Re: Sony A7RII - FIRST video sample!!!

[ReelWorksMedia]

This may be off topic. Any reason why you couldn't use a speed booster on the A7RII and get full frame in S35 mode? I believe this is commonly done on the A7S. I've never had a need for a Speedbooster so I'm not really sure of any of the oddities that may come up.

 

[wallrc]

This may be off topic. Any reason why you couldn't use a speed booster on the A7RII and get full frame in S35 mode? I believe this is commonly done on the A7S. I've never had a need for a Speedbooster so I'm not really sure of any of the oddities that may come up.

You can easily do just that.

 

[ectobuilder]

This may be off topic. Any reason why you couldn't use a speed booster on the A7RII and get full frame in S35 mode? I believe this is commonly done on the A7S. I've never had a need for a Speedbooster so I'm not really sure of any of the oddities that may come up.

I'm tired of using adapters. I want Sony to once and for all to give us good native e-mount lenses (gear driven, not fly by wire) and give us good image quality with minimal rolling shutter in FF + S35 mode.

 

[Samuel H]

OTOH, I want Sony to bring out an a7000 with full-readout internal 4K, slog2, XAVC-S, and maybe IBIS, so I can sell my a7S and get two of those, which I would use with a speedbooster. I already have a nice set of manual Leica-R lenses, and using adapters has never bothered me (much less on mirrorless mounts, where I just get one adapter and leave it on the camera; with Canon I had to buy a bunch of adapters and leave one on each lens).

 

[harddrive]

This may be off topic. Any reason why you couldn't use a speed booster on the A7RII and get full frame in S35 mode?

Extremely on topic, but the hope is that a speed booster becomes unnecessary - if you want to use a full frame lens, you simply use the camera in full frame mode, so a speed booster is pointless.

BUT, such an approach is only valid if the full frame mode is as good as the s35 mode. Initially, the assumption that was jumped to was that such wouldn't be the case, as the full frame mode was using a form of pixel binning, the s35 wasn't. Normally, that would be a reasonable assumption, but as I said earlier ( http://www.dvxuser.com/V6/showthrea...ple!!!&p=1986555224&viewfull=1#post1986555224 ) then with this pixel count that may not be true. We need to wait for accurate tests.

 

[Simon Shasha]

Pretty sure the FF mode won't be as good as the Super35 mode - I only say that because Sony themselves are pushing the Super35 mode for 4K shooting. If the FF mode was superior, they would have used it for their initial launch/promo video.

 

[harddrive]

If the FF mode was superior, they would have used it for their initial launch/promo video.

Even if not superior overall, it may be comparable - a bit better in some respects, worse in others. As for why it wasn't used in the initial launch it could (if both modes were comparable for quality) be a number of reasons of a more practical nature, such as lens availability?

I'm not saying you are wrong - just keeping an open mind until critical and independent tests start to be seen.

 

[Faymus Media]

I believe it only reads half the rows and half the columns, so, a fourth of the photosites. Because that leaves you a matrix on which you can do a 1-to-1 debayer to get a 4K image. But really, no idea. All I really know is that Sony says it doesn't line skip on APS-C mode, and says nothing about FF mode. From there on it's my speculation.

@FaymusMedia:
Please tell!!!
Do you see any codec issues? How is DR? How good are the high fps modes? Really, anything you can share, I'd love to hear

The thing with slog2 is you must expose so the shadows don't clip or they will be noisey. However, with the base ISO being 800 from what I saw, there was more forgiveness. Will you be able to shoot 51,200 as clean as the A7s? Not sure but probably not. However because the base ISO was 2 stops below the A7s it has a lot less grain in the shadows. The DR is about the same as the A7s, the body feels better to hold.

As far as S35 mode. Think of it different than APS-C mode on the A7s as its not the same. The A7s is 1.6 crop, and the A7rII is 1.3x crop (i believe). No rolling shutter while shooting 24/30p. Awesome camera.

 

[Samuel H]

1.3x???? Ha!!! So much for my last 3 or 4 posts on this thread XDDD

I'm not sure I like that fact, though: it's too much of a crop for FF glass (my 24mm gives the FoV of a 31mm; that's completely different), but at the same time it's too big for APS-C glass and for FF glass with speedbooster. It really calls for S35 cine glass, which is not what I will use (too big and expensive).


(thanks for all the other comments)

 

[ReelWorksMedia]

1.3x???? Ha!!! So much for my last 3 or 4 posts on this thread XDDD

I'm not sure I like that fact, though: it's too much of a crop for FF glass (my 24mm gives the FoV of a 31mm; that's completely different), but at the same time it's too big for APS-C glass and for FF glass with speedbooster. It really calls for S35 cine glass, which is not what I will use (too big and expensive).


(thanks for all the other comments)

does that mean the Sigma 18-35 is gonna vignette like crazy?

 

[Samuel H]

Since the 18-35 is fine on FF 16:9 as a 25-35mm, I would guess that on 1.3x it would work as a 20-35mm or something like that.

 

[Faymus Media]

I shoot with the Sigma 18-35, and Canon/Nikon 24-70 on my RED Dragon which is a 1.25x crop in 6k and it does great. No worries.

 

[dia3olik]

The 18-35 would be usable 20-35, on Dragon it's useable only from 20mm

 

[DLD]

... The A7s is 1.6 crop, and the A7rII is 1.3x crop (i believe). No rolling shutter while shooting 24/30p. Awesome camera.

1.3x???? Ha!!! So much for my last 3 or 4 posts on this thread XDDD ...

s35 is a 1.4 crop off FF (unless they're doing something different with this model).

 

[harddrive]

As far as S35 mode. Think of it different than APS-C mode on the A7s as its not the same. The A7s is 1.6 crop, and the A7rII is 1.3x crop (i believe). No rolling shutter while shooting 24/30p. Awesome camera.

1.3x???? Ha!!! So much for my last 3 or 4 posts on this thread XDDD

Well, well, well, how interesting!

I'm sure you know what I'm thinking, Samuel.......

My theory was that with the 42Mp photosite count, it may be doing a direct read from 7680x4320 for 4K (2x2) and HD (4x4), and for HD in s35 mode the same with a 3x3 read off a 5760x3240 window (3x 1920x1080 on each axis).

It follows that would predict a crop factor for s35 HD of 7680/5760 relative to the FF video modes. And 7680/5760 would be ...... 1.33! From Faymus comment of "I believe" I'm reading 1.3 wasn't a precise measurement (?) - the question is whether such an accurate measurement would give 1.33x?

More and more I am wanting someone to accurately do exact crop measurements of the four video modes relative to the full frame stills mode. We know the latter is 7952 horizontally, so such measurements would give the horizontal counts for the video modes. And if the numbers do work out to 7680 for FF and 5120 for s35, I know what conclusions I'll be drawing......

And if that is the case, the implications really are huge, and would mean Sony have just raised the bar much higher as far as hybrid stills/video cameras are concerned. I'd also expect a video-centric camera using the same sensor to follow in due course. If it is the case......

 

[DLD]

But - if my PC calculator's math is correct - 5760x3240 is 18.66 MPX, not the rounded off 15, as asserted by the Sony managers. A 8.3x2 would be 16.6, two MPX fewer but still over 15. So, why 15? 8.3 + 2.07 + 2.07 +2.07 ~ 14.5, which is closer to 15 but ... implies a different Bayer mask in the s35 crop?

 

[harddrive]

But - if my PC calculator's math is correct - 5760x3240 is 18.66 MPX, not the rounded off 15, as asserted by the Sony managers.

Yes, correct. But we can't be talking about a 1.3x crop and 15Mp - the two things are mutually incompatible. So which is the true statement? The only way to be sure (and can be easily checked by anyone with the camera) is by measuring the crop factors as mentioned above.

Though...... there is one way both statements COULD be true. Which is if the 15Mp figure relates to 4K s35 mode, and the 1.3x crop relates to 1080 s35 mode (and hence, yes, 18Mp, a bigger area of sensor). This can obviously be easily checked very quickly by anyone with access to the camera. With camera locked off, and without touching lens settings, does the angle of view change between s35 HD and s35 4K modes? If "yes" (the latter a tighter angle of view) then the modes must be using different sized crops of the sensor.

A 8.3x2 would be 16.6, two MPX fewer but still over 15. So, why 15? 8.3 + 2.07 + 2.07 +2.07 ~ 14.5, which is closer to 15 but ... implies a different Bayer mask in the s35 crop?

Mmmm, but something different to a Bayer mask would then have implications with the FF modes, and the idea of those video modes being based on 7680x4320 goes away. The whole reasoning behind why a binning approach *CAN* give a high quality output is based on it needing a multiple of the frame resolution on each axis to work - so reading 2x2, 3x3, 4x4 etc blocks. Hence multiples of the total count by 4,9,16 etc respectively. The figures of 14.5 and 16.6Mp you mention correspond to 7 & 8 respectively, which wouldn't really work well. Go to 9x and it becomes possible.

But we need somebody to test this in practice. See what the crop figures are for the different modes. My prediction would be:

mode ------------------------ crop factor --- hor res

FF stills ------------------------1.0 -----------7952
FF video (HD & 4K) -----------1.035 --------7680
s35 HD ------------------------1.38 ----------5760 (NB crop factor 1.33 relative to FF video)
s35 4K -------------------------1.55 ----------5120 (approx. - assuming 14.7Mp, and crop factor then 1.5 relative to FF video)

 

[paulcurtis]

FF stills ------------------------1.0 -----------7952
FF video (HD & 4K) -----------1.035 --------7680
s35 HD ------------------------1.38 ----------5760 (NB crop factor 1.33 relative to FF video)
s35 4K -------------------------1.55 ----------5120 (approx. - assuming 14.7Mp, and crop factor then 1.5 relative to FF video)

They mean UHD really, not 4K.

So surely the crops of the S35 would be the same for both HD and UHD, just that HD is down sampled by half again.

cheers
Paul

 

[Samuel H]

The 15 Mpix thing can just be a mistake by the Sony rep. Let's wait and see exactly how much crop there is, when somebody does an actual controlled test.

 

[harddrive]

They mean UHD really, not 4K.

So surely the crops of the S35 would be the same for both HD and UHD, just that HD is down sampled by half again.

cheers
Paul

To be totally precise I suppose it's really QFHD - which means 3840x2160 - because yes, 4K is more normally used for 4096x2160 nowadays. UHD tends to be used to mean any system of more than normal HD resolution, so QFHD, 4K, even 8K.

One other thing I should clarify is that my above predictions for the crop factors I'm expecting, are all thinking of in the horizontal sense only (not diagonal, which is how crop factor is normally defined). That's to eliminate complications with still mode being 3:2, video 16:9.

But in answer to the real question, then you've got to think exactly what is happening in the scenario I'm suggesting. In s35 mode, I'm assuming numbers of an exact multiple of 1920X1080, and the only one which gives a sensible s35 type size is 3 - which equates to a crop window of 5760x3240. Assume that, and bin on a 3x3 basis of photosites - each block forming directly one pixel of the 1920x1080 image - and it makes for a very easy way of directly reading a high quality HD image.

But that number wouldn't work for 4k (OK - QFHD!! ). The blocks would have to be "1.5x1.5", which is obviously impossible. The smallest whole number it could be is 2, which for 3840x2160 means 7680x4320 photosites. But that's the size of the full frame mode! It's what I'm thinking it does for QFHD in full frame mode. It cannot do any direct read mode for any sort of s35 size window - it has to either do a form of binning which is non optimum quality, or a full deBayer and scale to get the QFHD result. Which is what is assumed to be happening.

So the question is, why then not do a deBayer/scale on the same window that (I suspect) it uses for HD - the 5760x3240? Maybe it's as simple as over 18Mp is simply too much to deBayer and scale at framerate for this processor? It can manage 15MP maybe - but not any more?

But I'm conscious I may be adding 2 and 2 up to get 7 with all this hypothesising. Just looking at the numbers and using basic principles it would explain a lot - but it is (at the moment) hypothetical. Which is why I'm so interested to see what the horizontal crop factors are in reality. If they are as the table above

mode ------------------------ crop factor --- hor res

FF stills ------------------------1.0 -----------7952
FF video (HD & 4K) -----------1.035 --------7680
s35 HD ------------------------1.38 ----------5760 (NB crop factor 1.33 relative to FF video)
s35 4K -------------------------1.55 ----------5120 (approx. - assuming 14.7Mp, and crop factor then 1.5 relative to FF video)

it's strong evidence I'm right, and it's pretty obvious what is going on. And the practical significance would be to mean all four modes would give high quality - not just s35 4K, as was initially thought. We'll see!