Do you guys also adjust ISO and aperture for focal equivalency when you switch to APSC? Asking for a friend.
I usually set to max aperture on the lens and Auto ISO to never have to worry about equivalency
The simpler the better , esp. for birding for birds don't wait for you to fumble with adjustments
The reason I use APS-C mode on my A7R6 is for autofocus detection.
The closer the bird is within the frame, the better the detection works.
For example: a subject that is far away—where recognition might fail in full-frame mode—can be recognized in APS-C mode because the autofocus sees it more clearly.
It also results in a smaller file to process and saves space on the memory card, especially when I know the photo will be cropped anyway—as is the case with the majority of wildlife shots.
valmofab wrote:
The reason I use APS-C mode on my A7R6 is for autofocus detection.
The closer the bird is within the frame, the better the detection works.
For example: a subject that is far away—where recognition might fail in full-frame mode—can be recognized in APS-C mode because the autofocus sees it more clearly.
It also results in a smaller file to process and saves space on the memory card, especially when I know the photo will be cropped anyway—as is the case with the majority of wildlife shots.
That is indeed the case—it was always true with my old Canon R5 and Nikon Z8 as well. Eye detection doesn't work beyond a certain distance; switching to APS-C mode or using a longer focal length helps the camera identify the eyes.
valmofab wrote:
The reason I use APS-C mode on my A7R6 is for autofocus detection.
The closer the bird is within the frame, the better the detection works.
For example: a subject that is far away—where recognition might fail in full-frame mode—can be recognized in APS-C mode because the autofocus sees it more clearly.
It also results in a smaller file to process and saves space on the memory card, especially when I know the photo will be cropped anyway—as is the case with the majority of wildlife shots.
That would be surprising if the AF area is chosen the same. The sensor is seeing the same thing, but in APS-C the viewfinder is zoomed in so it looks larger to you. If you are using full area AF then zooming in will change the AF area.
In addition to shrinking the AF detection zone, more pixels, faster readouts, sharper lenses, and better viewing conditions helps AF.
If you must crop often, then you would benefit from a longer focal length or to get closer.
valmofab wrote:
The reason I use APS-C mode on my A7R6 is for autofocus detection.
The closer the bird is within the frame, the better the detection works.
For example: a subject that is far away—where recognition might fail in full-frame mode—can be recognized in APS-C mode because the autofocus sees it more clearly.
It also results in a smaller file to process and saves space on the memory card, especially when I know the photo will be cropped anyway—as is the case with the majority of wildlife shots.
That sounds like an urban myth to me, but if it works better for you, fair enough.
Your last paragraph is true. But cropping later in post processing gives you more freedom to choose the best composition. If you use APS-C mode while shooting, that framing is baked into the file; afterward, you can only crop tighter.
AGeoJO wrote:
That sounds like an urban myth to me, but if it works better for you, fair enough.
It shouldn't be an urban myth. As @EB-1 said above, assuming the AF area is 100% of the viewfinder, switching to APS-C would be equivalent to shooting full frame with a smaller AF area. Reducing the AF area theoretically should improve subject detection performance because the AI chip's throughput is finite. The model can analyze a fixed number of small square patches per AF cycle, reducing the AF area increases the level of detail within each patch.
In other words, the AF works on a tiny downsampled image. I asked Astra to estimate the theoretical performance of the latest BIONZ, and based on its power consumption it suggested its maximum resolution limit to be around 262K, i.e. about a quarter of a megapixel. Assuming 1:1 ratio for simplicity, this 262K image is split into a grid of 32x32 patches, and the resulting 1024 patches are analyzed in parallel in one pass. So yeah, "zooming in" by cropping should definitely improve subject detection.
In practice, this means we should be selecting the smallest possible AF area.
old-gregg wrote:
In practice, this means we should be selecting the smallest possible AF area.
One of banes of the MILC is that it needs a reasonable size to recognize something.
For example I find the XS to be worse at AF than the S in the a7rVI.
That camera also has several different AF calculation speeds depending on the compression level and bit depth, whihc could bve confounding factors.
EB-1 wrote:
That would be surprising if the AF area is chosen the same. The sensor is seeing the same thing, but in APS-C the viewfinder is zoomed in so it looks larger to you. If you are using full area AF then zooming in will change the AF area.
In addition to shrinking the AF detection zone, more pixels, faster readouts, sharper lenses, and better viewing conditions helps AF.
If you must crop often, then you would benefit from a longer focal length or to get closer.
EBH
It is very much true for all cameras that have the "AI chip".
It is very easy to replicate over and over again. Find a bird far enough away that the camera doesn't recognize it as a bird....swap into APS-C mode and it will recognize it unless it is still too small.
Cameras before the AI chip actually performed the opposite. They would recognize birds in full frame mode and then have trouble with the same bird in APS-C.
Canon MILCs also perform the same way as the newer Sonys and again easy to replicate over and over again.
arbitrage wrote:
It is very much true for all cameras that have the "AI chip".
It is very easy to replicate over and over again. Find a bird far enough away that the camera doesn't recognize it as a bird....swap into APS-C mode and it will recognize it unless it is still too small.
Cameras before the AI chip actually performed the opposite. They would recognize birds in full frame mode and then have trouble with the same bird in APS-C.
Canon MILCs also perform the same way as the newer Sonys and again easy to replicate over and over again.
Please explain this further. Assuming the same lens, say a 600, the bird covers the exact same number of pixels whether in FF mode or APSC mode. Yes the APSC view is now magnified in the finder and the bird appears larger to us, but nothing has changed at the sensor where the bird is being recorded, so how is af performance getting improved?
Jack Flesher wrote:
Please explain this further. Assuming the same lens, say a 600, the bird covers the exact same number of pixels whether in FF mode or APSC mode. Yes the APSC view is now magnified in the finder and the bird appears larger to us, but nothing has changed at the sensor where the bird is being recorded, so how is af performance getting improved?
Jack, I am not the OP but I explained that in my previous comment above. The key error in your reasoning is that nothing is "magnified in the finder". Quite the opposite: the full 66MP image gets downsampled for the finder. But it also gets downsampled for the AF model, **massively** downsampled. So it is indeed quite helpful to have the bird fill up maximum % of the AF area, this is why tricks like shrinking the AF area (which is essentially what the OP is doing by switching to APS-C) help with recognition.
Jack Flesher wrote:
Please explain this further. Assuming the same lens, say a 600, the bird covers the exact same number of pixels whether in FF mode or APSC mode. Yes the APSC view is now magnified in the finder and the bird appears larger to us, but nothing has changed at the sensor where the bird is being recorded, so how is af performance getting improved?
You'll have to ask the Sony engineers for the reason. What Gregg says sounds reasonable.
It is very easy to replicate. I deliberately make use of this "feature" every day I'm out shooting to get bird recognition working. Don't use an A7RVI to test though as that camera can find a bird's eye when it is covering 5 pixels so no need for crop mode to help it out. But if you have any other Sony camera it is very reproducible.
What I found more strange was how the older Sony cameras (like A1, A9, A9II) would recognize a bird in FF but not in APS-C
old-gregg wrote:
Jack, I am not the OP but I explained that in my previous comment above. The key error in your reasoning is that nothing is "magnified in the finder". Quite the opposite: the full 66MP image gets downsampled for the finder. But it also gets downsampled for the AF model, **massively** downsampled. So it is indeed quite helpful to have the bird fill up maximum % of the AF area, this is why tricks like shrinking the AF area (which is essentially what the OP is doing by switching to APS-C) help with recognition.
I must be missing something. If AF is on the sensor, we do not gain any AF sensor sites with a crop, nor do they change in size relative to the subject being recorded on them.
The only things changing are the VF effective magnification and now the AF processor has fewer sites to check, but only assuming we were using the full frame for focus. Even then I don’t see how that can alter AF performance to any significant fashion.
arbitrage wrote:
You'll have to ask the Sony engineers for the reason. What Gregg says sounds reasonable.
It is very easy to replicate. I deliberately make use of this "feature" every day I'm out shooting to get bird recognition working. Don't use an A7RVI to test though as that camera can find a bird's eye when it is covering 5 pixels so no need for crop mode to help it out. But if you have any other Sony camera it is very reproducible.
What I found more strange was how the older Sony cameras (like A1, A9, A9II) would recognize a bird in FF but not in APS-C...Show more →
I only have the A7R6 so I guess it’s irrelevant to me, but the logic of the claim still escapes me.
old-gregg wrote:
Jack, I am not the OP but I explained that in my previous comment above. The key error in your reasoning is that nothing is "magnified in the finder". Quite the opposite: the full 66MP image gets downsampled for the finder. But it also gets downsampled for the AF model, **massively** downsampled. So it is indeed quite helpful to have the bird fill up maximum % of the AF area, this is why tricks like shrinking the AF area (which is essentially what the OP is doing by switching to APS-C) help with recognition.
But how is the entire view relevant to the small area of the AF zone? The whole point of choosing a limited AF area (spot or zone) is to keep the AF concentrated there. I could understand that a crop mode would only scan part of the sensor so the refresh rate of the data before AF zone could be greater, though I'm not sure that would affect AF on a static target.
@EB-1 I suspect he uses "wide area AF" or whatever it's called, when the entire view is used for AF. By selecting APS-C he effectively shrinks the AF area as well.
Jack Flesher wrote:
I must be missing something. If AF is on the sensor, we do not gain any AF sensor sites with a crop, nor do they change in size relative to the subject being recorded on them.
Yes, the AF points are on the sensor, but we're talking about subject recognition which is a separate step. [1]
For the AI model to find an eye, it needs to analyze the image. The camera doesn't feed the entire image into the AI chip, it's way too big. Instead, it crops it to AF area, and then downsamples. After downsampling, it becomes a very, very small. If the subject you're trying to detect occupies a small % of this cropped image, it won't be detectable after downsampling.
[1] With subject recognition enabled, focusing happens twice: first, the camera focuses optimizing for the entire AF area, then it runs subject recognition, and if it finds a subject, it fine-tunes the focus optimizing for what's found.
