Video review of the Sony A7S from Philip Bloom Reviews & Tutorials on Vimeo.
Philip Bloom posted his Sony A7s video review. He really likes the A7s despite the camera having one major issue…Massive Rolling Shutter issues…
Sony A7s store links:
Sony A7s preorder in USA/CA at Amazon, Adorama, BHphoto, Sony USA, Sony Canada.
Sony A7s EU pages: Sony DE, UK, ES, FR, IT, NL, BE, AT, CH, SE, NO, FI, PL. And Wex UK.
A7s Gariz leather case in [shoplink 27569 ebay]Black (Click here)[/shoplink], [shoplink 27570 ebay]Brown (Click here)[/shoplink], [shoplink 27571 ebay]Red (Click here)[/shoplink] and [shoplink 27572 ebay]Orange (Click here)[/shoplink].
0 Responses
Nothing intelligent to add so I’ll just say first blood and call it a night! ;^)
I’m just amazed how many people I’m reading- trading the A7R in for the A7S. REALLY! IMO that is just crazy- absolutely crazy.
Not crazy if your primary photographic mode is video, the A7S by far the best at that.
Crazy? Why?
I looked closely at the A7r and passed on it. I was at NAB2013 the day it was first shown. The very second that I read that it had 8 micron photosites, I jumped into the waiting line to buy it. (especially after I heard it had SLog2)
With photosites that massive on a modern sensor design? I was sold before I saw the first test videos hit the web. I didnt even need to see a demo to know how this camera was going to perform.
Today? I friggin LOVE this camera!
Photosite size is not the whole story. There are also things called microlenses and more importantly gapless microlenses that can… close the gap… on sensors with lower fill factors.
Also, is there a report anywhere saying what the actual size of a photosite is on the a7s?
Sure…go to DXO. They have the pixel potch listed there. You can Google it too. Other sites have the numbers.
Yup…gaplless micro lenes are wonderfull. The A7s is not the first to have it.
“pixel potch” = “Pixel pitch” lol
Do you have any doubts about that a FF, 12mp sensor wouldn’t have 8+ micron photosites?
Seems you have misunderstood and also missed the point.
DxO doesn’t have the actual area of a photosite. Where do you think dxo gets their info from? Out of a magic hat?
Gapless microlenses benefit sensors with lower fill factors than one with higher fill factors… making the relationship between photosite size and noise/DR not very straightforward.
I suspect what you really want to know is what the full well capacity (pixel area is easily calculated – it’s about 64 microns). By reverse engineering from measurements of noise and DR it’s close to 157,000 electrons, which is about 2,500 electrons per micron squared. That’s pretty well exactly what you’d expect from a modern CMOS sensor.
http://www.dpreview.com/forums/post/53993974
Yes the full well capacity is the more relevant thing but how do you explain that to someone who thinks the calculated pixel pitch is the actual photosite area?
And whose measurements of noise and DR are you following for your calculations?
It’s not my calculations (although I’d estimated it would have been of that order simply by extrapolation from other cameras), I provided a link to the dpreview thread from whence this statistic came. In it the poster links to his own site.
http://legault.perso.sfr.fr/sony_a7s_measures.html
As to how these calculations are made, they come from a statistical analysis of noise figures. At high brightness levels, shot noise dominates and there’s a square root relationship between that and signal strength (and signal strength is the count of electrons). Get the SNR and you are on your way. Read noise can be calculated from measuring the dynamic ratio.
Of course it’s necessary to do a lot of measurements, but it’s fairly well established by now.
Another link for you as the A7S is now on the sensinfo site. They have a full well capacity figure of 155,000. It also explains how measurements at made.
http://sensorgen.info/
Thanks. will have a look.
Did you watch the video?
Crazy indeed. Unless you want to shoot video in dark there is no point…
The A7s video has more DR than EVERY OTHER CAMERA AT EVERY ISO — more DR means better handling of highlights, more detailed shadows and better color gradations – so no banding when shooting skies. There is nothing under $10,000 that can match the dynamic range of the A7s.
I don’t get where all the “useless unless you’re shooting in the dark” comments that always pop up with the A7s – its either trolling or pure ignorance. The 1080p out of the A7s is sharper than anything outside of GH4 4k scaled down, but the magic is the DR.
Sure, DR is important… but there’s also the matter of resolution. I really find it funny how some say the a7s is _the_ camera when not even the manufacturer doesn’t think so because there are two more bodies (a7 and a7r) that can do things the a7s can’t.
I focused on the A7s video, resolution is irrelevant. The codec, DR and picture profile gamma and color adjustments are unmatched.
If you want stills resolution there’s the A7/r, but the video out of those two leaves a lot to be desired – hence the A7s. Its nice to have choices.
IMO- it’s not like another two stops of DR negates the need for graduated filters in broad daylight. I think people have this idea they’ll never need to bracket or work around these shortcomings- not so.
Dude, do you have access to internet? I guess you do since you are posting BS here. It’s a common fact the A7s has lower DR than many modern cameras, including the A7r, especially at low ISO: http://www.dxomark.com/Cameras/Compare/Side-by-side/Sony-A7S-versus-Sony-A7___949_916. So please quit trolling and go educate yourself.
It simply depends on what you intend to do with the camera.
Well if you not doing massive crops or that your field of photography has loads of movements and ends up in areas where there’s less light source. I can see how the A7s would be really beneficial, especially that you can use an f4/f.5.6 without suffering huge consequences.
Oh the A7s is THE camera to get for your Leica or M Mount lenses, the Voigt 15mm f/4.5 works with little to no edge color cast and that lens is practically distortion free.
Ah, yes. I forgot about that. The Lieca wide angle advantage is pretty significant- can’t argue that one- it’s very apparent. I wonder how many that really applies to though?
Yea, tough to say as if you use UWA for atmosphere people related type of shots, then resolution usually is not that important, but for landscape usually they’ll want something like the A7r.
I’m just waiting for Sony to release their UWA Zoom solution, I hope it won’t disappoint.
going from the top resolution still image camera (a7r) to the low resolution still image camera with much better video capability (a7s) is like not really knowing what camera you need.
That might be the most insightful comment I’ve read on here in quite some time. Definetly something to think about.
another a7s advantage vs a7r
you can go down to f16 without difraction/coc kick in
with tha a7r you can only go to f8
in fact if you look at tiny apertures images from a7s and a7r
you could mistake the a7s as the more resolution camera
How is that possible, I mean, does the camera somehow impact diffraction ?
“You can go down to f16 without diffraction”
That’s simply ridiculous. The diffraction is identical at f16 on both cameras. It’s just that in the A7R you can see it if you pixel peek whilst the A7S doesn’t have sufficient resolution to detect it. If you are happy with the resolution you get from the A7S under such circumstances, then a down-sized A7R image will be at least as good. In terms of resolution there is no disadvantage to the A7R over the A7S at f16, it’s just that you simply aren’t using the full resolution capability of the camera.
A better way of putting it is that with the A7S there’s not resolution
gain from opening up wider than f16 whilst with the A7R there is.
I guess he did not meant it technically than visually.
But the A7S images simply won’t look better than the A7Rs at f16. It’s just that at f8 the A7R’s will start having more detail…
I am not sure – would proove by myself.
As far as my knowledge explains (please excuse if my grammar in the following is not perfect – I am no native english tongue):
Beyond f8 the abberation unsharpness turns into diffraction unsharpness which basically is a decrease in detail sharpness. This implies that higher amount of detail would show more visible diffraction unsharpness.
Maybe that was what he meant by saying “…without diffraction”.
You don’t “see” diffraction unsharpness. You just fail to see as much detail as the image becomes slightly more blurry. As that detail isn’t in the A7S photo to begin with, all that will happen is that the A7R photo will start looking more like that from the A7S.
In fact diffraction isn’t really an aberration in the sense it distorts the image. It’s just a fundamental limitation on the resolution of a lens at a given aperture.
Also, diffraction isn’t a sudden “cliff edge”. What happens is that edge contrast gradually fades.
It’s worth reading this article
http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
This bit is relevant. It actually states the camera with the smaller pixels (higher resolution) will exhibit fewer artifacts, even if the diffraction limit is passed. To an extent this probably reflects the pitch of the bayer filter pattern (which is twice as coarse as the native photosite pitch of course).
“Are smaller pixels somehow worse? Not necessarily. Just
because the diffraction limit has been reached (with large pixels) does
not necessarily mean an image is any worse than if smaller pixels had
been used (and the limit was surpassed); both scenarios still have the
same total resolution (even though the smaller pixels produce a larger
file). However, the camera with the smaller pixels will render the photo
with fewer artifacts (such as color moiré and aliasing).”
You must have lots of cool friends.
Welcome to the world of face palming 🙂
Not my first experience, and not my last I fear. Unfortunately I seem not to have the ability to just ignore such things.
There are surely a few PhD thesis to be written about the persistence and propagation of Internet myths when wholly out of context principles are generalised into items of faith.
Good video review, the Brighton Beach night scenes were truly awesome.
This is truly a beautiful video review, seriously.
Now he doesn’t quite get into the nitty-gritty depth that Kai manages does he?
As usual with Philip Bloom’s reviews, incredibly informative and putting things in context. There’s an immense amount of work in these reviews and I can only imagine how much the true cost would be if you had to pay for his time.
Anyway, I can’t help but think that some time Sony are going to release this sensor in another, somewhat large body, which would have more battery power and heat dissipation capabilities to overcome some of the issues.
I’d love to see an FE camera with a true one piece body with two grips (like a real, full-sized professional full-frame camera body a la the 1DX). I know it defeats the purpose of being small, but it would make things so much easier if they had more space to work with (and I hate how flimsy the battery grip feels on my A7R). And side-loaded batteries are easier to change in the field when you’re mounted up.
Hell, I’d love a plain old 1080p camera with a global shutter, to be honest.
He is a real pro, he shoots – among others – for the BBC.
fact: there is nothing “awesome” and will never be with Sony.
I don’t believe him. The guy which was cuddling his GH4 is far more credible!
😉
The A7s video has more DR than EVERY OTHER CAMERA AT EVERY ISO — 25600 is cleaner than the 5d3 at 6400 and the GH4 at 1600 – that’s two and four full stops. Amazing. You’re no longer forced to shoot at wide apertures just to keep the ISO down. More DR means better handling of highlights, more detailed shadows and better color gradations – so no banding when shooting skies. There is nothing under $10,000 that can match the dynamic range of the A7s.
I don’t get where all the “useless unless you’re shooting in the dark” comments that always pop up with the A7s – its either trolling or pure ignorance. The 1080p out of the A7s is sharper than anything outside of GH4 4k scaled down, but the magic is the DR. 60p is still very sharp and 720/120p looks great upscaled to all but the pixel peepers, 99.9% of the people out there will not notice a small drop in sharpness.
Add a Shogun for 4:2:2 4k and you’ll have even more DR and gain a bit of sharpness downscaled, what camera can match it in this price range? None. And it’s a pretty good stills camera as well.
I’m hoping the new big announcement is an Amount body with the A7s sensor and in camera stabilization…
Unless you are in SLog2 then you have one stop less DR than the a7r, and if you are in Slog2 then you can’t go below ISO 3200. So you are wrong in just so many ways.
There’s so many things incorrect in what you just said, I don’t even know where to start. Feel free to live in your own fantasy land.
As much as Sony tends to reuse some of their sensors in multiple bodies, I think the proper successor to the 12MP sensor is going to be a 39MP one. The latter will fit significantly higher detail in to a 4K frame because there’s no need for interpolation. It might also make the way for 8K at lower frame rates. The only issue when it comes to marketing at least is it needing an AA filter to avoid artifacts.
Do you think the next sony camera processor will be able to downsample 39mp to 4K on the processor instead of binning off the chip?
Whether on-sensor or off-sensor (in-camera) is an implementation detail… but yes… 39MP 3:2 to 4K 16:9 would give crazy detail like not many have seen before in a 4k video.
39MP only works if you exclude the DCI standard (which requires 4096). Also, both Sony and Panasonic have allowed for a “pixel border” beyond even that. Maybe it’s to allow room
Whatever the reason, I suspect if Sony ever did produce an 8K sensor (which is what 2×2:1 4K amounts to), then it will be closer to 48MP.
However, there’s a serious problem in what you suggest. It means a pixel read rate of 4x what the A7S achieves. Even if it is possible, it has huge implications for power consumption, heat and noise.
However, I’ll make a forecast now. The next FF resolution upgrade will be closer to 48MP than 39MP and video would use line skipping (to keep the read rate tolerable).
Of course you are right that a major advantage of high resolution is that AA filters become irrelevant as the required “blurring” comes from other parts of the optical system.
Of course a 39MP translates directly to 4K without the need for Bayer interpolation. This is not rocket science.
Of course. What on earth made you think I thought otherwise? However, the calculation is only true for the UHDTV version of 4K. The DCI (Digital Cinema Initiative) native version is 4096 x 2048.
You might ask why Sony chose a 12MP sensor for 4K when a 10MP one would have done. I’d suggest it is to allow for the DCI 4K standard (which requires 4096 columns) with room for some other options; maybe digital stabilisation. That may not be in the A7S, but might be used in other cameras – not necessarily Sony, but a customer. (The Panasonic GH4 supports the DCI 4K standard in addition to the UHDTV one). It’s the same thought with going for 48MP rather than 39MP (which is also quadruple the resolution of the A7S). It might be to maximise flexibility.
Also, I was also talking about the problem of the much higher read rate required for quadrupling the number of photosites read per frame (which is what will be required for the implied 4:1 pixel mapping). That, most certainly, is a very major technical challenge.
I was talking about the concept of having a 39MP sensor to get UHD by 4×4 binning. Same can be applied for what ever other formats they want to support. And if they do add more pixels (horizontally) than necessary for UHD, it would involve a slight crop to follow the same technique.
So what I said wasn’t limited to any particular resolution. It’s about the binning and having a good AA filter.
I’m not sure I understand why an AA filter is necessary. Once you get into such high resolutions, it becomes much less important.
As you say, a slight crop will be necessary, and I suspect that is what the A7S does. It would mean that HD is mapping each 2×2 bayer matrix to a single output pixel (which is what the Canon C300 does; that has a 4K capable s35 sensor with a resolution precisely matching UHDTV 4K yet is an HD output camera).
The GH4 apparently does 1:1 pixel mapping for 4K in video mode. I’d be surprised if the A7S doesn’t do so too (at the expense of a small crop which might go unnoticed). Of course 4K on the A7S does require interpolation, but it’s notable that the A7S (and GH4) produce 4:2:2 encoding, so chroma isn’t at full resolution anyway.
Yeah that’s why im more interested in canons for video because one way or another we can have full control from raw video.
Btw the AA filter does become necessary because the effective pixel pitch becomes that of a 9MP/8MP sensor although the actual is 39MP/33MP. Plenty of lenses can resolve more than “9MPs” on a FF sensor, especially in the center.
There’s no interpolation happening in the A7s. ‘Quad-HD’, which is 4x 1080p (i.e. 3840 x 2160) is used by all ‘4K’ TVs, and is only 8.3mp. Even full 4k, shot by such cameras as the RedOne, is only 9.3mp.
I understand that the Sony doesn’t do any line-skipping, so it seems to be using a 12mp sensor to generate an 8mp video. 39MP won’t make any difference if that’s the case. Certainly, it will help for 8K, but that’s a long way down the road.
Of course you need interpolation/demosaicing in the a7s for 4K… maybe they dont do it for 1080p but for 4K yes.
You need to have a read of this:
http://genotypewritings.blogspot.com.au/2012/04/more-resolution-without-going-foveon.html
I agree with sloaah. You are wrong and I don’t think you know what you are talking about.
Do you even know how cameras and sensors work?
Don’t be a waste of bandwidth. .. use a search engine, click on a link and learn something once in a while.
the ability to shoot concerts and festivals at night without low noise is pretty incredible.