• tripod which did not move between shots
• same lens (35GM)
• iso 6400
• 1/400s
• f/5.6
• processed the same way
• deep crop to centre of frame
• equal Lightroom NR applied to both.
The test shots have equal NR applied, but the A9iii file can be pushed harder (more NR) without skin looking plastic
Here is dpreview's take on the image quality loss that occurs with the A9III's global shutter:
Thank you chiron for the link, interesting.
I wonder whether DPReview aren’t a bit too skeptical about a global shutter.
Will a mechanical shutter eventually not go the same way as the horse buggy?
We’ll see.
DPReview:
“Sony told us the a9 III has dynamic range comparable with previous models, and our measurements show that it is comparable with cameras when operating at ISO 250.“
This is correct afaik
DPReview:
“The a9 III's is noisier than the a9 II at ISO 6400, but if you try brightening the low ISO files there doesn't appear to be an additional (electronic) read noise cost lurking in the shadows.”
Does the lack of electronic read noise mean there’s more headroom to brighten images?
They did not test running Noise removal and comparing the output unfortunately. Ultimately that’s what matters the most.
DPReview:
“What might be of more concern to sports shooters is that the high ISO performance appears to be as much as one stop noisier than its full-frame rivals, especially as you reach its highest ISO settings.”
Have not tested highest iso, my iso range cap is usually 12,800.
Even on the A1 the highest few isos are kinda unusable, highest iso performance doesn’t matter as much as “highest usable iso”
DPReview:
“There's a noticeable softness to the 'grain' pattern in the a9 III's images too, which we suspect is the result of noise reduction being applied in the Raws.“
High iso NR should be turned off when shooting RAW. I don’t think DPR turned it off.
aCuria wrote:
DPReview:
“Sony told us the a9 III has dynamic range comparable with previous models, and our measurements show that it is comparable with cameras when operating at ISO 250.“
This is correct afaik
DPReview:
“The a9 III's is noisier than the a9 II at ISO 6400, but if you try brightening the low ISO files there doesn't appear to be an additional (electronic) read noise cost lurking in the shadows.”
They did not test running Noise removal and comparing the output.
Does the lack of electronic read noise mean there’s more headroom to brighten images?
There is a lot of useful insight in dpreview's report on the A9III's image quality, but here is their summary:
"Summary
Examining the a9 III's images shows everything that you'd expect from it having a reduced capacity for light. The higher base ISO isn't inherently a problem for sports shooters, so it's simply a question of whether the noise penalty is worthwhile for all the things that super-fast 120fps shooting and global shutter bring. That's something we'll consider in more depth in our final review.
But what does this trade-off mean beyond the pro sports market? Our tests show that this sensor's performance comes with an image quality hit that might make less sense for general photography. Furthermore, this cost of up to a stop of image quality in return for added performance is likely to make global shutter less appealing in the smaller APS-C and Four Thirds formats, which don't have the luxury of so much IQ to give up.
Overall, the a9 III still looks promising, for its intended purpose, but it shouldn't be assumed to herald the future of cameras as a whole."
chiron wrote:
There is a lot of useful insight in dpreview's report on the A9III's image quality, but here is their summary:
"Summary
Examining the a9 III's images shows everything that you'd expect from it having a reduced capacity for light. The higher base ISO isn't inherently a problem for sports shooters, so it's simply a question of whether the noise penalty is worthwhile for all the things that super-fast 120fps shooting and global shutter bring. That's something we'll consider in more depth in our final review.
But what does this trade-off mean beyond the pro sports market? Our tests show that this sensor's performance comes with an image quality hit that might make less sense for general photography. Furthermore, this cost of up to a stop of image quality in return for added performance is likely to make global shutter less appealing in the smaller APS-C and Four Thirds formats, which don't have the luxury of so much IQ to give up.
Overall, the a9 III still looks promising, for its intended purpose, but it shouldn't be assumed to herald the future of cameras as a whole."
Donzo98 wrote:
Would anyone use an A9III instead of an A1II for daily shooting??
I could def live with 24mp vs 50, as I’m not a big cropper.
Just curious about real world use in terms of DR and general usability.
Like many said, DR would be a big advantage for the A1II. However, I went through this process and I picked the A9. For general usability, both would do well. But the A9iii would allow you to capture more spontaneous shots, especially when you pair it with Sony lenses with 2 or more motors (I would recommend those compact G lenses). Flash use is a huge advantage as well, especially with Sony flash.
I only use an A9iii currently. I used to think I’d swap once the A1ii supply became more established, but the decision is actually more difficult than I thought it would be.
As I’ve used it, I’ve gradually found more and more utility in applying its wide range of burst rates to different subjects and situations. While I don’t use 120fps too much, 60fps is actually wildly useful, and I alternate between 30 and 60 frequently.
In a weird way, it’s almost analogous to acclimating to higher MP sensor cropping capability. Despite being 24mp, and missing the 45mp of the Canon R5, I’ve found the IQ more than adequate - even landscapes.
That said, if the A1ii was capable of 60fps, the decision would be a no brainer. But it’s 30, so I constantly go back and forth between trading in to get an A1ii, or just selling a couple lesser used lenses to nab the 300/2.8gm.
In the meantime, the A9iii is a lot of fun to use, and the 4K oversampled is pristine.
• tripod which did not move between shots
• same lens (35GM)
• iso 6400
• 1/400s
• f/5.6
• processed the same way
• deep crop to centre of frame
• equal Lightroom NR applied to both.
The test shots have equal NR applied, but the A9iii file can be pushed harder (more NR) without skin looking plastic
What happens when you downrez the a1II file to match the a9III file? Do you still retain the advantage? Does the a1II file even require aggressive NR once downrezed?
I haven't used an A1ii, but I have both the A9iii and the A7Rv. I shoot with the Rv 90% of the time, but for event shooting with flash both indoors and outdoors, I really like what the A9iii can do. Being able to strobe reliably at 15 or 30FPS has been great in the right situations. Although I've been taking the Rv for my last few trips since acquiring the A9iii I might change things up in a few weeks when I go to Italy. While the cropping capability is fantastic on the high res bodies, being able to do video with no rolling shutter has me curious about trying out the A9iii as a travel cam.
freaklikeme wrote:
What happens when you downrez the a1II file to match the a9III file? Do you still retain the advantage? Does the a1II file even require aggressive NR once downrezed?
If I understand your question correctly, you allude to pixel binning as a result of downrezing a 50MP file to 24MP, with the expectation that this should improve the SNR vs. that of the 24MP image.
Bill Claff argued that "photosite size is not a factor in noise or diffraction as it applies to our photography" because "between 6 and 13 photosites are visually combined within the CoC values we typically want." He noted "The grouping of photosites within the CoC is the visual equivalent of "pixel binning". https://www.photonstophotos.net/GeneralTopics/Sensors_&_Raw/Resolution_Doesn't_Matter.htm
From this point of view, resizing from 50 to 24 MP should have no effect on the SNR, assuming that the images are viewed from typical viewing distances, and assuming that the printed/displayed images are of typical sizes (no pixel peeping)
freaklikeme wrote:
What happens when you downrez the a1II file to match the a9III file? Do you still retain the advantage? Does the a1II file even require aggressive NR once downrezed?
I’m cropping by percentage, so if I crop out 5 MP from the A1, I would only crop 2.4 MP from the 24 MP A9 III.
The picture viewer then does the up or downres for comparison of both images at the same size on screen
Therefore I am already doing downrez or the a1 file before comparison for a large or no crop, and a upres of both files for a small crop.
However, if you take a fixed-pixel crop from the A1, for example, 5 MP from its 50 MP file and 5MP from the A9iii's 24MP file, then the A1 will indeed show more noise. This is because 5MP from the A1 image comes from a smaller area of the sensor compared to 5MP from the A9iii file.
Well.. in the never say never category - I have added an A9 III to the A1 II.
Was perusing Craigslist and connected with a nice young fellow with an A9 III. Turns out he had an A1 previously. We worked out a swap for my A1 and a little cash and we’re both happy.
Looking forward to learning how to use some of the global shutter benefits for flash and of course birds.
liggy wrote:
Well.. in the never say never category - I have added an A9 III to the A1 II.
Was perusing Craigslist and connected with a nice young fellow with an A9 III. Turns out he had an A1 previously. We worked out a swap for my A1 and a little cash and we’re both happy.
Looking forward to learning how to use some of the global shutter benefits for flash and of course birds.
k-h.a.w wrote:
Congratulations. Both are great cameras.
K-H.
Fortunate to have what seems to be the best of both worlds! Thanks - looking forward to the journey. With fall approaching - lousy weather will make for more time indoors. Time to figure out how to get the Nano trigger to synch super fast on the A9III.
I'm still of the mind that if it had to be only one that the A1 II is the way to go but we'll see.
I think the 1.4xTC is so different in performance from the 2xTC that it probably accounts for the difference in color, contrast, and certainly in resolution. It would be very interesting to see the two cameras shooting a similar scene with the same lens/TC.
chiron wrote:
I think the 1.4xTC is so different in performance from the 2xTC that it probably accounts for the difference in color, contrast, and certainly in resolution. It would be very interesting to see the two cameras shooting a similar scene with the same lens/TC.
I wouldn't read much into that. The A9iii image was 1/4000 f9 iso6400 and the A1ii was 1/2500 f5.6 iso4000. I was waiting for the kestrel to fly hence the 1/4000s shutter speed.
Won't be much of a contest with the same lens on, A1ii will win that easily, unless shooting fast action.
duncangr wrote:
I wouldn't read much into that. The A9iii image was 1/4000 f9 iso6400 and the A1ii was 1/2500 f5.6 iso4000. I was waiting for the kestrel to fly hence the 1/4000s shutter speed.
Won't be much of a contest with the same lens on, A1ii will win that easily, unless shooting fast action.
But it would be interesting to actually see what the image quality differences were, perhaps with a variety of situations or settings (different ISOs, etc.).
chiron wrote:
But it would be interesting to actually see what the image quality differences were, perhaps with a variety of situations or settings (different ISOs, etc.).
IMO the bigger noticeable difference for wildlife and landscapes comes with the A7r5 images so if your pixel peeping that's the camera to use. If your definition of IQ includes capturing the right fast action frame then A9iii is likely to have the best IQ. Having higher resolution doesn't help if you don't have the right frame to begin with.
chiron wrote:
But it would be interesting to actually see what the image quality differences were, perhaps with a variety of situations or settings (different ISOs, etc.).
But that would be hugely dependent on a scene and the severity of a photographer's exposure error. A correctly exposed slide film tells us that even 6-7 stops is often enough for a fantastic photo. In fact, printed images are also confied within 7 stops due to the contrast limits of paper. Operating within 7EV is perfectly possible, but it requires good light and perfect metering.
Digital cameras with wide dynamic range allow for two things:
1. Photograph scenes that used to be prohibitive.
2. Be sloppy with our exposure. (easy to solve with essentially problem-free bracketing in high-FPS bodies)
Which means that comparing A1 and A9 is going to be inevitably subjective endeavor with questionable conclusions. I am currently renting a Z8 body and I keep trying to see the measured DR difference in real world photographs, and I keep failing to find a natural scene which allows me to see that extra 1EV that my Sony body supposedly offers. I am sure that if I buy high-granularity step wedges and run the full spectrum of flash power steps I'll see it. But that would be almost cheating. What's the point of XYZ if it can only be seen in a lab?
