p.1 #1 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
Results for Canon at this time is only in Jpeg
Would like to see the Fuji GFX100S with its 400mp mode & Panasonic Lumix S1R in high-res which can output 187 megapixel photos included in this shoot out
p.1 #4 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
I'd be curious to see the pixel shift differences between the a1 and a7R V. The new IBIS module on the a7R V supposedly has more precision (read it somewhere.....could be pure BS).
I did some 16 shot pixel shifts with the a7R V and it was pretty mind blowing.
p.1 #5 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
If you can’t check the file in the field, what’s the point? Come home to find every high res shot didn’t work out? No thanks. Sony, Canon and Fuji’s implementation is pointless. Olympus and Leica/Panasonic do it in camera. They produce a raw file and the Panasonic and Leica also give you a single shot standard resolution file along side the high res, just in case.
p.1 #6 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
Huh? Canon’s is in camera.
flash wrote:
If you can’t check the file in the field, what’s the point? Come home to find every high res shot didn’t work out? No thanks. Sony, Canon and Fuji’s implementation is pointless. Olympus and Leica/Panasonic do it in camera. They produce a raw file and the Panasonic and Leica also give you a single shot standard resolution file along side the high res, just in case.
p.1 #7 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
It is interesting technology. On the other hand, the number of situations in which having 8x the MP of an R5 would be more of a plus than the process of shooting this way is a minus would be pretty darned small.
p.1 #9 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
Jman13 wrote:
Huh? Canon’s is in camera.
Yes, but Canon says it won't use the full-res images during in-camera review, so you need something else if you want to inspect it at 1:1 for merging artefacts.
My phone handles the 400MP files fine, so you could use a phone/tablet in the field to check for artefacts. Maybe this will make Canon fix their "connect phone to camera" workflow to be more efficient and robust
p.1 #10 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
I once took a macro shot of machined metal and engineering plastic (e.g. a nice camera top) with the GFX 100S, 120mm GF and using the pixel shift mode and the results were impressive.
p.1 #11 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
Jman13 wrote:
Huh? Canon’s is in camera.
Sorry. I should have written that better. You can’t review the full file in camera. You’ll need to download the file to another device to see if it worked. Hardly ideal in the field.
And let’s not forget it’s just a jpeg. A lossy compressed 8 bit file. Not exactly the benchmark for a robust file for post processing. Panasonic and Leica get a raw file that you can review in camera with the same resolution and have done so for a couple of years.
Gordon
Apr 06, 2023 at 03:21 PM
osv2 Offline • • • • [X]
p.1 #12 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
flash wrote:
Panasonic and Leica get a raw file that you can review in camera with the same resolution and have done so for a couple of years.
i've never heard of a camera that lets you open and view raw files in-camera.
p.1 #13 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
flash wrote:
If you can’t check the file in the field, what’s the point? Come home to find every high res shot didn’t work out? No thanks. Sony, Canon and Fuji’s implementation is pointless. Olympus and Leica/Panasonic do it in camera. They produce a raw file and the Panasonic and Leica also give you a single shot standard resolution file along side the high res, just in case.
Gordon
This is why most people I have seen want the camera to save all the RAW sub-images as separate raw files (ideally in a folder or named/numbered as related). The last thing I would want for this type of operation is an in-camera with minimal resources compared to a PC and irreversible processing. By saving the RAW sub-images, the worse case is you have "only" multiple 45MP images. Then you could use software, now and in the future, to algorithmically combine the images and/or you could patch up tricky areas with overlays.
Checking a 400mp image in the field seems pointless unless you have a laptop with a big screen with you. You certainly will miss fatal flaws with the image scanning around it with the small LCD on the back of the camera.
p.1 #14 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
Whew! I almost clicked on that video ....but then I noticed who made it. People should really put TN warning right in the title of their post....just a little heads up for those of us not paying attention.
p.1 #15 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
After the initial excitement of hearing about 400mpx, I listened to the video and my reaction is - if you can't get a tiff and the raw files out of it - it (Canon implementation) fails. eg Sony is better even if Canon out resolves the Sony. As a mostly canon user with an R5, I am hoping that the next firmware update provides something similar to Sony for R5 - jpgs are lossy. However even Sony implementation where you have to - shoot near perfectly on a tripod and run it through software makes it questionable as being more than a theoretical product.
My view is that both Sony and Canon methods and pixel shifting in general sucks, but Canon sucks more. I need a 100mpx camera with less bayer affects and 21 stop dynamic range, not a 400mpx pasted together shot from multiple exposures. [beyond 100mpx, is there even a lens that keeps up]
400 vs 45mpx [canon or 2.5 for sony] is in theory capable of about 3x the resolution however:
I am not seeing 3 x the resolution
There are not many lens that resolve beyond 45mpx anyway particularly for landscape
Wildlife does not sit still and capturing a moment is capturing 1 frame in a burst of many. Just getting to the right spot waiting for the grizzly is hard enough with one shot. 9 shots is going to slow down fps.
Landscape light in golden hour are constantly changing - so is water, branches, ice, and atmosphere
Typically I will return more than a dozen times to the same shot to get the perfect cloud, light, time of year.... and I change my mind as the light arises and shift my composition
Typically I have a nd filter on for landscape [reducing the resolution]
The majority of the time I am at f8 or higher [further reducing the resolution of the lens]
I usually shoot AEB to have more dynamic range and blend or HDR/blend - throwing in more processing would be challenging [an order of magnitude increase in processing]
My compositions for wildlife and landscape are constantly changing with the scene, thus putting the camera in the exact same spot for the exact same composition is not practical - for multiple compositions. I would have to shoot almost all in pixel shift and ask nature or the bear to sit still.
Shooting 9 or more shots, processing x 3 and blending would be a nightmare, not to mention the disk space.
In order for this to be useful for me, I would need the camera to save the raws it creates and create a Tiff in body. Or create a raw with the 9 pictures embedded.
I am curious. I have not visited the Sony forum for a while and my A7r2 is collecting dust - not worth selling and I still have lens in case Sony does something spectacular. Sony has had this feature for a while, has anyone had success with it in Landscape photography - eg is it practical for shooting a dynamic scene. It feels like it delivers higher resolution but it interferes with the artistic (composition) flow of the process. Is it just a studio product useful feature? Should I be jealous of Sony?
p.1 #16 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
Scott Stoness wrote:
After the initial excitement of hearing about 400mpx, I listened to the video and my reaction is - if you can't get a tiff and the raw files out of it - it (Canon implementation) fails. eg Sony is better even if Canon out resolves the Sony. As a mostly canon user with an R5, I am hoping that the next firmware update provides something similar to Sony for R5 - jpgs are lossy. However even Sony implementation where you have to - shoot near perfectly on a tripod and run it through software makes it questionable as being more than a theoretical product.
My view is that both Sony and Canon methods and pixel shifting in general sucks, but Canon sucks more. I need a 100mpx camera with less bayer affects and 21 stop dynamic range, not a 400mpx pasted together shot from multiple exposures. [beyond 100mpx, is there even a lens that keeps up]
400 vs 45mpx [canon or 2.5 for sony] is in theory capable of about 3x the resolution however:
I am not seeing 3 x the resolution
There are not many lens that resolve beyond 45mpx anyway particularly for landscape
Wildlife does not sit still and capturing a moment is capturing 1 frame in a burst of many. Just getting to the right spot waiting for the grizzly is hard enough with one shot. 9 shots is going to slow down fps.
Landscape light in golden hour are constantly changing - so is water, branches, ice, and atmosphere
Typically I will return more than a dozen times to the same shot to get the perfect cloud, light, time of year.... and I change my mind as the light arises and shift my composition
Typically I have a nd filter on for landscape [reducing the resolution]
The majority of the time I am at f8 or higher [further reducing the resolution of the lens]
I usually shoot AEB to have more dynamic range and blend or HDR/blend - throwing in more processing would be challenging [an order of magnitude increase in processing]
My compositions for wildlife and landscape are constantly changing with the scene, thus putting the camera in the exact same spot for the exact same composition is not practical - for multiple compositions. I would have to shoot almost all in pixel shift and ask nature or the bear to sit still.
Shooting 9 or more shots, processing x 3 and blending would be a nightmare, not to mention the disk space.
In order for this to be useful for me, I would need the camera to save the raws it creates and create a Tiff in body. Or create a raw with the 9 pictures embedded.
I am curious. I have not visited the Sony forum for a while and my A7r2 is collecting dust - not worth selling and I still have lens in case Sony does something spectacular. Sony has had this feature for a while, has anyone had success with it in Landscape photography - eg is it practical for shooting a dynamic scene. It feels like it delivers higher resolution but it interferes with the artistic (composition) flow of the process. Is it just a studio product useful feature? Should I be jealous of Sony?...Show more →
On the a7RV the app now does the smart stuff to compensate to camera/subject movements Gary Friedman demonstrated that doing a hand-held pixel-shift bracket, and the software built the image without any issue, still a post-processing workflow on a PC but at least they are working on it
p.1 #17 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
Scott Stoness wrote:
My view is that both Sony and Canon methods and pixel shifting in general sucks...
To me, this seems like one of those things that is sort of fascinating in a conceptual way, but which is unwieldy and of extremely limited value in real world photography.
While I generally think that more MP is a fine thing as long as it doesn't raise the cost too much or depress other aspects of image quality too much, there's not a lot of point in jumping through the hoops of a process like this to get a 400MP image when we can already produce really beautiful 40" x 60" prints from existing cameras.
p.1 #18 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
I never played around too much with pixel shift on my a7R III, the one or two times I did, I didn't really feel it added much. On the a7R V though I've used it once or twice and was impressed. The a7R V is supposed to be more precise when shifting than my a7R III was.
Any of the shots that make up the 4 or 16 shot pixel shift set can be processed as a normal RAW file. So lets say you let it take 4 or 16 pixel shift shots and maybe there was change in lighting, or maybe too much movement that the software couldn't correct (which the new software tries to account for movement), you can just process one of the individual files where you liked the lighting or things were still, etc. In a way I kind of like that you get the RAW files and process them later. You can always decide whether to use the pixel shift combined version, or one of the frames that was part of the set. It doesn't take very long to shoot a set, and you can shoot conventional ones easily as you would normally.
Also, even if you feel you shoot at diffraction limited apertures (which I very often do)....you have the added penalty that you get with bayer interpolation when the RAW file is demosaic'd. With either 4 or 16 pixels shift, you're getting sampling of all 3 colors at each pixel. So you are going to get more detail than you would on a conventional shot.
Now whether you *need* or *want* that extra resolution...no argument there.
Here is a 16 shot pixel shift from the a7R V. Sharpening these large files is not something I'm used to, but hopefully you can get a feel for the results. It's a measly 240 MP I guess.
Here is 16 shot pixel shift that I will post the right 70% of the image at 100% size (see the full size link). Processing is always to someone's taste (in this case mine), but hopefully it's useful to look at.
p.1 #19 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
tsdevine wrote:
"even if you feel you shoot at diffraction limited apertures (which I very often do)....you have the added penalty that you get with bayer interpolation when the RAW file is demosaic'd. With either 4 or 16 pixels shift, you're getting sampling of all 3 colors at each pixel. So you are going to get more detail than you would on a conventional shot."
Here is 16 shot pixel shift that I will post the right 70% of the image at 100% size (see the full size link).
-Tim
Thanks Tim - for me, I find 1 image to be limiting vs AEB, in capturing colours and dynamic range. But I appreciate the sample and description. But I note what you are saying - 45mpx is not really 45mpx because of the Bayer approach. Perhaps in the future with computing power, the pixel shift can cause 45mpx to be 45mpx and turn it into a raw.
The good news for me is that the approach is still not good enough for me, that I don't feel tempted to Sony again (I moved there(a7r,a7r2) and back (5dsr) when canon was slow to match 36mpx). But I am still using on AEB/HDR blending because I find a single image still lacking for landscape vs what my eyes see. [not saying canon is better but I have some expensive long and uwa lens and am familiar with canon ]
p.1 #20 · R5 - 400 MEGAPIXELS Vs Sony A1 pixel shift
No problem, it should be more than possible. The 4 shot mode on the a7R V is 4 RAW files that are demosaiced into a single ARQ file which conceptually is like a linear DNG. So you get native resolution but with color being sampled at each pixel. No reason Canon can’t do something like that in the future.
Scott Stoness wrote:
Thanks Tim - for me, I find 1 image to be limiting vs AEB, in capturing colours and dynamic range. But I appreciate the sample and description. But I note what you are saying - 45mpx is not really 45mpx because of the Bayer approach. Perhaps in the future with computing power, the pixel shift can cause 45mpx to be 45mpx and turn it into a raw.
The good news for me is that the approach is still not good enough for me, that I don't feel tempted to Sony again (I moved there(a7r,a7r2) and back (5dsr) when canon was slow to match 36mpx). But I am still using on AEB/HDR blending because I find a single image still lacking for landscape vs what my eyes see. [not saying canon is better but I have some expensive long and uwa lens and am familiar with canon ]
