New Sony Alpha Camera | January 26th

I downloaded an A1 sample from Gordon Laing on Vimeo, 8K 24p, 10b 4:2:0 S-Log3, 400 Mbps, and graded it to HDR10 in Resolve, played on 8K QLED, it honestly made my P6K upscaled to 8K look soft by comparison but it was still difficult to see much if any change from a 4K playback of same 8K file on same tv, even from 1 foot where I could scrutinize the pixels.

I would note, I could break XAVC apart too easily for HDR. I tried to place the sun in Gordon’s clip at 2000 nit but it was breaking apart by 1200 nit. It didn’t appear the sun (diffuse) was majorly clipped if at all on the luma exposure waveform but data in the blue channel was rolling off starting about 10b code value 775, approximately 1200 nits. As I continued to raise it, the highlight swung grossly to yellow which is an overdone characteristic I see in a lot of YouTube HDR videos, presumably for artistic effect as yellow rays of warmth, so I let it ride, but it is fundamentally still just channel clipping. If you kept going the green channel would clip too , leaving you with a mosaic red sun spilling into nearby areas as well. The footage was quite busy (euphemism for noisy) on patterned areas, smooth textures and gradients alive with swarming winds.

When I compare BRAW footage from P6K, it would be unfair say anything less than BRAW killing it, denser, solid richness to the gradients, minus the busyness and swarming winds of dust, also lacking the fine crisp details. I’m not sure how much of this is sharpening, edge enhancement but I did notice black and white outlines to edges, but they are very small at 8K. I can push highlights like the sun in BRAW to a much greater degree, but these are HDR points. Straight from the camera of normal contrasts the A1 is most certainly great, but Gordon’s beach shot was a torture test of ordinary flat tones shot into a flat S-Log3 gamma. You have to do something with it. And my result was very nice looking, just couldn’t push into the extremes with it for HDR.

I’m undecided on the A1, trying too hard to fall in love perhaps. If not for the breathtaking price, an easy win for its AF and stabilization but no matter what imatest said about its great DR measurement, the small pixels are noisy in SLog3 and it begs for raw, which to me is an expensive royal pita to franken-rig. It means that P6K with internal BRAW and a stabilized EF/EFS lens kills it even without AF when price is factored, certainly for HDR and maybe not at all for straight from the camera output.
 
I downloaded an A1 sample from Gordon Laing on Vimeo, 8K 24p, 10b 4:2:0 S-Log3, 400 Mbps, and graded it to HDR10 in Resolve, played on 8K QLED, it honestly made my P6K upscaled to 8K look soft by comparison but it was still difficult to see much if any change from a 4K playback of same 8K file on same tv, even from 1 foot where I could scrutinize the pixels.

I would note, I could break XAVC apart too easily for HDR. I tried to place the sun in Gordon’s clip at 2000 nit but it was breaking apart by 1200 nit. It didn’t appear the sun (diffuse) was majorly clipped if at all on the luma exposure waveform but data in the blue channel was rolling off starting about 10b code value 775, approximately 1200 nits. As I continued to raise it, the highlight swung grossly to yellow which is an overdone characteristic I see in a lot of YouTube HDR videos, presumably for artistic effect as yellow rays of warmth, so I let it ride, but it is fundamentally still just channel clipping. If you kept going the green channel would clip too , leaving you with a mosaic red sun spilling into nearby areas as well. The footage was quite busy (euphemism for noisy) on patterned areas, smooth textures and gradients alive with swarming winds.

When I compare BRAW footage from P6K, it would be unfair say anything less than BRAW killing it, denser, solid richness to the gradients, minus the busyness and swarming winds of dust, also lacking the fine crisp details. I’m not sure how much of this is sharpening, edge enhancement but I did notice black and white outlines to edges, but they are very small at 8K. I can push highlights like the sun in BRAW to a much greater degree, but these are HDR points. Straight from the camera of normal contrasts the A1 is most certainly great, but Gordon’s beach shot was a torture test of ordinary flat tones shot into a flat S-Log3 gamma. You have to do something with it. And my result was very nice looking, just couldn’t push into the extremes with it for HDR.

I’m undecided on the A1, trying too hard to fall in love perhaps. If not for the breathtaking price, an easy win for its AF and stabilization but no matter what imatest said about its great DR measurement, the small pixels are noisy in SLog3 and it begs for raw, which to me is an expensive royal pita to franken-rig. It means that P6K with internal BRAW and a stabilized EF/EFS lens kills it even without AF when price is factored, certainly for HDR and maybe not at all for straight from the camera output.

Very useful info Tom. I agree with you, for bang for the buck, the Pocket 6K seems to pretty much be in a league of it's own. This also confirms for me that from a practical standpoint, 8K seems to be wasted in 2021.
 
Yes, great post Tom. Yes 8k seems like overkill but the photo world wants the high MPX cameras, so 8k is really coming along for the ride with hybrids. The A1 proves that the processing power is up to the task these days. I stopped by a camera store and handled an A7sIII the other day and I just can't get past Sony's ergonomics for photo use. I put the camera up to my left eye, so my nose covers or is too close to the AF-ON button. The next button over is kind of on the edge, so a button remap results in an awkward angle for your thumb. Also handled an R6 and everything was comfortable as usual with Canon.

Sony is killing it with the tech but sadly the design takes it out of the running for this hybrid user. Felt great for video use though.
 
Thanks Guys! I didn’t intend to make the Pocket 6K into a competitor per se. The important takeaway is the comparison of 400 Mbps XAVC Slog3 4:2:0 for 8K versus BRAW as it applied to what I saw grading Gordon’s clip in HDR10. You can make beautiful HDR with the A1, and you can see the extra resolution but I had to look very hard and from very close. I’m not sure at all that I could distinguish 8K from 4K if I didn’t already know which was which, but it was sharper than my Pocket 6K, but what role was played by edge enhancement I can only speculate.

But one point of certainty, not in dispute in my observation, is the disadvantage XAVC puts you in for 8K HDR, a place where the A1’s extra dynamic range is mitigated. It will take the external raw to unleash its full potential. That was the point for the Pocket 6K, internal BRAW.

And so having great dynamic range from the sensor is a feature of the A1, but XAVC hobbled me in manipulation of the highlights. I still got very nice images overall, but with that sensor, the potential is there to make HDR really pop when unleashed, and BRAW is how the Pocket 6K is able to punch above its class.
 
I am curious - without the 8K upscaling (or, to be more accurate, the upscaling into 8K), is the difference more pronounced? In other words is the image from P6K benefiting disproportionately from the latest TV technology?

Side note - from the reviews, the camera does overheat from the continuous 8K shooting but is capable of resetting itself in a few minutes ... unlike the early iteration of R5. Which is more of a video-first/portraits-second hybrid and a better buy overall, unless one needs an action/sports/BiF cam.
 
I am curious - without the 8K upscaling (or, to be more accurate, the upscaling into 8K), is the difference more pronounced? In other words is the image from P6K benefiting disproportionately from the latest TV technology?

I don't think so. First of all, Gordon's A1 footage just looks sharp. You might extend that to "sharpened." Or not. There are additional factors in the perception, lens, focal length, dof, shutter speed. In his clip he was shooting wide, at 1/200th shutter. I didn't get the same sharpened impression from the official Sony footage. For me usually, 6K is down-sampled to 4k in Resolve for output, upscaled to 8k by the tv. As an experiment, I upscaled 6K to 8K in Resolve and fed that to 8K tv. Between the two, I can't see any difference, at least so far I haven't been able to. On the 8K tv, I can see (barely) the pixel structure from almost 3 feet out, (65 inch set). It's not a good way to watch in any reasonable sense from that close. I sit a more normal 6 ft from the screen in my small studio, nearly perfect distance. No difference 4k-6K-8k from that distance. 8K tv sets present an opportunity to get very close or an opportunity to have a very large screen.

I adamantly maintain that 8k camera footage stresses XAVC 400 Mbps for strong HDR color grading. For motion, it holds up very well. Very helpful of Gordon to make available the native footage for that analysis. BRAW kills it. Fortunately, there are plenty of reasons to recommend Alpha-1 besides just 8k video mode.

The XAVC 8k ingest into Resolve stresses real time playback and fast scrubbing, but is easy on rendering. The BRAW 6k ingest is super easy with fast real time playback and scrubbing, super fast 4k rendering, stresses on 8k rendering output.
 
I am curious - without the 8K upscaling (or, to be more accurate, the upscaling into 8K), is the difference more pronounced? In other words is the image from P6K benefiting disproportionately from the latest TV technology?

The riddle is this; why does my Pocket 6K look softer upscaled to 8k than a 4k presentation of the 8k original from the Alpha 1 when it is upscaled to 8k on the same tv?

It’s the same answer for why the BM 12K output isn’t sharper than 8K from the Canon R5. In both cases, the smaller BM sensor is out resolving the best lenses. The best lenses have resolving power of about 83 line pairs per millimeter, the 6k sensor is resolving 133. The lens can’t keep up with microscopic demands of tiny pixels. In the case of the A1, the 8k sensor is resolving just 90 lp/mm. It’s a hopeless case for the 6K, 12k and MFT of an 8k GH5. The 12k sensor is resolving a whopping 227 lp/mm.

It’s the net resolution of the sensor and lens as a system that matters.
 
Last edited:
The best lenses have resolving power of about 83 line pairs per millimeter, the 6k sensor is resolving 133. The lens can’t keep up with microscopic demands of tiny pixels. In the case of the A1, the 8k sensor is resolving just 90 lp/mm. It’s a hopeless case for the 6K, 12k and MFT of an 8k GH5. The 12k sensor is resolving a whopping 227 lp/mm.

How soon do we get into "who cares?" territory?

I mean, sure, there are different applications. There was a photographer who shot focus stacked and panorama stitched macros of coins. Something like hundreds or thousands of exposures per coin. To render all the little details. I think Alec Baldwin bought a print. When detail is the purpose of the shot, then I guess that the sky is the limit. But for me, personally, affect is usually the purpose of a shot and having too much detail can be a bad thing.
 
This “problem” of having a sensor that out resolves the lens has some benefits as well. A single ray of light, the smallest that a lens can convey, spills onto more than 1 pixel of the sensor. This results in a denser pattern, thicker paint film. As you move the spray gun away from the canvas, the pattern enlarges, the applied droplets become finer, the space between droplets enlarges, the image becomes more grainy, more noisy. Move the gun closer for more color saturation and density, which is a characteristic of the Ursa 12k. As we increase the number of pixels we only gain resolution up to the limit of the lens. When the lens can no longer support the density of the sensor, if we want to keep increasing resolution we have to increase the size of the pixels.
 
... It’s the net resolution of the sensor and lens as a system that matters.

On the other hand, most reviews assert that there's discernible difference when shooting in a pixel shifting mode with S1/S1R and A7RIII/IV. And that gets you into 187 MPX on S1R and 240-something on 7R4.
 
On the other hand, most reviews assert that there's discernible difference when shooting in a pixel shifting mode with S1/S1R and A7RIII/IV. And that gets you into 187 MPX on S1R and 240-something on 7R4.

This has been mostly folly. There are lenses that resolve 200 line pairs/millimeter @ MTF10. The 83 lp/mm used for a reference is @ MTF50. MTF is not arbitrary, but it is fungible. In other words, there will be more details visible to the sensor by lowering the threshold of what constitutes a visible detail.

That said, for any given MTF or lp/mm, the bigger sensor puts less stress on the lens. I know that is obvious to everyone, until you start trying to put numbers to it. Thanks for letting me get my head around it.
 
Last edited:
This video has a bit of an interesting discovery on how the Sony Alpha cameras (at least the most recently released versions) handle their HDMI output based what the internal resolutions settings are. It looks like the Sony A1 is downsampling the 8K sensor for the 4K HDMI output which is something that is does NOT do internally.

Skip to 10:28 min:sec
https://www.youtube.com/watch?v=MJfqW7_mvlg
 
I think that is how many cameras work. They are like computers that function a certain way based on their programming. (Ironically it's an annoyance with some Canon cameras as the output settings are forced to match the internal resolution, which some don't want.)

So when the A1 is set to its highest-quality, it is outputting the highest-quality image from that juicy high-resolution sensor in each lower mode (4K and HD).

You would think it would be doing that at all times no matter if you're in 4K or 1080p, but you have to manually set it to its highest-quality internal setting.
 
Back
Top