Adobe Super Resolution for TIFF and JPEG?

[Jestertheclown]

So, in reality, for all your constantly harping on about Dxo and how superior it is to everything else, you've produced an image, using some Topaz gear, which you're constantly slagging off as being inferior to, yes, you've guessed it, DxO, to achieve a result that's no better; if it's better at all, than the image that you started with.

You really are a twit as well as a clown, aren't you?

Ah! Name calling again.

Your favoured tack when you can't find anything vaguely relative to the subject to post.

You're incapable of taking part in a thread; any thread, without turning it into the DxO appreciation society.

Just for once, would you please just accept that someone, other than yourself, might actually be right and that you're not.

For all your efforts to show us how much better your enlargement was to the detriment of anything that Adobe can achieve, it failed quite miserably.

My mistake was to do an enlargement at all: the OP's image turns out to be a downsize. Perhaps you can explain why Adobe's Super Resolution AI upsizer produced a downsized image?

Regardless of your mistake or whatever the OP has done with his image, the fact remains that you immediately decided that noise was a huge problem and that, as always, DxO must dash to the rescue.

I pointed out that I didn't think that noise was the issue in this case and I still say it, don't forget that I have sone experience with this software, you've only ever read dubious articles about it on line but you simply can't accept that I might have a point.

In which case, of course, DxO won't be required and that, for some reason, upsets you.

I'm used to software that does what it's supposed to,

The software does do what it's supposed to.

I have to assume that the OP's done something else.

but perhaps this is normal in your confused world.

Nothing confused in my world.

You're the one with the software fixation.


"It's good to be . . . . . . . . . Me!"

 

[Denis de Gannes]

Thanks for the comments concerning my post, this will help me with further testing. One other point I wish to make there was a very light rain the time of capture.

Ok so got a lot of info here yesterday and decided to adopt a different approach to my tests. First off I decided to give Topaz AI a try so I downloaded a trial version.

The second issue a few responders commented on the size of my uploads and the fact is I thought to upload limit to the galleries was 160 pixels on the long side.

I used ACR to work the use of the Super-Res feature a minimum of edits were used on the raw file which was saved and then sent the resulting image to PS to create the jpeg for upload to my DP Review galleries.

LrC was used to apply the same edits as ACR to the raw file and forward the tiff for the Topaz AI Sharpening at jpeg was created from Topaz to disk.

See the results, comments are welcome. First image is the LrC and the second ACR.





--
Denis de Gannes

 

[jerry1]

I am pleased with this new addition for LrC/ ACR,. I am thinking I could use this instead of purchasing a longer zoom lens for my travels to complement. After doing a few tests I am satisfied I will shelve the thought of purchasing an additional lens. See the attached files the original was created from an export from the original raw file the Crop is from the Super-Resolution DNG crop.


Jpeg from the original raw file.


Jpeg from Super-Resolution DNG with crop applied.

Comments are welcome.

These programs are good but they are not going to be a good substitute for a longer lens for photos like this. They may be great in addition to a longer lens but not a substitute imo.

 

[Reilly Diefenbach]

I am pleased with this new addition for LrC/ ACR,. I am thinking I could use this instead of purchasing a longer zoom lens for my travels to complement. After doing a few tests I am satisfied I will shelve the thought of purchasing an additional lens. See the attached files the original was created from an export from the original raw file the Crop is from the Super-Resolution DNG crop.


Jpeg from the original raw file.


Jpeg from Super-Resolution DNG with crop applied.

Comments are welcome.

These programs are good but they are not going to be a good substitute for a longer lens for photos like this. They may be great in addition to a longer lens but not a substitute imo.

After a few days of pixel peeping the new feature, I would say that Adobe Super Resolution is neck and neck with a lot of longer prime lenses and definitely better than some big zooms which are soft at distance. Which is most of them. It's all about pixels on target and it's tough to beat 4X data with 1.4X glass.

 

[Jestertheclown]

These programs are good but they are not going to be a good substitute for a longer lens for photos like this. They may be great in addition to a longer lens but not a substitute imo.

After a few days of pixel peeping the new feature, I would say that Adobe Super Resolution is neck and neck with a lot of longer prime lenses and definitely better than some big zooms which are soft at distance. Which is most of them. It's all about pixels on target and it's tough to beat 4X data with 1.4X glass.

I still think that to some extent, it depends on the content of the image.

Like Reilly, I've been experimenting with this and I'm getting mixed results although they're mostly very positive.

One thing that I did try was to enhance a shot twice; I can upload the results if anyone's interested and although the shot wasn't the most challenging; no feathers etc., it did show just how well this tool can work given the right subject.


"It's good to be . . . . . . . . . Me!"

 

[Reilly Diefenbach]

These programs are good but they are not going to be a good substitute for a longer lens for photos like this. They may be great in addition to a longer lens but not a substitute imo.

Which long telephotos did you compare with Super Res?

After a few days of pixel peeping the new feature, I would say that Adobe Super Resolution is neck and neck with a lot of longer prime lenses and definitely better than some big zooms which are soft at distance. Which is most of them. It's all about pixels on target and it's tough to beat 4X data with 1.4X glass.

I still think that to some extent, it depends on the content of the image.

Like Reilly, I've been experimenting with this and I'm getting mixed results although they're mostly very positive.

One thing that I did try was to enhance a shot twice; I can upload the results if anyone's interested and although the shot wasn't the most challenging; no feathers etc., it did show just how well this tool can work given the right subject.

"It's good to be . . . . . . . . . Me!"

 

[gardenersassistant]

These programs are good but they are not going to be a good substitute for a longer lens for photos like this. They may be great in addition to a longer lens but not a substitute imo.

Which long telephotos did you compare with Super Res?

I've been wondering exactly how to test that. There are two scenarios I can think of, both with their own problems. (Please forgive me if I have missed something obvious here.)

• Use a long telephoto. Then capture the scene with a different, shorter focal length lens, centred and focused on the same element of the scene, Then crop the wider angle shot to give the same field of view, upsize and compare the results. One problem I see with this is the differences in the optical quality of the two lenses. How to be sure that any differences, or some element of any differences, are not to do with one lens being better than the other one? Depending on the scene there may also be complications in doing comparisons away from the focused element, because the depth of field may be different for the two shots. For example if the shots both use the same f-number, won't the cropped shot have greater depth of field? And perhaps the two lenses' sweet spot apertures are different, further complicating things. Or do you adjust apertures so as to equalise depth of field, taking at least one of the lenses away from its sweet spot aperture and resolution?
• Alternatively, use the same lens for both shots. It might be used with a teleconverter for the longer focal length. Similar to the first case, will the quality of the teleconverter confuse the issue as far as any differences in image quality go? Or perhaps use a telezoom lens used at two different focal lengths. But the lens might be, for example, softer at the long end, confusing the comparison. And in either case there is a potential cropping/depth of field complication. (And in the telezoom case, its maximum aperture may be significantly smaller at the long end. More complications. And in the other case a teleconverter would reduce the maximum aperture.)

From the testing I've done, my experience has been that getting a really convincing like for like comparison can sometimes be rather tricky (or in fact impossible, requiring some compromises).

What testing method would you recommend in this case? That is a genuine question btw; I'm quite interested in trying it, but I don't know what approach would be worth spending time on. (Of course I might not have suitable equipment, That would depend on what the recommended testing method involves.)

--
Nick
Summary of photo activity and output since 2007 https://fliesandflowers.blogspot.com/2019/01/when-i-retired-in-2006-i-had-it-in-mind.html
Flickr image collections http://www.flickr.com/photos/gardenersassistant/collections/

 

[Reilly Diefenbach]

These programs are good but they are not going to be a good substitute for a longer lens for photos like this. They may be great in addition to a longer lens but not a substitute imo.

Which long telephotos did you compare with Super Res?

I've been wondering exactly how to test that. There are two scenarios I can think of, both with their own problems. (Please forgive me if I have missed something obvious here.)

• Use a long telephoto. Then capture the scene with a different, shorter focal length lens, centred and focused on the same element of the scene, Then crop the wider angle shot to give the same field of view, upsize and compare the results. One problem I see with this is the differences in the optical quality of the two lenses. How to be sure that any differences, or some element of any differences, are not to do with one lens being better than the other one? Depending on the scene there may also be complications in doing comparisons away from the focused element, because the depth of field may be different for the two shots. For example if the shots both use the same f-number, won't the cropped shot have greater depth of field? And perhaps the two lenses' sweet spot apertures are different, further complicating things. Or do you adjust apertures so as to equalise depth of field, taking at least one of the lenses away from its sweet spot aperture and resolution?
• Alternatively, use the same lens for both shots. It might be used with a teleconverter for the longer focal length. Similar to the first case, will the quality of the teleconverter confuse the issue as far as any differences in image quality go?
• Or perhaps use a telezoom lens used at two different focal lengths. But the lens might be, for example, softer at the long end, confusing the comparison. And in either case there is a potential cropping/depth of field complication. (And in the telezoom case, its maximum aperture may be significantly smaller at the long end. More complications. And in the other case a teleconverter would reduce the maximum aperture.)

From the testing I've done, my experience has been that getting a really convincing like for like comparison can sometimes be rather tricky (or in fact impossible, requiring some compromises).

For sure, you have put your finger right on the difficulties. Having shot all of the big Nikon primes as well as the Tamron and Sigma 600mm zooms, I feel I have a pretty good handle on how the focal length vs cropping equation works out, but more controlled samples would be good.

What testing method would you recommend in this case? That is a genuine question btw; I'm quite interested in trying it, but I don't know what approach would be worth spending time on. (Of course I might not have suitable equipment, That would depend on what the recommended testing method involves.)

I'm thinking my 300PF with and without 1.4 t.c. or my 70-300 at some different focal lengths. I could also do 28mm prime vs. 60 prime. What fun!

In the meantime, I'll continue to enjoy my 16512 × 9288 bird shots :^)

 

[gardenersassistant]

These programs are good but they are not going to be a good substitute for a longer lens for photos like this. They may be great in addition to a longer lens but not a substitute imo.

Which long telephotos did you compare with Super Res?

I've been wondering exactly how to test that. There are two scenarios I can think of, both with their own problems. (Please forgive me if I have missed something obvious here.)

• Use a long telephoto. Then capture the scene with a different, shorter focal length lens, centred and focused on the same element of the scene, Then crop the wider angle shot to give the same field of view, upsize and compare the results. One problem I see with this is the differences in the optical quality of the two lenses. How to be sure that any differences, or some element of any differences, are not to do with one lens being better than the other one? Depending on the scene there may also be complications in doing comparisons away from the focused element, because the depth of field may be different for the two shots. For example if the shots both use the same f-number, won't the cropped shot have greater depth of field? And perhaps the two lenses' sweet spot apertures are different, further complicating things. Or do you adjust apertures so as to equalise depth of field, taking at least one of the lenses away from its sweet spot aperture and resolution?
• Alternatively, use the same lens for both shots. It might be used with a teleconverter for the longer focal length. Similar to the first case, will the quality of the teleconverter confuse the issue as far as any differences in image quality go?
• Or perhaps use a telezoom lens used at two different focal lengths. But the lens might be, for example, softer at the long end, confusing the comparison. And in either case there is a potential cropping/depth of field complication. (And in the telezoom case, its maximum aperture may be significantly smaller at the long end. More complications. And in the other case a teleconverter would reduce the maximum aperture.)

From the testing I've done, my experience has been that getting a really convincing like for like comparison can sometimes be rather tricky (or in fact impossible, requiring some compromises).

For sure, you have put your finger right on the difficulties. Having shot all of the big Nikon primes as well as the Tamron and Sigma 600mm zooms, I feel I have a pretty good handle on how the focal length vs cropping equation works out, but more controlled samples would be good.

What testing method would you recommend in this case? That is a genuine question btw; I'm quite interested in trying it, but I don't know what approach would be worth spending time on. (Of course I might not have suitable equipment, That would depend on what the recommended testing method involves.)

I'm thinking my 300PF with and without 1.4 t.c. or my 70-300 at some different focal lengths. I could also do 28mm prime vs. 60 prime. What fun!

This afternoon, with 24 mpix Canon 70D, I did

• 55-250 at 100mm with Super Resolution then cropped vs 200mm.
• 100mm macro with Super Resolution then cropped vs 100mm macro with 2X TC
• 55-250 at 70mm with Super Resolution then cropped vs 55-250 at 250mm. (This was the first one I did, it was a mistake, but turned out to be interesting anyway)

(btw I don't have a telephoto prime. In fact the only primes I have are macro lenses.)

I need to do some other stuff. If anyone cares I can write up the method and observations later. For now I'll just post the images.

It turns out that as well as the complications I mentioned above (with depth of field differences very evident for example), there is an added complication that Super Resolution appears to change saturation/vibrance and the balance of light/dark, and possibly some other stuff too such as clarity. I tried but failed to match the tonality between the images in each pair. Also, the geometry looks different as between the two versions in each pair, with different amounts of barre/pincushion distortion. All this makes a precise "like for like" comparison an even more remote possibility than I suspected it might be.

#1 55-250 at 70mm, Super Resolution (applied just once) and then a big crop to match the field of view of #2 (This image is only 1937 pixels high compared to 3648 pixels high for #2. The pairs are the same size for the other two pairs.)



#2 55-250 at 250mm



#3 105mm macro, Super Resolution and then a crop to match the field of view of #4.



#4 105mm macro with 2X teleconverter



#5 55-250 at 100mm, Super Resolution and then a crop to match the field of view of #6



#6 55-250 at 200mm.

In the meantime, I'll continue to enjoy my 16512 × 9288 bird shots :^)

--
Nick
Summary of photo activity and output since 2007 https://fliesandflowers.blogspot.com/2019/01/when-i-retired-in-2006-i-had-it-in-mind.html
Flickr image collections http://www.flickr.com/photos/gardenersassistant/collections/