position varied slightly. I might attempt an experiment with more variables controlled later.

Rawtherapee with no demosaic. The highest contrast seemed to be the reflected light on the curved surface of the cup, so I attempted to crop to that.

It still seems to me like the finedetail style brightens and dims some of the raw pixels slightly to achieve greater contrast when some of the settings are auto instead of manual. But, I might be fooling myself.

I put png files resized to 400% using graphicsmagick ( convert -filter point -resize "400%" ) on my web server at: https://www.rsok.com/~jrm/raw/?C=M;O=D

	IMG_8310_faithful.png_x4.png	2023-12-27 06:18	611K
	IMG_8309_standard_x4.png	2023-12-27 06:18	619K
	IMG_8308_finedetail_x4.png	2023-12-27 06:18	613K
	IMG_8307_finedetail_x4.png	2023-12-27 06:18	614K
	IMG_8306_standard_x4.png	2023-12-27 06:18	612K
	IMG_8310.exif.html	2023-12-27 06:17	38K
	IMG_8309.exif.html	2023-12-27 06:17	35K
	IMG_8308.exif.html	2023-12-27 06:17	38K
	IMG_8307.exif.html	2023-12-27 06:17	38K
	IMG_8306.exif.html	2023-12-27 06:17	38K

exiftool -s -picturestyle -iso -shutterspeedvalue /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_830[6-9].CR3 /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8310.CR3             
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8306.CR3
PictureStyle                    : Standard
ISO                             : 640
ShutterSpeedValue               : 1/512
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8307.CR3
PictureStyle                    : Fine Detail
ISO                             : 800
ShutterSpeedValue               : 1/664
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8308.CR3
PictureStyle                    : Fine Detail
ISO                             : 640
ShutterSpeedValue               : 1/512
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8309.CR3
PictureStyle                    : Standard
ISO                             : 800
ShutterSpeedValue               : 1/664
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8310.CR3
PictureStyle                    : Faithful
ISO                             : 800
ShutterSpeedValue               : 1/664
    5 image files read

for i in 06 07 08 09 10
do
exiftool -G0:2 -s -h IMG_83${i}.CR3 > IMG_83${i}.exif.html
done

[Peter · Accepted solution]

Thanks! I will try myself later this week with my R6.

[johnrmoyer]

I tried again, this time with sun through a window, but the window had a screen. There was wind and moving clouds. The tripod was more stable, but might have moved when I pressed menu button to change styles. Remote trigger was Canon camera connect on an old iPad. The image is the edge of a geranium leaf in sunlight and shadow behind the leaf.

I guess I will need to get out my old light box to get constant light.

exiftool -s -picturestyle -datetimeoriginal -iso -shutterspeedvalue /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_831[1-5].CR3 ======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8311.CR3
PictureStyle : Standard
DateTimeOriginal : 2023:12:27 14:17:34
ISO : 1250
ShutterSpeedValue : 1/166
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8312.CR3
PictureStyle : Fine Detail
DateTimeOriginal : 2023:12:27 14:18:07
ISO : 1250
ShutterSpeedValue : 1/128
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8313.CR3
PictureStyle : Fine Detail
DateTimeOriginal : 2023:12:27 14:18:10
ISO : 1250
ShutterSpeedValue : 1/128
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8314.CR3
PictureStyle : Standard
DateTimeOriginal : 2023:12:27 14:18:27
ISO : 1250
ShutterSpeedValue : 1/128
======== /Volumes/EOS_DIGITAL/DCIM/100CANON/IMG_8315.CR3
PictureStyle : Faithful
DateTimeOriginal : 2023:12:27 14:18:59
ISO : 1250
ShutterSpeedValue : 1/128
5 image files read

	IMG_8315.exif.html	2023-12-27 15:46	35K
	IMG_8314.exif.html	2023-12-27 15:46	35K
	IMG_8313.exif.html	2023-12-27 15:46	35K
	IMG_8312.exif.html	2023-12-27 15:46	35K
	IMG_8311.exif.html	2023-12-27 15:46	35K
	IMG_8315_faithful_x4.png	2023-12-27 15:38	1.6M
	IMG_8314_standard_x4.png	2023-12-27 15:38	1.6M
	IMG_8313_finedetail_x4.png	2023-12-27 15:38	1.6M
	IMG_8312_finedetail_x4.png	2023-12-27 15:38	1.6M
	IMG_8311_standard_x4.png	2023-12-27 15:38	1.6M

[Peter · Accepted solution]

I did a compartion between raw files with Fine details and Neutral. At least I couldn't see any difference.