Extreme dynamic scene loses most details

[Floessie]

Hi Ingo and DrSlony,

I've shot a scene using Magic Lantern's auto HDR bracketing, which correctly resulted in eight shots all 2EV apart. I naively saved them to a 16b DNG, knowing the embedded preview would be crap. Only later, when I struggled developing the DNG in RT, I learned that this DNG was only 1MB in size. 😮

Then I saved the series in 32b, this resulted in a 16MB DNG, again undevelopable. Only when I reduce the set to the "innermost" three to four shots I get a decent DNG.

What's wrong with the scene, and why does HDRMerge lose that data instead of accumulating all possible data into the DNG?

You can find the files here.

Best,
Flössie

[Beep6581]

Would be good to first exclude Magic Lantern as the source of the issues. I've never processed Magic Lantern raws in HDRMerge. Could you try the maximum bracketing set allowed by standard Canon firmware?

I renamed the files in order of EV, 1-8.cr2.

Worth nothing that --batch mode detects 3 sets, and the black levels and wb are not constant.
See #1 and #6 (comment)

The white balance is passed untouched to the output DNG

$ ~/programs/hdrmerge-master/hdrmerge --batch -b 16 --no-align --no-crop -p half -w 11432 -vv *.cr2
Using LibRaw 0.18.5-Release
Set 0: 4.cr2
Set 1: 3.cr2 5.cr2
Set 2: 2.cr2 6.cr2 8.cr2 1.cr2 7.cr2
Skipping single image 4.cr2
[  0%] Loading 3.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
3.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/100sec f8 EV:-12.6439
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1601 1024 2456 1024, cblack: 2048 2048 2048 2048
[ 33%] Loading 5.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
5.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/6sec f8 EV:-8.58496
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1579 1024 2411 1024, cblack: 2048 2048 2048 2048
Load files: 5.68887 seconds
Using custom white level 11432
Compute response functions: 0.496783 seconds
Generate mask: 0.948835 seconds
[100%] Done loading!
Writing result to /home/morgan/downloads/floessie/3-5.dng
Writing /home/morgan/downloads/floessie/3-5.dng, 16-bit, 5202x3465
[  0%] Rendering image
Adjusted white balance: 1.56348 1 2.39844 1
Fatten mask (SSE version): 0.0883232 seconds
Blur: 0.465352 seconds
Compose: 0.843472 seconds
[ 33%] Rendering preview
Render preview: 1.12768 seconds
[ 66%] Writing output
Write output: 7.23871 seconds
[100%] Done writing!
[  0%] Loading 2.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
2.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/400sec f8 EV:-14.6439
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1586 1024 2422 1024, cblack: 2048 2048 2048 2048
[ 16%] Loading 6.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
6.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/1.66667sec f8 EV:-6.73697
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1586 1024 2422 1024, cblack: 2049 2048 2048 2048
[ 32%] Loading 8.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
8.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/0.1sec f8 EV:-2.67807
b4b4b4b4 RGBG, sat 13584, black 2049, flip 5, wb: 1799 1024 1893 1024, cblack: 2050 2050 2049 2050
[ 48%] Loading 1.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
1.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/1600sec f8 EV:-16.6439
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 2166 1024 1584 1024, cblack: 2048 2049 2048 2048
[ 64%] Loading 7.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
7.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/0.4sec f8 EV:-4.67807
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1799 1024 1893 1024, cblack: 2049 2049 2048 2049
Load files: 14.29 seconds
Using custom white level 11432
Compute response functions: 2.16348 seconds
Generate mask: 1.04476 seconds
[100%] Done loading!
Writing result to /home/morgan/downloads/floessie/1-8.dng
Writing /home/morgan/downloads/floessie/1-8.dng, 16-bit, 5202x3465
[  0%] Rendering image
Adjusted white balance: 2.11523 1 1.54688 1
Fatten mask (SSE version): 0.076665 seconds
Blur: 0.471906 seconds
Compose: 1.0446 seconds
[ 33%] Rendering preview
Render preview: 1.11075 seconds
[ 66%] Writing output
Write output: 1.26277 seconds
[100%] Done writing!

$ ~/programs/hdrmerge-master/hdrmerge -o 1-8.dng -b 16 --no-align --no-crop -p half -w 11432 -vv *.cr2
Using LibRaw 0.18.5-Release
[  0%] Loading 1.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
1.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/1600sec f8 EV:-16.6439
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 2166 1024 1584 1024, cblack: 2048 2049 2048 2048
[ 11%] Loading 2.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
2.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/400sec f8 EV:-14.6439
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1586 1024 2422 1024, cblack: 2048 2048 2048 2048
[ 22%] Loading 3.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
3.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/100sec f8 EV:-12.6439
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1601 1024 2456 1024, cblack: 2048 2048 2048 2048
[ 33%] Loading 4.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
4.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/25sec f8 EV:-10.6439
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1584 1024 2422 1024, cblack: 2048 2048 2048 2048
[ 44%] Loading 5.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
5.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/6sec f8 EV:-8.58496
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1579 1024 2411 1024, cblack: 2048 2048 2048 2048
[ 55%] Loading 6.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
6.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/1.66667sec f8 EV:-6.73697
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1586 1024 2422 1024, cblack: 2049 2048 2048 2048
[ 66%] Loading 7.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
7.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/0.4sec f8 EV:-4.67807
b4b4b4b4 RGBG, sat 13584, black 2048, flip 5, wb: 1799 1024 1893 1024, cblack: 2049 2049 2048 2049
[ 77%] Loading 8.cr2
Number of frames : 1
Invalid aperture: 0 replaced by aperture: f8
8.cr2: 5202x3465 (5344x3516+142+51, by Canon EOS 600D, 100ISO 1/0.1sec f8 EV:-2.67807
b4b4b4b4 RGBG, sat 13584, black 2049, flip 5, wb: 1799 1024 1893 1024, cblack: 2050 2050 2049 2050
Load files: 20.7374 seconds
Using custom white level 11432
Compute response functions: 2.15621 seconds
Generate mask: 0.398727 seconds
[100%] Done loading!
Writing result to 1-8.dng
Writing 1-8.dng, 16-bit, 5202x3465
[  0%] Rendering image
Adjusted white balance: 2.11523 1 1.54688 1
Fatten mask (SSE version): 0.0238369 seconds
Blur: 0.249987 seconds
Compose: 0.407139 seconds
[ 33%] Rendering preview
Render preview: 0.396647 seconds
[ 66%] Writing output
Write output: 0.502799 seconds
[100%] Done writing!

Fixing the aperture made no difference:

exiftool -P -overwrite_original_in_place -ApertureValue="8.0" -FNumber="8.0" *.cr2

[Floessie]

@Beep6581

Would be good to first exclude Magic Lantern as the source of the issues. I've never processed Magic Lantern raws in HDRMerge. Could you try the maximum bracketing set allowed by standard Canon firmware?

The Canon EOS 600D can only do three bracketed shots (max. +-2EV) with stock firmware. That's why I use ML. There's nothing special about the CR2s (apart from not having the bracketing sequence tag AFAIK), as ML only changes the shutter time the way I would do manually. So you can take the center three images and basically get what Canon bracketing would have shot. Worth to mention, I never had a problem with ML auto bracketed shots, but this here is extreme in it's dynamic range. Most of my bracketed scenes causes four to five shots 2EV apart.

Worth nothing that --batch mode detects 3 sets, and the black levels and wb are not constant.

About the three sets: The longest exposure might be longer than the default set gap of HDRMerge. Plus, I had to take the series twice, because someone was walking through the shot. Maybe I made a best-of selection, can't remember and don't have the full series here to check.

About BL and WB: I can't tell you anything about BL, but WB was set to AutoWB because I wasn't aware that this would matter. Does it?

Fixing the aperture made no difference.

This was shot with my manual Samyang 35 1.4, most probably at f/2 or maybe f/2.8. I had no problems with the combination of this camera, lens, and ML auto bracketing in the past, but the scenes weren't that extreme.

HTH,
Flössie

[Floessie]

@Beep6581 In your batch run, why is the 10s exposure (1/0.1) still rated -2.67807EV? Is that normal? I would have expected at least 0 for the brightest image.

[Beep6581]

@Floessie I'm "under the impression" that it doesn't matter as long as the step size is right, but I will test both the EV and WB on some other shots this weekend.

[Floessie]

@Beep6581 @heckflosse Any idea where to look? Could it be some cancelation problem like the one Alberto fixed for Fattal? I'd really like to develop that shot...

Best,
Flössie

[heckflosse]

@Floessie I have an idea where to look. Have to check that. I will tell you later today.

[heckflosse]

@Floessie I would start here :

hdrmerge/Image.cpp

Line 99 in febf9ee

void Image::computeResponseFunction(const Image & r) {

[Floessie]

@heckflosse The problem is the darkest image.

diff --git a/src/Image.cpp b/src/Image.cpp
index da5dfec..5fea6e7 100644
--- a/src/Image.cpp
+++ b/src/Image.cpp
@@ -153,6 +153,7 @@ void Image::computeResponseFunction(const Image & r) {
         alglib::spline1dfitreport rep;
         alglib::spline1dfitpenalized(values, adjValues, i, 200, 3, info, response.nonLinear, rep);
         response.linear = alglib::spline1dcalc(response.nonLinear, response.threshold) / response.threshold;
+        printf("algo 1: %lf\n", response.linear);
     } else {
         response.threshold = 65535;
         // Fallback method for dark images:
@@ -171,6 +172,7 @@ void Image::computeResponseFunction(const Image & r) {
             }
         }
         response.linear = numerator / denom;
+        printf("algo 2: %lf\n", response.linear);
     }
 }

With the whole set I get:

algo 2: 0,234110
algo 2: 0,074496
algo 2: 0,017766
algo 2: 0,004300
algo 2: 0,001041
algo 2: 0,000252
algo 2: 0,000062

And most of the resulting image is black. Without 20170724-230458.cr2, everything is fine:

algo 2: 1,535183
algo 2: 0,366117
algo 2: 0,088617
algo 2: 0,021460
algo 2: 0,005202
algo 2: 0,001274

I tried this and that inside computeResponseFunction(), understanding the code only fragmentally, but without success.

[Beep6581]

@Floessie is this still reproducible?

[Floessie]

@Beep6581 Last time I checked, it was still reproducible and there were no changes concerning computeResponseFunction(). I can reevaluate, though (if I can find the stack in question again).

[Floessie]

@Beep6581 Problem still exists. Range is 0, 4, 6, 8, ..., 16 EV.