When comparing two lenses I want to ensure any differences are due to the lenses, not the way the images were processed. What is the best way to ensure consistent processing?
1. Shoot in JPG with WB, exposure, other processing like sharpness & color manually set.
2. Shoot in raw with manual exposure, then apply the same processing to both images on the PC.
And what to do about automatic lens corrections? Let them happen or turn them off? Can you even turn them off for JPG images?
Shoot with identical settings (aperture, shutter speed, ISO, focus point, lighting, etc) and in raw mode. Do not do ANY post processing on either file or do exactly the same post-processing.
Any time you shoot jpg the camera is going to make some decisions based on what image hits the sensor, and they won’t necessarily be identical for both lenses.
As to the automatic lens corrections, that is actually a more complex question than it might first appear. Obviously, if you want to see the lens’s native performance you would turn them off. But this also depends on what post-processing software you use. I am not familiar with all of the options here, but some automatically apply “lens optimization” settings, including Adobe products.
Beyond that, imagine this scenario. Let’s say that you think Lens A performs better than Lens B without optimization. But let’s say that after optimization is applied they look identical… and in both cases, better than the un-optimized files. Wouldn’t you want to just use the optimized files?
- - -
Back to add one more thing.
My answer above assumes that you are looking to do a careful comparison of the LENS performance itself, and that you are of a technical mind.
However…
If you just want to know which you prefer on your camera another approach is reasonable, too. Just shoot both lenses exaclty as you usually do — in other words, in the way you are most likely to use them — and compared that result. If that is .jpg files in automatic mode, so be it.
I would say that this approach is less technically definitive (by far) but that it could be right for YOU depending on… stuff.
Given that you can't adjust white balance in post on a JPEG with freedom, JPEG only shooting is a total nonstarter.
You'd do best to develop your own LR or FXRS presets based on your preferences and then you're down to simply adjusting WB for each shot for the most part.
gdanmitchell wrote:
... Do not do ANY post processing on either file ...
As I understand it, LR and C1 both do default processing on the raw files as they are imported, otherwise you don't even have an image to view. So I interpret your statement to be "don't do any additional processing".
darwinphoto wrote:
As I understand it, LR and C1 both do default processing on the raw files as they are imported, otherwise you don't even have an image to view. So I interpret your statement to be "don't do any additional processing".
Well... technically that is true of anything represented by digital data... including the letters I am currently typing on my computer. The letters are, of course, not stored as "letters," but as numerical encodings representing them if they are decoded appropriately.
A digital raw file is also just bunch of numerical data representing brightness levels recorded by the camera's photo sites. You cannot see those until your software converts that back to a visual representation.
If you regard that as "processing" and don't want to do it... you can't do digital photography! ;-)
So, yes, the shorthand "don't process/post-process the images" doesn't mean "inspect the binary digital data." It means work with identical (probably default) settings that are identical for both samples.
- - -
Since I have your eyes again, let me ask you a question that I raised in my post.
What are you trying to determine? Are you intending to do some kind of technical comparison of two lenses to determine which has the "best" or "preferred" optical performance? Or are you just trying to see which you like best on your camera?
in the first case, do what I suggested first — raw files, keep all camera settings identical, be extremely careful with camera stability and so forth, input the raw files with the same converter settings and so forth. Basically try to eliminate ALL variables other than the lenses themselves.
If the second case, just set up the camera the way you'll likely shoot it — even if this is .jpg files in a fully automatic mode — and make some pictures with both lenses and see which you prefer... or if you can see a meaningful difference.
darwinphoto wrote:
When comparing two lenses I want to ensure any differences are due to the lenses, not the way the images were processed. What is the best way to ensure consistent processing?
1. Shoot in JPG with WB, exposure, other processing like sharpness & color manually set.
2. Shoot in raw with manual exposure, then apply the same processing to both images on the PC.
And what to do about automatic lens corrections? Let them happen or turn them off? Can you even turn them off for JPG images?
If you are looking for simple technical analysis, don't bother, leave to Fred here to do practical tests or consult the test sites who have the expertise to measure mtf etc. (NOT YouTube warriors with opinions and prejudices but no technical competence)
If you are looking for less measurable characteristics,rendering etc, then you won't get useful for info straight out of camera with identical settings, take the shots identically (tripod etc) and set to work on tghe RAW files with your favourite sofware and see what YOU can get, choose the most pleasing. The range of adjustments is so great I bet you can get pleasing results with ANY modern lens anyway.
What are you talking about GDan? I only ever look at photos by checking the binary digital data? Personally I like Fuji's 1s and 0s better than Sony's.
Geoff D F wrote:
What are you talking about GDan? I only ever look at photos by checking the binary digital data? Personally I like Fuji's 1s and 0s better than Sony's.
Well, I was writing about what others should do.
Personally, I like to sit for hours with a glass of wine and take in the grandeur of a string of binary data. ;-)
(Speaking of processing, even those data have “processed” the original binary states into alphanumeric representations.)
Previously, I shot in raw and applied consistent processing. Then I tried to compare results from two different cameras (X-T3 and X-T5). That did not work well since the processing for each camera was different and the same scent shot with the same settings and same lens on each body looked noticeably different on the PC. I found I got closer to consistent results if I shot in jpg with each body. Closer, but still not truly consistent.
Now I'm back to doing these sorts of things on the same body and just swapping lenses.
Oh, and no way I'm going to just let Fred do it. I want to see the results for myself with my lenses, and my gear.
darwinphoto wrote:
Previously, I shot in raw and applied consistent processing. Then I tried to compare results from two different cameras (X-T3 and X-T5). That did not work well since the processing for each camera was different and the same scent shot with the same settings and same lens on each body looked noticeably different on the PC. I found I got closer to consistent results if I shot in jpg with each body. Closer, but still not truly consistent.
Now I'm back to doing these sorts of things on the same body and just swapping lenses.
Oh, and no way I'm going to just let Fred do it. I want to see the results for myself with my lenses, and my gear....Show more →
X-T3 and X-T5 have very different sensors. 3rd party programs on the PC will demosaic these files differently.
Shooting JPEG lets Fuji do that work for you, and the results from each camera should be roughly the same if Fuji is doing the processing in-camera. That's why your results were more consistent. Alternately, you can use Fuji X Raw Studio for the same look you'd get in camera, but against your RAW files. Too much work for me, personally.
As long as you get the exposure pretty much right in camera, there's no reason you can't lightly edit a JPEG, I'll do that from time to time if it's faster or easier.
darwinphoto wrote:
Previously, I shot in raw and applied consistent processing. Then I tried to compare results from two different cameras (X-T3 and X-T5). That did not work well since the processing for each camera was different and the same scent shot with the same settings and same lens on each body looked noticeably different on the PC. I found I got closer to consistent results if I shot in jpg with each body. Closer, but still not truly consistent.
Now I'm back to doing these sorts of things on the same body and just swapping lenses.
Oh, and no way I'm going to just let Fred do it. I want to see the results for myself with my lenses, and my gear....Show more →
Yeah, using two different cameras introduces even more variables — unless you are going to compare both lenses on camera A and then compare both on camera B. (If that give you a different outcome regarding the relative “quality” of the two lenses, you have discovered that it is, in the end, just about impossible to separate out small lens differences when doing actual photography, as differentiated from controlled tests analyzed carefully.)
Since in-camera jpg processing is designed (at least within a brand and camera category) to converge on relatively consistent results (but not perfectly identical), it isn’t too much of a surprise to hear that you got similar result when shooting the jpg files on the two cameras.
But since you are testing lenses, not bodies, pick one body, shoot raw, use the same settings for both lenses — all the settings, including on the camera and in post. Make sure that you can control the consistency of your shotogubject, too — preferably in good, consistent lighting and with no change in the subject. Again, the ideal would be to eliminate all variables except for the two lenses.
Aside from trying to find tiny differences between lenses (which I think is an over-rated pursuit, unless lens testing is your thing more than photography), there are other values in doing some careful tests with your lenses — not to find out which is “the best,” but just to better understand the characteristics of the lens and how settings choices and more might affect your outcome in real photography.
darwinphoto wrote:
When comparing two lenses I want to ensure any differences are due to the lenses, not the way the images were processed. What is the best way to ensure consistent processing?
1. Shoot in JPG with WB, exposure, other processing like sharpness & color manually set.
2. Shoot in raw with manual exposure, then apply the same processing to both images on the PC.
And what to do about automatic lens corrections? Let them happen or turn them off? Can you even turn them off for JPG images?
Interesting question, are you testing lenses or correction profiles.
Everyone seems to be using lens corrections, across brands, and can you even get rid of them. On the older stuff you could, but even then, there's a LOT going on in cameras and who knows.
On Canon, you might not even be able to get rid of corrections using jpegs, but apparently in 3rd party raw processors you can see the ugly truth, and the actual fl. You'd probably be better off using 3rd party like Adobe or DxO in raw not that that guarantees there aren't any lens corrections
