Rumor: Sony A7s will have 4K video

Art Adams on CML on the A7s, and its codec/lowlight noise.

You don't really get extended DR from a log choice, as log only dictates
how the image is stored. If the camera can't see it then log isn't going to
help. SLog2 is good at protecting highlights as there's more bits dedicated
to the top of the curve, but that's not all that's going on here.

The shot is largely static: the camera doesn't move, and not much within
the frame moves. Noise tends to mess with compression considerably, and low
bit rate codecs tend to smooth it over a bit by turning blocky, so the best
way to hide the blocks is to keep everything still so as not to stress the
codec much. If the camera moved we'd probably see some blockiness as the
codec tried to keep up not only with the camera move but the shimmering
noise.

I've seen this kind of thing in the F55: in XAVC noise becomes almost
invisible in flat fields unless the frame moves, and then edges get really
dodgy and blockiness appears on close analysis, while SStP shows much
sharper edges and no blockiness. SStP, though, has a MUCH higher bit rate.

High ISO cameras require good codecs at high bit rates in order to be
useable. Sony traditional segments their market by dumbing down the low end
gear considerably. The camera may be phenomenal under low light but if the
codec can't handle the noise the results will be... less than perfect. I'm
curious to see whether the recording side can keep up with the camera side
at a low price point.
Click to expand...
 
IanTheM said:
Art Adams on CML on the A7s, and its codec/lowlight noise.
Click to expand...
The first point is not true, the second one is completely valid, the third one only applies if high-ISO means lots-of-noise.

* "log curve doesn't expand DR": well, it depends; if your previous curve was crushing 4 stops of the DR that comes out of the sensor (as is the case in my 5N), then switching that for a curve that doesn't do this is giving your 4 extra stops of DR. It is true that this does not increase the DR out of the sensor, but for all practical purposes it means you have 4 stops more than you had before (just an example, we still have to see how the details work for the A7s).

* "the codec is not getting a hard time in these tests": true. The codec in my 5N is really good in general, then breaks down pretty badly if you stress the camera too hard (e.g. super-fast motion, the kind you don't want to have anyway with such slow rolling shutter)

* "high ISO camera require good codecs at high bit rates to handle the noise": only applies if the footage actually has noise; what I like about the A7s is the mostly-clean ISO 12K, I don't intend to shoot ISO 400K anytime soon
 
Samuel H said:
* "high ISO camera require good codecs at high bit rates to handle the noise"
Click to expand...

Which is why I intend to record externally to either a pix240i (prores4444) or a hyperdeck shuttle (uncompressed). Prores422HQ doesn't perform very well in the dark.

Sony should sell an A7gs with a global shutter for another $1000.
 
global shutter is a really big redesign, it's a completely different pixel structure
like comparing a car and a plane: let's put wings on this and make it fly

(though, if someone can do it, that's Sony: they have a stacked-sensor tech that could allow for a back-side layer full of capacitors to store the charge of each pixel, without affecting the light-gathering side too much; but it's still really major)
 
Last edited:
I think we have to wait for some REAL tests before we trush the camera for its RS performance. Im not an expert, but those ''rolling shutter test'' videos can not be considered as serious!
I really doubt SONY would release a video-oriented mirrorless with unacceptable RS...
This is from http://www.dslrfilmnoob.com/ : A few of you asked about rolling shutter. We were able to do a little bit of fast motion testing on the Sony a7s before we left NAB on friday. The rolling shutter on the a7s is a bit worse than the Canon 6d I had with me. I spoke with the Sony rep about rolling shutter and they said that the firmware on the display cameras still needed a lot of work. They seemed to be confident that rolling shutter would be as good or better than the Canon 5d mark III by the time of the a7s release in the next 3 to 4 months.
 
lyrov said:
I spoke with the Sony rep about rolling shutter and they said that the firmware on the display cameras still needed a lot of work. They seemed to be confident that rolling shutter would be as good or better than the Canon 5d mark III by the time of the a7s release in the next 3 to 4 months.
Click to expand...

:2vrolijk_08:

Insert dodgy Sony slogan
 
lyrov said:
I think we have to wait for some REAL tests before we trush the camera for its RS performance. Im not an expert, but those ''rolling shutter test'' videos can not be considered as serious!
I really doubt SONY would release a video-oriented mirrorless with unacceptable RS...
This is from http://www.dslrfilmnoob.com/ : A few of you asked about rolling shutter. We were able to do a little bit of fast motion testing on the Sony a7s before we left NAB on friday. The rolling shutter on the a7s is a bit worse than the Canon 6d I had with me. I spoke with the Sony rep about rolling shutter and they said that the firmware on the display cameras still needed a lot of work. They seemed to be confident that rolling shutter would be as good or better than the Canon 5d mark III by the time of the a7s release in the next 3 to 4 months.
Click to expand...

As 'good' as 5DmkIII? Damning with faint praise? The 5D was always primarily a stills camera that shot video and Canon had their arm twisted to add more video features but it remained primarily a stills camera, the A7s is being marketed as a 4K video camera that also does stills. One should expect a general purpose video camera and rolling shutter of a level commensurate with general video activities.


The UK dealers have 'available' dates of 15th May, hmm, maybe that's when Sony will formally announce the price and official orders can take place. I doubt if there'll be a rush now.
 
:p


If they really can make it as good as the 5D3, that's definitely good enough in my book: at 20ms, it's not far from real video cameras like the FS100, RED ONE or C100 (all around 15-17 ms).

Awaiting to be pleasantly surprised by Sony, both by the RS and by the price of this thing (can't believe we don't have that yet).
 
Samuel H said:
The first point is not true, the second one is completely valid, the third one only applies if high-ISO means lots-of-noise.

* "log curve doesn't expand DR": well, it depends; if your previous curve was crushing 4 stops of the DR that comes out of the sensor (as is the case in my 5N), then switching that for a curve that doesn't do this is giving your 4 extra stops of DR. It is true that this does not increase the DR out of the sensor, but for all practical purposes it means you have 4 stops more than you had before (just an example, we still have to see how the details work for the A7s).

* "the codec is not getting a hard time in these tests": true. The codec in my 5N is really good in general, then breaks down pretty badly if you stress the camera too hard (e.g. super-fast motion, the kind you don't want to have anyway with such slow rolling shutter)

* "high ISO camera require good codecs at high bit rates to handle the noise": only applies if the footage actually has noise; what I like about the A7s is the mostly-clean ISO 12K, I don't intend to shoot ISO 400K anytime soon
Click to expand...


So you still intend to purchase an a7s?
 
I'm gonna be 1st in line to beg/borrow/buy/steal one just because I'm a mad bastard. For the doubters: these aren't the droids you're looking for.
 
philiplipetz said:
So you still intend to purchase an a7s?
Click to expand...

I haven't decided yet, I need a lot more info (PRICE!!), but so far it's the only one that has me interested: it seems it will fit my budget (<$3K) and tick all the boxes I care about (full frame or APS-C+Speedbooster, sharp 1080p without aliasing/moire, lots of DR, nice colors, decent codec, at least 60fps). There could still be a dealbreaker, like price or RS, but I'm optimistic by nature...
 
squig said:
I'm gonna be 1st in line to beg/borrow/buy/steal one just because I'm a mad bastard. For the doubters: these aren't the droids you're looking for.
Click to expand...

+1

The low light capabilities and a more robust codec should make for some nice 1080p footage. This has the potential to be THE one man band doc camera. I'm not concerned about the rolling shutter on a pre-prod cam being whip panned at the long end of a zoom. I doubt it'll be any worse than the 5d2/3, and people writing checks can't seem to get enough of that image. The indoor and low light possibilities have me salivating for the A7s. I'm not canceling my GH4 order either. Can't wait!
 
dcloud said:
If the 2,500 pounds is true, thats a major dealbreaker.
Price should be around a7 not twice :-/
Click to expand...

For 2500 it'd be a terrible deal compared to the A7, and preclude me from thinking of switching over to the A7s as my primary stills & video camera.
 
Many (perhaps most) of us creating video documentaries (or whatever we shoot) have more than one camera. The Sony A7s looks to be a great low-light camera that is sharp as well. We will use it for its strengths, not its weaknesses. The fact that it does not skip lines when reading the sensor is a big advantage for many shots, especially for those of us who are shooting documentaries. We plan to combine it with our FS-700, Blackmagic Pocket Camera, various DSLRs, to have a range of options. Also looking at the GH4 for use in a 3-gimbal stabilizer. Like lenses, cameras also can have different tasks. I am more interested in what it can do than what it cannot do. It would never be our only video camera.
 
You must log in or register to reply here.
Back
Top