p.31 #2 · Sigma 14-24mm f/2.8 DG DN first impressions
Playing around mimicking a long exposure. I took 64 images @1/3s each and averaged them in PS using a Mean Stack. The result mimics a 6-stop filter but without any color shift or any IQ degradation. In fact, the shadows look so clean. (look at the 1:1 crop)
p.31 #3 · Sigma 14-24mm f/2.8 DG DN first impressions
Yes mean and median stacks do a great job cleaning up noise. It's a shame it's so painfully slow in photoshop, especially with R4 files.Trying to do 64 files at once is asking for trouble on my machine.
I'm guessing the problem is in how PS handles the files, as it tries to load all the files before doing any calculations. I suspect you could create a much faster application if you did it differently. It would be really nice if someone wrote a stand-alone program to do just that, even if it were just for DNG files.
p.31 #4 · Sigma 14-24mm f/2.8 DG DN first impressions
rvh23 wrote:
Yes mean and median stacks do a great job cleaning up noise. It's a shame it's so painfully slow in photoshop, especially with R4 files.Trying to do 64 files at once is asking for trouble on my machine.
I'm guessing the problem is in how PS handles the files, as it tries to load all the files before doing any calculations. I suspect you could create a much faster application if you did it differently. It would be really nice if someone wrote a stand-alone program to do just that, even if it were just for DNG files.
It's very slow but the results are always impressive. Photoshop mainly uses RAM when loading the images as layers so if you only have 64GB memory, it will definitely be an issue with 64-128 images. I don't see GPU being used and only one CPU core seems to be activated when running my PS Mean Stack action. Adobe should definitely modernize 'smart object' and 'image stacking' algorithms.
However, there is a faster way but it's more troublesome: Load "32" images at a time (or 16 depending on your system)
It will still take about 10min (from loading the images as layers until the mean stack is complete), so 20 minutes for 64 images in 3 steps (32 + 32 + 2) or 40 minutes for 128 images (32 + 32 + 32 + 32 + 4).
The Smooth Reflections app (from previous cameras) has a gap of 1.5s between exposures and that caused some stepping with clouds or water if shutter speeds were too short. With the new built-in intervalometer at 1 sec gap, I don't see much stepping but I have not tried with fast moving clouds with short shutter speeds yet. If you want almost no gap between images, the Sony wireless remote can take X number of shots using "Hi" continuous shooting mode.
I thought about creating a Lightroom plugin the uses imagemagick. Not sure if it could be faster than PS though.
p.31 #5 · Sigma 14-24mm f/2.8 DG DN first impressions
Fred Miranda wrote:
It's very slow but the results are always impressive. Photoshop mainly uses RAM when loading the images as layers so if you only have 64GB memory, it will definitely be an issue with 64-128 images. I don't see GPU being used and only one CPU core seems to be activated when running my PS Mean Stack action. Adobe should definitely modernize 'smart object' and 'image stacking' algorithms.
However, there is a faster way but it's more troublesome: Load "32" images at a time (or 16 depending on your system)
It will still take about 10min (from loading the images as layers until the mean stack is complete), so 20 minutes for 64 images in 3 steps (32 + 32 + 2) or 40 minutes for 128 images (32 + 32 + 32 + 32 + 4).
The Smooth Reflections app (from previous cameras) has a gap of 1.5s between exposures and that caused some stepping with clouds or water if shutter speeds were too short. With the new built-in intervalometer at 1 sec gap, I don't see much stepping but I have not tried with fast moving clouds with short shutter speeds yet. If you want almost no gap between images, the Sony wireless remote can take X number of shots using "Hi" continuous shooting mode.
I thought about creating a Lightroom plugin the uses imagemagick. Not sure if it could be faster than PS though....Show more →
Fred, for landscape, is there any disadvantage of using the electronic shutter to take multiple images? Thanks!
p.31 #6 · Sigma 14-24mm f/2.8 DG DN first impressions
AGeoJO wrote:
Fred, for landscape, is there any disadvantage of using the electronic shutter to take multiple images? Thanks!
No disadvantage with new cameras (starting with A7R3). In fact you should use silent shutter and save your shutter. Just remember to use uncompressed RAW if you want 14-bit precision. I have a feeling that the SR app (from previous cameras) yielded a 8-bit ARW resulting file as I could see some posterization in solid areas.
p.31 #7 · Sigma 14-24mm f/2.8 DG DN first impressions
Fred Miranda wrote:
No disadvantage with new cameras (starting with A7R3). In fact you should use silent shutter and save your shutter. Just remember to use uncompressed RAW if you want 14-bit precision. I have a feeling that the SR app (from previous cameras) yielded a 8-bit ARW resulting file as I could see some posterization in solid areas.
p.31 #8 · Sigma 14-24mm f/2.8 DG DN first impressions
Given what I understand to a relatively slow scan time for the A7RIV in silent shutter mode, I have been concerned that this may result in more apparent camera shake. I understand that mirror slap/anticipation would be another factor affecting camera shake, but taking that out of the equation, is my concern wrong? I'd love to be told yes that I am wrong. Thanks!
Fred Miranda wrote:
No disadvantage with new cameras (starting with A7R3). In fact you should use silent shutter and save your shutter. Just remember to use uncompressed RAW if you want 14-bit precision. I have a feeling that the SR app (from previous cameras) yielded a 8-bit ARW resulting file as I could see some posterization in solid areas.
p.31 #9 · Sigma 14-24mm f/2.8 DG DN first impressions
httivals wrote:
Given what I understand to a relatively slow scan time for the A7RIV in silent shutter mode, I have been concerned that this may result in more apparent camera shake. I understand that mirror slap/anticipation would be another factor affecting camera shake, but taking that out of the equation, is my concern wrong? I'd love to be told yes that I am wrong. Thanks!
We were referring to image averaging which should optimally be done on a tripod which means no issue with camera shake or sensor readout speed.
p.31 #11 · Sigma 14-24mm f/2.8 DG DN first impressions
Fred Miranda wrote:
It's very slow but the results are always impressive. Photoshop mainly uses RAM when loading the images as layers so if you only have 64GB memory, it will definitely be an issue with 64-128 images. I don't see GPU being used and only one CPU core seems to be activated when running my PS Mean Stack action. Adobe should definitely modernize 'smart object' and 'image stacking' algorithms.
However, there is a faster way but it's more troublesome: Load "32" images at a time (or 16 depending on your system)
It will still take about 10min (from loading the images as layers until the mean stack is complete), so 20 minutes for 64 images in 3 steps (32 + 32 + 2) or 40 minutes for 128 images (32 + 32 + 32 + 32 + 4).
The Smooth Reflections app (from previous cameras) has a gap of 1.5s between exposures and that caused some stepping with clouds or water if shutter speeds were too short. With the new built-in intervalometer at 1 sec gap, I don't see much stepping but I have not tried with fast moving clouds with short shutter speeds yet. If you want almost no gap between images, the Sony wireless remote can take X number of shots using "Hi" continuous shooting mode.
I thought about creating a Lightroom plugin the uses imagemagick. Not sure if it could be faster than PS though....Show more →
Are you saying the in-built intervalometer has a minimum gap of 1 sec between the end of one exosure and the start of the next (even with electronic shutter)? That's much longer than I would like, and way too long when you want to smooth out fast moving water but want to retain some structure. If the wireless remote can take faster shots, why can't the intervalometer?
In any case, it would be much nicer if we had an in-camera option like the smooth reflections app restored in the R4. When focus stacking, dealing with multiple shots per focal setting becomes a bit of a nightmare.
p.31 #13 · Sigma 14-24mm f/2.8 DG DN first impressions
rvh23 wrote:
Are you saying the in-built intervalometer has a minimum gap of 1 sec between the end of one exosure and the start of the next (even with electronic shutter)? That's much longer than I would like, and way too long when you want to smooth out fast moving water but want to retain some structure. If the wireless remote can take faster shots, why can't the intervalometer?
In any case, it would be much nicer if we had an in-camera option like the smooth reflections app restored in the R4. When focus stacking, dealing with multiple shots per focal setting becomes a bit of a nightmare. ...Show more →
Their previous intervalometer app (Time-lapse) started at 30 second gap so it's an improvement.
Yes, for their current built-in intervalometer, it would be great if they added a "shortest" option below 1 second but Sony does not listen to feedback like this. I do think that 1 second is not bad though and works better than the SR app which is 1.5s gap.
The SR app was great but looking at my files, I'm pretty sure we were dealing with 8-bit images as result even though Sony called it ARW. There was no latitude for the highlights and lots of posterization in solid areas...It was convenient but IQ was not the best. (We could not use silent shutter either)
I found a trick to Mean Average 64 images (A7R4 Uncompressed RAW) in only 4 minutes and my computer is not very fast. (Mac Mini 3.2 GHz 6-Core i7 with 64GB RAM)
This would mimic a 6-stop ND filter
128 images (7-stop) in 12min and 256 (8-stop) in 28min. (It completes the task without issues)
p.31 #14 · Sigma 14-24mm f/2.8 DG DN first impressions
Here is a sample of a 256 image averaged using Mean Stack from uncompressed RAW files.
It was pushed in post (shadows, colors, etc.) and still there is zero noise. It's incredible. If this was done with the SR app, the swimming poll water would have posterization.
One image from the 256 sequence (not averaged and not edited)
p.31 #15 · Sigma 14-24mm f/2.8 DG DN first impressions
Fred Miranda wrote:
Their previous intervalometer app (Time-lapse) started at 30 second gap so it's an improvement.
Yes, for their current built-in intervalometer, it would be great if they added a "shortest" option below 1 second but Sony does not listen to feedback like this. I do think that 1 second is not bad though and works better than the SR app which is 1.5s gap.
The SR app was great but looking at my files, I'm pretty sure we were dealing with 8-bit images as result even though Sony called it ARW. There was no latitude for the highlights and lots of posterization in solid areas...It was convenient but IQ was not the best. (We could not use silent shutter either)
I found a trick to Mean Average 64 images (A7R4 Uncompressed RAW) in only 4 minutes and my computer is not very fast. (Mac Mini 3.2 GHz 6-Core i7 with 64GB RAM)
This would mimic a 6-stop ND filter
128 images (7-stop) in 12min and 256 (8-stop) in 28min. (It completes the task without issues)...Show more →
I guess the other thing you can do to get a minimal gap for smooth motion blurring is to use the high-speed continuous shooting mode rather than the intervalometer. It has almost no gap as far as I can tell. I checked this morning and it shot 30x 1-second exposures in about 35 seconds. Only down side is that you can't specify exactly how many shots to take as easily, but it's not really critical for this purpose.
p.31 #16 · Sigma 14-24mm f/2.8 DG DN first impressions
Fred Miranda wrote:
Here is a sample of a 256 image averaged using Mean Stack from uncompressed RAW files.
It was pushed in post (shadows, colors, etc.) and still there is zero noise. It's incredible. If this was done with the SR app, the swimming poll water would have posterization.
The water looks pretty "dirty" on my screen. I dont see this as having an advantage over the SR app.
p.31 #18 · Sigma 14-24mm f/2.8 DG DN first impressions
pdmphoto wrote:
The water looks pretty "dirty" on my screen. I dont see this as having an advantage over the SR app.
I'm not sure what you meant by 'dirty' but this is a pebble tec poll and therefore there is texture at the bottom. The image is completely free of noise and posterization...The latter is very common with images from the SR app.
Here is a screenshot at 1:1 showing the water area:
p.31 #20 · Sigma 14-24mm f/2.8 DG DN first impressions
Fred Miranda wrote:
Two more sample tests from last night. (No filters used)
There was no wind and the clouds were pretty still.
First image:
70 images averaged for the background at f/16 (infinity distance)
16 images averaged for the foreground at f/8
Focused stacked
Second image:
256 images averaged for the background at f/16 (infinity distance)
16 images averaged for the foreground at f/8
Focused stacked
Both images were severely pushed in post and have pretty much zero noise.
Been following this thread, and I'm kind of excited about this mean stacking that your talking about. Those images look fantastic.
Just to clarify on how you did this... For the first image, is this what you did?
1. Take 70 photos with the background in focus and exposed as well as possible
2. Take 16 photos with the foreground in focus and exposed as well as possible
3. mean stack the 70 background photos to get a single clean background image
4. post process that single resulting image, focusing on the background
5. mean stack the 16 foreground images to get a single clean foreground image
6. post process the single resulting image, focusing on the background
7. then combine the resulting two clean images
