Rumor: Sony A7s will have 4K video

I do have a FS100 and find the RS performance acceptable and would not use a camera with a worse RS skew.

Long life to rolling shutter! Imagine now this made with a global shutter... : D

lartiguecar.jpeg

Jacques Henri Lartigue [ LINK ]
Car Trip, Papa at 80 kilometers an hour, 1913
aka Grand Prix de l'A.C.F Automobile Delage - 26 juin 1912
 
I do have a FS100 and find the RS performance acceptable and would not use a camera with a worse RS skew. I do rig my camera and shoot on supports so I'm not a run n gun merchant, I was a professional Camera Operator.

The point about RS is not just shooting hand held it's also a question of shooting life and life has a habit of throwing fast moving subjects at the lens, it's 2014 overt jello-vision should be consigned to history and I don't want to see skews on anything other than whip pans and certainly I don't want to see skews on moving objects and vehicles unless they're formula 1 cars.

EOSHD comments to the size of the sensor as being cause of RS, not sure if I believe that. My understanding is that the boat shots were all shot in HD (correct me if I'm wrong) which uses the crop section so may show less skew than full sensor 4K. If Sony has gone backwards wrt to RS skew they won't be able to sell me this camera at any price.

The price the dealers have been told to work with is £2500 inc VAT in EU and ME regions that's a fact I have spoken to one directly myself. I think Sony were waiting for after NAB to set the price in stone save for surprises and unfortunately the A7s will be premium priced even if they take a hint that £2500 is too steep.

It's probably not the camera I was hoping it to be but that doesn't make it a bad camera or one that others will find great for their needs.

Bigger sensors, and sensors with more pixels, are slower to read, for a same basic design. This one is big and does a full read (skipping half the lines cuts the read-out time in half), but has few pixels. So: wait and see. We'll probably know before the camera hits the shelves.

The camera allows 4K full sensor, 4K APS-C (weird because it's an upscale from 2.7K), HD full sensor, and HD APS-C. AFAIK, the Scotland video was shot in internal HD using the full sensor.



edit:
Sigma just announced the price for the 50mm f/1.4 Art: 850 GBP, 900 EUR, and 950 USD
http://www.dvxuser.com/V6/showthread.php?321965-Rumor-Sony-A7s-will-have-4K-video/page45
With those implicit exchange rates, 2500 GBP is 2650 EUR and 2800 USD
 
Last edited:
EOSHD comments to the size of the sensor as being cause of RS, not sure if I believe that. My understanding is that the boat shots were all shot in HD (correct me if I'm wrong) which uses the crop section so may show less skew than full sensor 4K. If Sony has gone backwards wrt to RS skew they won't be able to sell me this camera at any price.

I'm with you there, the 5D MK3 rolling shutter is as much as I can tolerate but I don't expect the A7s will be any worse, but if it's on the level of the 5D that's quite a bit more than the FS100. I wasn't implying you're not a professional, I was just making the point that DSLR's don't perform well handheld. I've left a msg for Den asking how the camera was rigged on the boat.
 
I'm completely perplexed about the amount of "hate" the 5dmarkIII is usually getting. The image is absolutely breathtaking when shot the Magic Lantern RAW. The amount of colors and latitude and dare I say, even the DR, is really highend. And it is astonishing how well the lowlight RAW stuff gets cleaned up with neatvideo. Way better than any compressed material. I had to make a decision about shooting a bigger shortfilm with either the 5dmarkIII (which meant purchasing a bunch of 1000x cards) or renting a c300/fs700. I'm so glad I went the 5d way. Fullframe is just something else than even s35. What people kinda miss is that anamorphic s35 has more surface area than fullframe.

All that said. The A7s looks amazing. If the rolling shutter is ok, I'm soooo getting one.
 
No hate for the 5D3+RAW from me. The H.264 video is soft and has much worse DR, but with the hack it's probably the best camera I could afford today. It's just not compelling enough for me right now. If the A7s doesn't beat it handily for DR, I'd rather get the 5D3+RAW than the A7s. But it seems it will beat it by quite a margin.
 
I'm completely perplexed about the amount of "hate" the 5dmarkIII is usually getting. The image is absolutely breathtaking when shot the Magic Lantern RAW. The amount of colors and latitude and dare I say, even the DR, is really highend. And it is astonishing how well the lowlight RAW stuff gets cleaned up with neatvideo. Way better than any compressed material. I had to make a decision about shooting a bigger shortfilm with either the 5dmarkIII (which meant purchasing a bunch of 1000x cards) or renting a c300/fs700. I'm so glad I went the 5d way. Fullframe is just something else than even s35. What people kinda miss is that anamorphic s35 has more surface area than fullframe.

All that said. The A7s looks amazing. If the rolling shutter is ok, I'm soooo getting one.

+1 this, except I don't care about the FF aesthetic.
The appeal of the A7s for me is 4k shot for 2k. I am sure the camera will be at least a stop better than 5dmk3 in low light and DR, possibly more. On the other hand, as you say high iso ML raw cleans beautifully in NV. Add to this the hassle of LUTs for monitoring s-log (Atomos don't have monitoring LUTs AFAIK?), the lack of the excellent ML exposure and focus goodies which I KNOW I will miss, the 5d allure of raw, and the fact that at 700+ mbps 4K prores is practically the same bitrate as 1080p ML raw. Now take into account that rental prices for 5dmk3 are pennies a day these days. Consider the price rumors (yes, A7s is a specialty camera, but still - going higher than A7r just based on specialty? Dunno, doesn't sound right.). And suddenly A7s doesn't look THAT good. :)
 
Bigger sensors, and sensors with more pixels, are slower to read, for a same basic design.

Yeah, I get that but this sensor is bigger in physical size but is only a 12MP part and one would've thought that readout would've been effectively no worse than the FS700 for example which is a 11.6MP part. The comment by EOSHD regarding RS would likely indicate the A7s is indeed worse than a Sony sensor with a comparable pixel density.
 
* Both things matter: total size, and pixel count. The a7s should be slow because it's a big sensor, but should be fast because it has few pixels. Which one is more important, we'll know soon enough.
* Architecture matters a lot more than any of those. You can make an awfully slow small-and-lowres sensor, and a really fast huge-and-highres sensor. How much each costs, and what you win or lose in other areas, that's another story.
* The a7r doesn't get horrible RS in video because it skips lines. The a7s won't.
 
* Both things matter: total size, and pixel count. The a7s should be slow because it's a big sensor, but should be fast because it has few pixels. Which one is more important, we'll know soon enough.
* Architecture matters a lot more than any of those. You can make an awfully slow small-and-lowres sensor, and a really fast huge-and-highres sensor. How much each costs, and what you win or lose in other areas, that's another story.

I'm not sold on the idea a marginally larger sensor with a similar number of pixels should be much different in RS performance.

That's what I'm trying to understand is why Sony appear to have produced a camera for the shoot by candle light brigade yet produced a sensor with troubling levels of RS (if the reports are to be believed).

I don't know enough about sensors to know what are the limiting factors and the ones which necessitate a slower readout and hence increase rolling shutter but I would've thought a video camera would be best with as little RS as far as possible at the cost of light sensitivity if they are both linked. Maybe the bigger photo sites holding more charge takes longer to drain and that reduces the readout time, I dunno, I'm guessing.
 
Yeah, I get that but this sensor is bigger in physical size but is only a 12MP part and one would've thought that readout would've been effectively no worse than the FS700 for example which is a 11.6MP part.
It's 12 Mpixel - but that's for the whole sensor as used for stills - which are 3:2 aspect ratio.

Just do a simple 16:9 crop (losing top and bottom of the 3:2 frame) and that then becomes more like 10Mpixel.

In fact, I strongly suspect that when they say they are "reading the whole sensor in video mode", they are reading 3840x2160 - hence the thought of the 1.1 crop. Be very easy to confirm - does the horizontal angle of view tighten slightly between stills mode and video. (I suspect it does.)

In which case, same number of photosites/frame as the FS700. I won't speculate on rolling shutter until I've seen some tests that I'd consider reliable.
 
oh, my...

I don't get it: Den Lennie said it was not a problem
The only way to reconcile RS being fine for him and awful for others is that he shot full-frame 1080p internally, then external 4K with static shots. The sensor could have high-quality in-sensor binning, that allows for a much faster read-out without aliasing/moire artifacts when you read it at a lower resolution (it's not common, but for example the CMV12000 has this).
 
oh, my...

I don't get it: Den Lennie said it was not a problem
The only way to reconcile RS being fine for him and awful for others is that he shot full-frame 1080p internally, then external 4K with static shots. The sensor could have high-quality in-sensor binning, that allows for a much faster read-out without aliasing/moire artifacts when you read it at a lower resolution (it's not common, but for example the CMV12000 has this).

Den also used lenses that have Sony's OS technology - Sony have some of, if not the best, lens OS on the market.

Regardless, that rolling-shutter is horrific. I guess you gotta take into account that the camera they were filming the A7s with probably suffers from rolling-shutter too, however, even then, its probably not a major factor/inconsequential.
 
I guess you gotta take into account that the camera they were filming the A7s with probably suffers from rolling-shutter too, however, even then, its probably not a major factor/inconsequential.

In this case, the point by macgregor a few pages back is 100% valid. If you stop the video, you will see some ghosting in the front LCD, which is jumping between two grames; if the back camera has rolling shutter, the top part of the image will show more of the first frame, the bottom more of the second frame, but the skew is only determined by the A7s. And it's bad.
 
Back
Top