Dx mode or crop?

[bleirer]

That is a good point about the system noise. However, when using data from Photonstophotos there is another factor that comes into play. From my understanding of the explanation of his methodology the results are normalized for a standard output size. There is specific mention of correcting for circle of confusion when calculating results for testing full frame sensors in APS-C crop mode. So not sure how much of the difference in his "photo dynamic range" numbers are measured vs calculated.

I believe measured, but the starting size is normalized so viewers would view equal image size side by side.

 

[NorthernFocus]

I believe measured, but the starting size is normalized so viewers would view equal image size side by side.

Then there's math involved i.e. adjusted for the circle of confusion.

 

[Chris K]

I tried that approach, but I was surprised that my autofocus on FX mode was much better than I had expected. When I shot I was thinking I am not sure I got anything but then when I brought up the image in post I found the camera/lens did a lot better than I had thought. I was then free to work on the image.

Nice photo! I think subject detection makes AF point manipulation much less challenging. Oftentimes with bird detection on the Z9 I simply allow the camera to use the entire area, and figure out the scene itself. I'll start to use smaller areas when the scene is very challenging (like in your photo, or with BIF with tricky backgrounds), or when there are many potential subjects (ie: a pond full of geese). In that situation, "pre-cropping" with a DX viewfinder is really helpful.

 

[nmerc_photos]

No difference in IQ.


No.2 above is camera dependent. Improved AF in DX mode is unique to Nikon Z8/9. Makes no difference on Sony cameras(up through A1 at least). Not sure about Canon.

I'm not sure how it would be camera dependent.

All these crop modes do is just "delete" some of the frame. Which makes everything else in frame appear larger, relatively. Also less pixels to scan, so AF should always improve.

It certainly works this way for both Nikon and Canon. I have very little experience with Sony, but imagine it would operate the same way.

 

[ARClark]

I'm not sure how it would be camera dependent.

All these crop modes do is just "delete" some of the frame. Which makes everything else in frame appear larger, relatively. Also less pixels to scan, so AF should always improve.

It certainly works this way for both Nikon and Canon. I have very little experience with Sony, but imagine it would operate the same way.

It’s my understanding that it actually is camera dependent. As noted earlier, there is an AF advantage when shooting small/ distant subjects in DX mode with a Z8 or Z9. I don’t believe this is the case for their other ML cameras, and it was not the case for their DSLRs.

 

[bleirer]

Then there's math involved i.e. adjusted for the circle of confusion.

I think it just makes it a fair comparison. The CoC is just the size in the image where a sharp point would start to look like a blur at a certain distance by someone with a certain visual acuity. The assumption is that a fair comparison involves viewing the same size image. In my basic way I think of it like taking a 8x10 paper print and cutting out a 4x5 print from the center. If I stay viewing at the same distance then both have to be the same since they literally are the same, but if I move closer to get a better look at the 4x5 or if I enlarge the 4x5 section to 8x10 then I'll notice a difference. Same with a crop sensor. I'm not going to look at them at 2 different sizes, but I'm going to normalize the sizes then compare. The dp review studio image does the same if you click that little compare box.

 

[Nimi]

I believe measured, but the starting size is normalized so viewers would view equal image size side by side.

Then there's math involved i.e. adjusted for the circle of confusion.

On Photonstophoto, the Dynamic Range is normalized for size so comparing, for example, Z9DX to Z9FX is legit. The noise chart is NOT normalized, so putting two cameras with different sensor sizes is possible but completely meaningless. There are slight differences in sensor sizes for the A1, Z9 and R5.

 

[nmerc_photos]

It’s my understanding that it actually is camera dependent. As noted earlier, there is an AF advantage when shooting small/ distant subjects in DX mode with a Z8 or Z9. I don’t believe this is the case for their other ML cameras, and it was not the case for their DSLRs.

do you have any literature on this?

it could very well be the case, but it doesn't seem very intuitive.

we'd expect that with any AF setting where a camera is scanning for a subject, if you only have to cover 100 pixels instead of 250, you should be able to find your target faster and more accurately.

 

[NorthernFocus]

I think it just makes it a fair comparison. The CoC is just the size in the image where a sharp point would start to look like a blur at a certain distance by someone with a certain visual acuity. The assumption is that a fair comparison involves viewing the same size image. In my basic way I think of it like taking a 8x10 paper print and cutting out a 4x5 print from the center. If I stay viewing at the same distance then both have to be the same since they literally are the same, but if I move closer to get a better look at the 4x5 or if I enlarge the 4x5 section to 8x10 then I'll notice a difference. Same with a crop sensor. I'm not going to look at them at 2 different sizes, but I'm going to normalize the sizes then compare. The dp review studio image does the same if you click that little compare box.

I understand the use/purpose of the CoC. Applied in the same way when analyzing noise performance it will have the same net effect i.e. detail is enlarged on the cropped image. Which would result in the Photonstophotos data indicating DX mode performs worse than the same sensor in FX mode.

 

[Viathelens]

As to the question of which format, both, I have a toggle set on the joy stick of my Z8. But many of my wildlife shots are taken in DX mode as I can get a better fix on what is happening and I don't generally print larger than 16 x 20 which is really what pixel image size is all about. DX in Nikon is a native size of 10 x 15, so I can print in a 1:1 ratio with a DX crop. I believe the quality of an image is unaffected by an in-camera crop.

 

[NorthernFocus]

On Photonstophoto, the Dynamic Range is normalized for size so comparing, for example, Z9DX to Z9FX is legit...

The question at hand from the OP is a comparison of shooting in DX mode vs cropping the FX image in post.

From the website discussion of FX vs DX mode:

The PDR target SNR is always 16000/(height of image). This is the normalization for print size and viewing distance.

Per above the calculation apparently assumes that system noise is constant for FX/DX mode and is simply normalizing for the same size output image. So the published DX curves apply whether shooting in crop mode or cropping in post.

At any rate this is an intellectual rabbit hole. Anyone who owns the camera in question can easily do their own test comparing crop mode vs cropping in post. For practical application an eyeball test is as/more meaningful than the mathematical derivation. For my part I did my own crude testing with the D850 and was satisfied there was no notable difference in performance. I came to the same conclusion when compared against the D500 which had theoretically better PDR characteristics. In fairness I haven't run a similar test with the Z8/9 and have assumed results would be similar.

 

[NorthernFocus]

I'm not sure how it would be camera dependent.

All these crop modes do is just "delete" some of the frame. Which makes everything else in frame appear larger, relatively. Also less pixels to scan, so AF should always improve.

It certainly works this way for both Nikon and Canon. I have very little experience with Sony, but imagine it would operate the same way.

do you have any literature on this?

it could very well be the case, but it doesn't seem very intuitive.

we'd expect that with any AF setting where a camera is scanning for a subject, if you only have to cover 100 pixels instead of 250, you should be able to find your target faster and more accurately.

It depends on what information the camera is using for the AF circuitry and how/when the pixels are discarded to produce the cropped image. The first mention of this I believe was by Thom Hogan based on his testing/observations. The Z8/9 achieve blackout free/real time EVF by running a parallel data stream to the EVF rather than using the image capture data. He speculates that the difference in the Z8/9 is that AF uses the EVF data stream therefore there is less data to analyze when in DX mode. Other Nikon cameras don't have this advantage nor apparently does Sony. For a year I was shooting both Z9 and Sony A1 and did my own informal testing. There was a notable improvement in AF performance with the Z9 in DX mode and no apparent difference with the A1. That said back in that time frame the A1 performance was better overall and particularly when the subject is small in the frame. Z8/9 have improved significantly in the past year with the FW updates.

Aside from Hogan's discussion of this I've not seen any verification from Nikon nor any sort of formal testing/validation by anyone else. But I've not followed up and looked for it. My own experience satisfied my curiosity and I moved on. I believe @Steve has commented previously that he noticed improved AF/eye detection when using DX mode.

 

[arbitrage]

I'm not sure how it would be camera dependent.

All these crop modes do is just "delete" some of the frame. Which makes everything else in frame appear larger, relatively. Also less pixels to scan, so AF should always improve.

It certainly works this way for both Nikon and Canon. I have very little experience with Sony, but imagine it would operate the same way.

Funny thing is it actually doesn't work this way with Sony. IME with Canon R5/R3 and Nikon Z9, going to crop mode can increase the likelihood of the camera recognizing a bird and showing the bird eye-af box. However with Sony, I find that it is more likely to show Eye-AF when the camera is in FX mode. It can be showing the eye-af in a fairly small in the frame bird in FX but then I switch to DX and the camera just dances the green squares on the subject but doesn't activate the eye-af. This has puzzled me to no end for the last 3 years as it doesn't make much sense.

 

[Winston Churchill (Sir)]

From my experience on the weekend - I was shooting swallows in flight - and couldn't even lock on when on FX - much better in DX
And that was with the Z8 and the 800. Who would have thunk - certainly not me.
I have my Video Record button dedicated to flip the switch - it's instant - never thought it would be useful - I am eating my words.

 

[nmerc_photos]

It depends on what information the camera is using for the AF circuitry and how/when the pixels are discarded to produce the cropped image. The first mention of this I believe was by Thom Hogan based on his testing/observations. The Z8/9 achieve blackout free/real time EVF by running a parallel data stream to the EVF rather than using the image capture data. He speculates that the difference in the Z8/9 is that AF uses the EVF data stream therefore there is less data to analyze when in DX mode. Other Nikon cameras don't have this advantage nor apparently does Sony. For a year I was shooting both Z9 and Sony A1 and did my own informal testing. There was a notable improvement in AF performance with the Z9 in DX mode and no apparent difference with the A1. That said back in that time frame the A1 performance was better overall and particularly when the subject is small in the frame. Z8/9 have improved significantly in the past year with the FW updates.

Aside from Hogan's discussion of this I've not seen any verification from Nikon nor any sort of formal testing/validation by anyone else. But I've not followed up and looked for it. My own experience satisfied my curiosity and I moved on. I believe @Steve has commented previously that he noticed improved AF/eye detection when using DX mode.

Funny thing is it actually doesn't work this way with Sony. IME with Canon R5/R3 and Nikon Z9, going to crop mode can increase the likelihood of the camera recognizing a bird and showing the bird eye-af box. However with Sony, I find that it is more likely to show Eye-AF when the camera is in FX mode. It can be showing the eye-af in a fairly small in the frame bird in FX but then I switch to DX and the camera just dances the green squares on the subject but doesn't activate the eye-af. This has puzzled me to no end for the last 3 years as it doesn't make much sense.

Wild. I can't imagine why Sony would be the only one not to utilize the advantage of crop mode for AF.

 

[RichF]

Cropping in post has the exact same result as DX mode in terms of IQ and pixels.

The only value in using DX mode is following:

1. since the file size is smaller, it gives you more time before you hit an FPS buffer
2. if you are having difficulty obtaining AF on a subject, DX mode reduces the area the camera must scan, so it can increase AF accuracy/speed
3. if you know you will crop in post, it can save you time and file size

I crop often in post, but I very rarely use DX mode. I reserve it only for times when subjects are really far out and the AF is struggling.

Storage is so cheap these days, I do not mind having bigger files with more options for how to crop in post.

Risk of Dx mode is that you might clip the subject which would not happen if you cropped in post. Of course you give up better AF (in Nikon and smaller file size)

 

[RichF]

I have my Video Record button dedicated to flip the switch - it's instant - never thought it would be useful - I am eating my words.

Better add some salt and pepper

 

[bleirer]

I didn't believe it either until I did a simple test with my R5. I used a human eye since the bird wouldn't stay still. In full frame mode I counted my backward paces until I reliably lost the lock on the eye. Then I switched to crop mode and counted my backward paces until I lost the eye lock again. I forget the exact result, but I got about 1/3 more distance with the camera in crop mode than in full frame mode.

I imagine it would be easy to test it more rigorously if anyone wanted to bother.

 

[Larry S.]

Hi, When you are out of range do you change to Dx-mode on the camera or do you crop your image in postprocessing? Or does it not matter on the image quality? Have a nice day!

If you are that far out of range do you expect to have an image that will be sharp? You will likely be cropping anyway, maybe smaller than the DX format. If a critter I’m trying to nail down is really far away I touch the lens barrel function button that I have programmed for “DX” mode. Your target image is then larger and you can hold on your focus point easier…. Your left hand is already supporting the lens so programming that button makes sense…. At least it works for me… My right hand is multitasking enough..

 

[Mark-]

Hi, When you are out of range do you change to Dx-mode on the camera or do you crop your image in postprocessing? Or does it not matter on the image quality? Have a nice day!

I prefer to select DX mode if shooting wildlife and know I would crop in post as I feel making an "eye" larger helps the focus points hit. But as others have said I have no noticed a difference in IQ.

 

[Rectory Photography]

Hi, When you are out of range do you change to Dx-mode on the camera or do you crop your image in postprocessing? Or does it not matter on the image quality? Have a nice day!

There is another situation where DX mode is useful with a zoom lens. If the light is dropping, you can switch to DX mode but zoom out to where the fstop drops.....allowing a lens such as the 180-600 to drop from f6.3 to f5.6..... The extra light has enabled me to capture action shots in lower light with a higher shutter speed or lower iso than if i had zoomed the lens in to the max in FX mode. I use this quite often to good effect.

 

[bleirer]

There is another situation where DX mode is useful with a zoom lens. If the light is dropping, you can switch to DX mode but zoom out to where the fstop drops.....allowing a lens such as the 180-600 to drop from f6.3 to f5.6..... The extra light has enabled me to capture action shots in lower light with a higher shutter speed or lower iso than if i had zoomed the lens in to the max in FX mode. I use this quite often to good effect.

I'm wondering though, wouldn't the noise from cropping about equal the advantage of the wider aperture? Since you lose more than one stop in noise going from full frame to crop.

 

[Rectory Photography]

I'm wondering though, wouldn't the noise from cropping about equal the advantage of the wider aperture? Since you lose more than one stop in noise going from full frame to crop.

I have noticed times where there is no advantage as you say, but at others I just look at my iso and shutter speed and see an improvement. In those moments I can achieve my required shutter speed at my highest acceptable iso in dx mode zoomed out whereas in FX mode i can't.... i've no idea about the mathematics behind it.... I really just rely on comparing what i can achieve in the field.

 

[DRwyoming]

I have noticed times where there is no advantage as you say, but at others I just look at my iso and shutter speed and see an improvement. In those moments I can achieve my required shutter speed at my highest acceptable iso in dx mode zoomed out whereas in FX mode i can't.... i've no idea about the mathematics behind it.... I really just rely on comparing what i can achieve in the field.

If you achieve the same ISO in DX mode as you'd tolerate in FX mode you'll actually have more noise if the resulting images were displayed at the same output size. If you want to achieve roughly equal noise you'd want the camera's ISO in DX mode to be about a stop and a third lower than the ISO in FX mode.

This video explains why cropping (for the same output size image) increases visible noise. It doesn't matter whether that cropping occurs in post or in the camera:

 

[Rectory Photography]

If you achieve the same ISO in DX mode as you'd tolerate in FX mode you'll actually have more noise if the resulting images were displayed at the same output size.

This video explains why cropping (for the same output size image) increases visible noise. It doesn't matter whether that cropping occurs in post or in the camera:

Thanks for that, i'll give it a watch. What about if you achieve a lower iso in dx mode than fx? Is it probable that in every case even when achieving a lower iso, i am not gaining any advantage in DX mode?