You can’t see enough of those A7s videos? Here comes the next one from Den Lennie! It shows how the Sony A7s video quality easily outperforms the Canon 5D MK III in Low Light conditions.
Only 18 days until the A7s ships at Sony Store US (June 30) and Amazon US (July 1). All A7s preorder links here:
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.
0 Responses
what a surprise. A7S is coming 2 years after 5Dmkiii….
2 years doesn’t really matter when Canon has been using basically the same sensors on all their cameras for the last 4 years…
Las Vegas with my lowly D-Lux 4 was perfect already … at ISO 80 … as it looks a bit like daytime !
Comparing this at high ISO, dunno, what for ?
Yeah but don’t forget its only 12MP. It looks nice but if you want to take a closer look at something you can’t. You are limited in zooming in on the picture. As soon as you pass the 12MP resolution when zooming it will begin to pixelate. :'(
If we are talking about still photography it’s true, 12Mpx gives less option for cropping and big prints (not “zooming”, though).
If we talk about video quality, there’s no difference as long as the maximum video resolution is the same and this is indeed the case – except A7s supports 4K via external recorder, and here the A7s looks quite better indeed.
That’s what you have interchangeable lenses for.
We can have both.
with 12 mp vs. 36 mp you need lens with 73% more FL.
I’m talking about the picture not through the camera. You miss understood me. When you zoom in(or crop) on an image on you laptop or whatever device(after you have taken the shot).
try to compare it when you have 10 more megapixels…
the fact is: sony can’t even get close to Canon.
Lame excuse; the difference is the underlying technology, not the number of megapixels. The image on the A7s is sharper all through the ISO range, beating the 7DMIII even at base ISO.
🙂
Correction; that’s 5DMIII
its compare the video quality, nothing about pixels, in fact, a7s do full pixel read out on video is outperform the 5D3 crap video quality.
while the G1XII only 12mp then compare to RX100III 20mp??
i’m comparing nothing. nice nickname though 😉
Its video man! And far superior to mkiii.
Canon EOS 1Dx versus Sony A6000 would be a much closer fight. Sad how stagnant Canon has become these days. Of course, they’ll still sell like crazy because all those Canon users out there don’t want to replace their lens collections. (That, and they probably like Canon’s overly-exaggerated color.)
we call it “gg”
Do you think Canon shooters don’t switch to Sony/Nikon because of the lens collection? They don’t do that because they need to make sure the customer support is good and fast, they can get loan equipment if they want, etc. Can Sony/Nikon offer that?
And yet admin fails to mention that 99.9 percent of wedding and portrait photographers and video professionals shoot with Canon and Nikon cameras not Sony gadgets with inferior flash system.
I love my Sony gear but I hate the bashing admin and others do on other camera gear.
The titles hear on SAR make ne laugh. Meanwhile the pro’s are out making a living while others here reference test charts on websites.
Pathetic
Let the burning begin
The usual refrain, claim you are a fan, then bash on it. If you hate bashing why do it?
Btw, pros aren’t much of a reference.
There are plenty of wedding photographers using Sony cameras, and their excellent Zeiss lenses, and more are converting to the A7r, A7s, A77II, and A99 daily.
You’ll need to check your numbers because Sony now controls over 15% of the interchangable lens market with both A, and E-mounts, and unlike Canon, or Nikon, Sony’s market share is growing with both mounts.
Canon is a dinosaur and your obviously living in the past, cause the future is SLT, and mirror-less.
Hey Grandpa,
I shoot weddings with a A77, A99, and A7r with a bag full of G-Series, and Zeiss lens. In the industry most wedding photographers are moving away from the old camera systems and embracing 21st Century technology like the A7s, and A99.
It seems the only the old photographers (50+) are interested in DSLRs these days. The rest of us are using modern mirror-less cameras, or SLT (which is closer to a mirror-less camera than a DSLR).
I shoot weddings with a99- check em out http://www.limelight-films.com
I’ll hop on, another pro photog using sony gear (well actually, whatever gear suits the job. nikon, canon, sony, hasselblad, olympus, fuji, they’re all just tools to a good photographer) – quite honestly if you can’t use sony cameras to make your images, you better blame something else besides the flash system mate. I use the A99s and A900s + Zeiss and G glass for weddings, portraiture, commercial and assignment work, when getting published they don’t ask you what you shoot p.s.
Get your head out of your arse.
http://www.frankwithersphoto.com
Actually, my colleagues and I were discussing today how Andrea bashes Sony a considerable amount for a Sony forum creator. I don’t see him as a fanboy by any means.
well, even the A7 Sensor trumps the 5D3 easily (i have both) – so no surpise here.
i wonder how big the advantage for still pictures will be – keeping in mind you can always downscale the images from higher megapixel sensors to 12 mpixel and get less visible noise as a result…so i guess the a7s will need to be 2 steps better to make a significant better picture at default print sizes.
Not at all… noise is generally proportional to the sensor area. Comparisons should only be made at the same effective resolution to keep things fair.
Good, now make a comparison with base ISO, daytime and use RAW on the Canon.
+1111111111
The only thing that bothers me is that these kind of videos tend to increase the exposure together with the ISO, thus not showing a fair comparison between different ISO levels, since higher exposure contributes to mitigate the effect of noise in the most blown up parts.
So, even if it’s still impressive to see how A7S surpasses 5DMKIII, I’d like to see a series of comparisons with the correct exposure in every shot, so to be able to better notice the differences.
What are you talking about? The exposure was kept the same in this comparison.
No it’s not- The exposure is going up as the iso is going up! Much more interesting if exposure had been kept the same.
Exposure is the amount of light the sensor is exposed to, and it stayed the same. Shutter speed and aperture contribute to exposure. Sensitivity does not.
He means the highlights are blown out because they didn’t compensate the high gain with lower exposure after ISO 6400. He means in real life we would increase the shutter time to get our highlights back, thus the whole frame of the footage would be darker, so noise would be visible in the highlights and less visible in the shadows.
In short: He thinks ISO is shutter time + aperture + luminance + ISO
well, that was comparison in the 4k mode. this mode will likely not be used by most of A7S users.
more common use and meaningful comparison would be a comparison of the FullHD mode of A7S with FullHD mode of canon 5d
where it state its on 4K mode? i think its on Full HD only
he says in the youtube comments that it was shot in 1080p internally.
then the level of details on A7S is impressive. noise suppression in video most likely is also quite substantial, but overall impression is quite positive anyway.
Very impressive. I’m sure there’s some noise suppression but mostly there’s the inherent advantage of scaling 4K in camera to hit 1080P as opposed to the 5D which line skips. The detail in the A7S is astounding and it stays up until the very last bump (though 100K+ it gets almost too noisy)
Anyone know how to get LR to be fooled into reading RAWs from A7S? I know people have done it with other early RAW files.
Anyone? Really want to try another RAW developer as I am getting some weird results with the electronic shutter mode (though not a big issue).
How can you have a camera that’s yet to be released?
I don’t have the camera, though I have had a play with the camera and shot a bunch of test shots.
You could try some of alternative raw converters like Raw Therapee (free and much more powerful than LR). And if I’m not mistaken, Sony’s own IDC is sometimes capable of rendering files from new cameras without an update so that might work as well.
I guess you must be using some pre-series firmware and are not allowed to distribute the RAW files. Otherwise I think people here would be curious to play around with the files.
Well it was running latest FW and no problems sharing files I was told. Want to have a bit more of a play myself… but have to say they look amazing! King of low light is an understatement. Sony’s IDC destroys these files, they look horrible! Iridient Developer seems to work well, but a little bit of weirdness with shots taken using electronic shutter. RAW Photo Developer (RPP) seems about the same are Iridient at this stage. Might try RAW Therapee though and see how that goes.
To fool the editing software people change the camera name in the exif.
Thanks
I don’t understand these tests which display under and over exposure. Why not show proper exposure for each ISO setting?
What am I supposed to be judging by the under/over exposure displays?
Hahaha, so true. I was watching thinking “well, it’s bright white, what then?”
Yes, utterly ridiculous. Especially considering that the last few stops beyond 100k are extended anyways, i.e. digitally pushed.
Newsflash, “extended ISO’s’ are mostly marketing speak, regardless of brand. Beyond ISO 1000-1600, most Sony sensor equipped cameras for example already use digital gain instead of analogue gain. Most Pentax cameras even use digital gain for the majority of their ISO scales.
Okay, so the Sony and the Canon reach together “poor” level, but then the Sony outperforms: It can even reach “very poor”, “crap”, and “utter crap”. What a progress.
Again: This is senseless. Show shadowy RAW images at the same ISO level, scaled down to 12MP for both cameras.
Or leave it.
I have some, they look amazing! 51,200 looks better than my A7 at 12,800 for noise, DR is better across the board (well from 3200 up, didn’t get to try below).
Why not upload your test with a picture of ur A7S …..?
I’m going to.
It isn’t senseless it is a VIDEO TEST. For people shooting video this is a great comparison. Shadows and high contrast areas are typically places where DSLRs get very noisy/soft shooting video. The A7S clearly excels in this area. The test was not made as a comparison between the cameras shooting stills or RAW. Although incidentally I have some A7S RAW files and they are pretty decent. I don’t have a 5Dmk3 so I can’t compare. I’d say the A7S is at least a stop better than the A7 even scaled to 12MP.
Video is not as demanding as still photography. So this doesn’t tell anything about the limit, NEITHER for video NOR for still.
In what way is video not MORE demanding on a camera than taking stills? Also this is clearly not a test of limits it is very clearly stated to be a comparison of high ISO video between the A7S and a 5Dmk3. That’s it and that’s all it pretends to be. For what it is the video is clear and compelling. It is not exhaustive or scientific but I don’t think it has to be to have value. I’m not really sure where your comments are coming from except a strange place of generalized anger and frustration. If you want to generalize about stills quality you will have to wait for review sites with comparison shots like Imaging Resource and DPReview to evaluate the camera.
The 5DIII is a 22MP (!) allround camera. Comparing the A7S to the current low-light reference, the 16MP D4S, would make more sense!
Because the 5D3 is more use for video?
LOL?!
Than the D4s?
I’ve just had a look at Df RAW and A7S RAW…. A7S win by a country mile!
in what, i detail or high iso noise? A7S sensor has 33% bigger pixels compared to Df. Try to make the window of your room 33% bigger and you will see what effect you get. If I compare DF or D4 detail and finesse on full sized shots, they are just disappointing compared to those you get from an A7, and here, A7S will be the same, excellent in video and at full size a disappointment in stills. You can’t have it all, what you win on one side, you lose on the other, video and stills in one box will forever remain a compromise, making both to perfection side by side will never work. In every game there is a winner and a loser.
Sorry but at higher sensitivities the A7S produces more real resolution than my A7 does.
Exactly what program did you use to convert the A7s Raw files?
Iridient Developer, RAW Therepee, RAW Photo Processor and Sony IDC. ID developer is in contact with me and I’m sending him RAWS to correct support. There are some slightly weird results when using the electronic shutter vs mechanical but otherwise looks good and better than Df using the same developer.
Because it’s a a7s vs 5d3 test, not a a7s vs d4s test. Sooner or later there will be one out there
Because the 5Dmk3 is still the undisputed DSLR video king in many people’s mind. Yes, they are very different cameras but many people buy the 5D to primarily shoot video. For those people the A7S makes a LOT of sense. For everyone else it is a bit more niche.
Two Words: Magic Lantern
Again, it’s video. And the way how high ISO still images look may not be the same as the same high ISO video. What’s the problem with Sony? Why doesn’t it show the RAW still images but only video?
Another point is that 5diii is already more than a 2yo camera, which probably will be replaced next year.
because it’s a “video” test?
we know, but I think he’s saying we’d like a still test. why no still tests?
they will come, and the few samples published yet show clearly that without any noise rework, this camera noises at 3200 iso, same as other cameras do. The samples shown on another post yesterday, proved that at 40000 and higher the shots are blotchy and not useable. On video you do not look at same shot long time, it passes in a fraction of a second, with stills that game is different, and we will discover truth quiet soon, good high ISO shots due to an excellent Bionz engine and best possible rework. Now, if the new Bionz engine is able to compete with Nikon’s expeed found in D4 and Df, is another point of discovery and discussion.
No, because the title says that A7s kills 5diii in low light comparison. Since majority of people here are photographers and primarily shoot still images, one would think there’ll be a comparison between still images (+video). But when all you see is video, and this is not the first time, what would you think? 400k ISO, which nobody has seen in still images?
Yes it is, it’s a youtube “video”.
One thing you forget, the A7s is a camera that will be used mostly by videographers (almost 75% I think) rather that photographers, unless you want very low light situation (again that’s a small portion of photographers…). Others will use A7/A7r or A-mount for sports (A77ii have very usable high ISOs)
because it is a video camera. Sony has never published stills at departure, since there they can in now way compete with image quality of A7 and A7R. This lack of stills publishing shows us that this camera is a video camera par excellence and the side effect is that it also can take photos. Until now we had still cameras that could make video, now we have video cameras in still camera size and housing that can make photos as well.
When a Canon 5D3 is such a good camera, why can´t we see it ? Or why can´t DOX Mark measure it ?
Is Dox score the only indicator how good/bad a camera is?
….better than the feelings of some canonfans…
What does it have to do with canonfans? You’re either objective or subjective.
The 5D3 was measured by DXO Mark and got a score of 81. 24 bits for the colors, 11.7 EVs for the dynamic range and 2293 for ISO.
The this is only an indicator, not a measure of how good a camera is when you use it.
Sony 7r DXO Mark “””93″”””
Canon’s DSLR’s are very good at high ISO. A 6D is used for astrophotography by many pros (the overall score doesn’t reflect this as all ISOs are weighted and Canon’s sensor is not as good at ISOs below 1600 compared to Sony sensors).
Sony’s new cam is brand new and has larger pixels, it should of course be better at high ISOs. But the heading used for this article is typical, a language I don’t like. Why can’t we say it’s better, isn’t that sufficient?
The A7s is not better in low light compared to the 5DmIII because it uses large pixels but because it uses all (16:9 crop) of them for video. The Canon uses about 1/4 of the pixels for video due to line skipping perhaps not even that much. So to start off it is 2 stops behind and loses a bit more due to the more optimized Sony sensor. In the case of stills it is only going to be behind due to the difference in sensor tech level, so maybe 1/3 to 1/2 stop when the images are scaled to the same final output. The 5DmIII does very well with what it has to work with, but more area to gather light is always going to help in low light.
As soon as the CPUs get to the level to read entire sensors and scale the images on the fly the 12 MP sensor will lose its advantages.
Full sensor readout isn’t just about the CPU speed. It’s about the sensor’s readout speed, which ’til now hasn’t been the fast enough on sensors much higher than 12 megapixels. It’s why we haven’t seen silent shutter for stills on so many cameras even though that is not a real time application and it doesn’t require a fast CPU.
It is not s big surprise that low light video is better on the a7s.
Very true. It’s no surprise that with gigantic photosites like that, of course DR is going to be huge, of course it’s going to be a very high output sensor. If you are willing to trade in resolution to get all that in return…it’s a great stills camera too.
I wonder what the voltage measurement is on each pixel of this sensor vs a 36mp sensor pixel with the same light exposure.
watch out with what you claim, you will get enemies here. I can tel you about that an prove it. Let’s wait the camera to be available and see what comes out of it, after that we will reopen this file. This comparison is a donkey game, just hilarious.
It’s the full sensor read out that’s more responsible for better noise in video.
Does anyone know if this will be a good camera for stills only shooters? I like the idea of increased DR and low light and noiseless shutter….but would otherwise an A7R be better? (I have no use for video)
this is a video camera that looks like a camera camera
If 12mp is enough for u, and accept contrast detect AF only. Then this is a good stills camera
or maybe even the A7… it’s the least expensive of the 3 and 24MP is a good compromise between 12MP and 36MP. IMHO A7S is for crazy low light and A7R is for crazy big prints, it all depends on what you want to do with it 🙂
A lot of great professional work has been done with 12MP and lower. Yes, by today’s standards that is low but it is very good unless you are printing very large or need to crop a lot.
Yes
When a Canon 5D3 is such a good camera, why can´t we see it ? Or why can´t DOX Mark measure it ? http://0rz.tw/9BIIY
if Sony would get some lens out for that camera (like fast and wide) i’d switch, but they don’t so I’ll make due
Would have been nicer to have seen them shooting at the same time the same subject. I wanted to see when the a7s would blow out the whit of the fountains but was only able to see when the canon started crapping out.
I don’t know what this should have told me.
The last FF sensor that really was something big news, I would say was the
one used in the Nikon D3. That was a big step up.
Later we have only seen small improvements.
It goes in the right direction, but can’t any new sensors the last 1-3 years has
made a big jump in image quality. Just my thoughts.
Major improvements/breakthroughs are not commonplace. They are usually few and far between.
Sensor improvements are always incremental, but 4-5 years is a long time for digital technology. The current crop of sensors, are much improved over those available 4-5 years ago.
No… the D3 sensor wasn’t that much better than the 5D sensor.
Sony made this camera for the sole purpose of shooting video in low light. When you no longer shoot video in low light situations, you can use it has a pedestal for your canon 5d.
I think you can use the Sony for taking pictures too.
No.
After you purchase the A7s you can place the 5DM3 on a shelf where it can retire, or maybe send it to the Smithsonian. The 5DM3 has had it’s day in the sun, that day has passed.
Umm not if you need raw video, 22MP stills or a camera that van actually AF and be responsive.
This camera is made for video. The 5d3 is made for everything and can use many more lenses.
The 5d3 can use many more lenses? wtf? lol
The A7&A7r&A7s, can use almost any lens via adaptors. The 5D3; not so much. As far as still photography the A7s has much more dynamic range and much less noise. Haven’t compared the two directly but would likely prefer the output from the A7s, It just has a much more advanced sensor. The 5D3 is faster, both FPS and focusing.
There hasn’t been a proper dr comparison between them… or most other cameras for that matter.
Please no more boring videos…
Please no more stupid comments.
Please no more waffles.
Comparing a camera that has almost twice as many pixels to A7S is ridiculous. The high ISO video performance of A7S is due to lower pixel count what makes them almost twice as big as the Canon 5D pixels. Here it is just a physical law phenomenon. If you want a true comparison between both, then test also still picture quality and then you will probably find out quite fast that even here, high ISO performance in A7S is much better, but IQ is always on the side of Canon’s 5DMk3. The finer detail will take it over on A7S that makes a picture quality similar to Nikon’s D700, a better dynamic range shot with the “finesse” and detail of Nikon’s D70 or D40.
Even Nikon’s Df which is midway between the both, lacks of IQ when it comes to detail in the shot. Stop praising A7S qualities, that camera is aiming video use. In Still photography it is a snail and will remain a snail in every point of view. I still stick to the meaning that D700 sensor is the departure point for the one we find in A7S, sames size, same pixel count, just a remake for a different purpose, and as we see, it works excellent at the task it has been created for, what is video.
We believe you Shamael, not our lying eyes. You know Shamael, even a broken clock is right twice a day. You are always wrong, and not just a bit wrong but totally wrong.
This. +409,600 😉
well, truth hurts, that is a known phenomenon in a world where all is build on a card house of lies, isn’t it? Humans lie to themselves permanently and what offend them most is to admit they where wrong, so keep trucking that way, you are just the sucker that the system needs to keep working. I wish you good luck with 12 mpix FF, hope you get happy with that.
If some day it happens that I find out I was wrong, I will say sorry, it happened to me as much as it happened to others, but probably less often and I doubt it will happen this time. Maybe this is due to the fact that I experienced much more in my life than all of you pantpissers. 54 years behind the lens is a long way, many of you rookies do not come that far. This, for sure, does not mean that I am a better photographer, digital has brought stunning progression in photo shooting and some do mind blowing things with it.
54 years behind the lens doesn’t make you understand things better either. See my reply other reply to you.
I hear BMW has a new 4 litre engine, they had a 4 litre engine in the past so this MUST be the same one, just slightly different. BMW claim it is a completely new design, but we know better because it has the same basic capacity right? Wrong.
Stop dissing it when you haven’t even seen the output yet. And STOP saying it is a D700 sensor, it isn’t. Utterly sick of the same comment from you over and over again. Fine you have expressed your opinion but you constantly push it as ‘fact’. So let me ask you this, again, why do you think Sony, the biggest sensor manufacturer in the world would need to ‘borrow’ a Nikon design manufactured by someone else over 5 years old? Give me ONE legitimate reason that isn’t “well the pixel count and size are similar”.
There is nothing else like this sensor, it is completely new… Your constant commenting of the same thing over and over do not make what you say true.
It is certainly not the D70 sensor, but a remake of the same principle, it is basically the same amount of pixels, same size, different conception, but for that there is no reason that the IQ in stills could be in anyway better. 12 mpix FF is 12 mpix FF, and here all the technology used can not change anything to it’s resolution that equals a 6 mpix in APS-C. This camera might be a huge step forward in video technology, but despite all you wanna find to defend it, it is for stills a nut when it comes to resolution, it has 12 mpix, and you can not do anything better in stills photography than with a D700, except that latest software will give better results and more impressive out of camera jpegs.
When D700 came to the market, it was also mind-blowing in high ISO ability, in that time, and still today, it is a good performer. Now, the D700 sensor is not a Nikon sensor, it is a Sony sensor, sorry about that, but this has been proven often enough, despite of Nikon’s eternal lies on this. In this matter, Sony does not needs to borrow anything from Nikon, they have it since the made it.
This said, I think that I am free to express my opinion on every post concerning this camera, even if it could be disturbing to you, others do the same as much as I can read.
Well this just shows how ignorant you really are. 1) 12MP FF into APS-C is approx 5MP, not 6MP. 2) The D70 is CCD not CMOS. 3) Despite your claims that nothing ever gets any better, this is absolutely false and has been proven over and over again – Just look at NEX-7 vs A6000 or any other improvement in sensor quality over generations, let alone over many generations. The A7S has gapless micro lenses and full sensor read out… These are enough to know it is a completely new design.
Gap-less micro lenses are new? Really now? :
I believe they are relatively new and were not present on the D700 sensor.
The reason why the a7s is better than the 5d3 in video is not because of bigger pixels. It’s because of the full sensor read out. If the 5d3 also had a full sensor read out it would come close if not match it.
just add this to it’s quality, but do not think that the larger pixels do not contribute to the high ISO ability, specially at full sensor read out, Even if the 5D had full sensor read out, what is a challenge with that large amount of pixels, A7S would beat it in high ISO ability. There is no secret, larger windows, more light.
No…
1. The difference between full sensor read out and no full sensor read out can be just as much as FF vs. APS-C or worse. With no full sensor read out, not all photons are utilised and they go to waste.
2. Look up what happens to noise when images are re-sampled down.
E.g. 22MP -> 12MP.
3. Look at the quality difference between a higher resolution image that is down-sampled vs. an image produced by a sensor that outputs an image at the same smaller size.
E.g. [22MP sensor->12MP image] vs. [12MP sensor->12MP image].
4. Look in to why there are such things called micro lenses. Especially their relationship to what’s called “fill factor”.
That’s not to say there won’t be differences… but it’s not the way you say it is.
The “high pixel count is better” people refuse to acknowledge
that larger photosites collect more photons and produce a higher charge. At the
pixel level, they over saturate less quickly and generally have a higher signal
to noise ratio. They operate higher off the circuit noise floor too. They are
brighter because each pixel “bucket” generates a hotter “charge”…or “voltage”
if you will.
These are super inflated numbers but are easy to work with.
Imagine one huge 10 micron photosite that produces 1 volt at
“X” amount of exposure.
May people think that FOUR 2.5 micon photosites can be added
together to produce that same exact 1 volt of charge. In a black and white only
sensor, maybe it’s possible but not with a color Bayer sensor. No,..no,..no. I don’t
claim to know Sony’s math equations on this. Maybe only 10 or 20 people on
Earth know those formulas.
Pixels are basically not “added” together, they are AVERAGED
together and that completely different.
Large pixels will always produce cleaner and brighter
signals….period.
Forget about me, call everything I say crap. Just listen to
Sony’s own people. THEY TELL YOU HOW THEY DO IT. It’s not rocket science….They use very LARGE
photosites and they are very direct and blunt about it.
The Sony A7s is trading high resolution for greatly improved
“individual” pixel performance.
Forget what I say,…Listen to “Sony” tell you why it works.
That’s only one side of the coin.
The noise level in an image is mostly proportional to the number of photons that goes in to it. If comparing two sensors that have different MP counts, resolution must be equalised in post (via resizing the higher MP image) or it’s not an apples to apples comparison.
Once that’s done, the size of the individual photo sites isn’t the main thing. See how the 21MP 1Ds3/5D2 had the same level of noise as the 12MP D3/D700, when properly compared. There are lots of other examples too.
You do have a point about the dark current but that’s again at the pixel level. They also get averaged out at the end of the day. Again, many examples exist.
And what math equations? 🙂 Haven’t had any such issues when I code my raw conversion algorithms. If you want to get manufacturer spec exposure, that can be done extremely easily if you know what you’re doing.
Finally, the reason why the a7s is 12MP is very likely because of its relationship to 4K, bandwidth limitations when doing a full sensor read out and marketing reasons. Ideally, the sensor should have been a 39MP (7680×5120) one with an AA filter and full read out so that binning could be performed to get 4K without demosaicing. This would have given mind blowing amounts of resolution over the native 12MP sensor. But you need 3x the internal bandwidth for that and output data rate would also be higher because of the information density although the dimensions (4K) are the same.
Yeah, Canon 5DMIII is dead! Good bye Canon…
Other than the additional ISO settings available on the a7s… I’m not really seeing benefits in low noise at the compared ISOs.
Sharpness wouldn’t be an issue if the 5D3 was shot with raw video. But hey… if you want to go and put your money down, don’t let anyone else talk sense in to you 😉
I too am very interested in a comparison between ML RAW video and the A7S but you have to remember that RAW video comes with its own issues. I’d say this looks cleaner and at least as sharp as the RAW video samples I’ve seen and this is 50Mb/s in a standard codec and file format. ML RAW video is amazing but very daunting, unsupported by the manufacturer, and very, very expensive to shoot. RAW takes over 10X the data and requires the fastest and most expensive CF cards. Yes, it has more flexibility in post but if Sony can even match the IQ then that’s amazing I think. I suspect that as far as artifacts, sharpness, and dynamic range are concerned they may have surpassed ML RAW but that has yet to be seen.
Also, when I view the video at full resolution I can see the difference in noise quality well before 51k ISO. Both cameras are impressive in that respect but in noise and detail the A7S is clearly superior. Will Canon come out with something new? Sure. Will it be better? Maybe. What matters is that we don’t know when that will be and right now Sony’s camera is very superior in the qualities that matter to a lot of video DSLR shooters.
Yes the card prices are high but if the best quality is what one needs, it’s the available technology that will be the limiting factor. A 128GB 1000x CF card can store 18 minutes of raw video and costs USD285. 18 minutes might not seem a lot for a holiday video or for a wedding but then again you wouldn’t go through the trouble of raw video for that sort of thing.
exactly, I can record 18 minutes in XAVC-S for $20. On a documentary or corporate shoot where I’ll record 2-3 hours of footage in a day that matters. Not only people shooting holiday videos or weddings need more than 18 minutes.
Plus you’ve got longer encoding time and the extra space needed for the footage on the NLE. RAW video is an amazing option but I think more and more people are discovering it is pretty niche. I see a number of people recording to external recorders from their RED cameras because RAW takes too much time and money. There are still a lot of unknowns about the A7S but I think there’s now compelling evidence that it shoots better footage than a stock 5Dmk3 (which costs the same or more). Most 5D shooters are still shooting stock cameras.
Exactly. Never understand the ‘if you want absolute best you must suffer, or else you go back to shooting weddings’ mentality. False dichotomy when its all compromises. A7s is just more practical for hd and even does 4k. And can use leica m mount lenses.
You missed the context… we’re talking about quality. Also the 5D series has been used for plenty of serious work, more than any Sony Alpha. A 5D3 can certainly be used for regular shooting with the exception of 4K. If it’s just 4K you’re after, why not get a GH4 that can actually record it internally?
Then would you need an a7s for corporate videos? Must the entire documentary be shot in raw video?
Need is a funny thing. The A7S would be useful and seems to offer a higher video quality without the cost and hassle of RAW. Do I “need” it? Probably not but it would add a lot of value.
I’m not really sure of the point you are trying to make. With an A7S I could shoot everything at this higher quality standard. Why is that a bad thing? If I have to pick and choose which bits I shoot in RAW how is that an advantage? Sure it is for those bits but that extra bit of latitude is not worth the trade offs for me (and I suspect many others).
I guess you haven’t seen what’s exactly possible with raw video vs. compressed video.
Anyway, the point I’m making is this is probably not the best demo of the a7s or its comparison to a 5d3. With proper use, more is possible from both of these cameras… judging these two from this video is not a good idea.
If “best quality is what one needs” you shouldn’t be using either of these cameras. Even ML RAW suffers from a myriad of artifacts and quality issues. If you NEED the best you’ll be shooting with an ARRI or F65 or Dragon or …. something. If you start compromising based on price there will always be trade offs. While I wish Sony hadn’t made quiet so many trade-offs on the A7S I think their formula makes a lot of sense. If you’d rather live with the limitations of shooting RAW that’s your choice. We don’t really know yet the quality difference between the two but I suspect it will be a lot less than any other existing DSLR plus the A7S should have less moire and aliasing with very comparable dynamic range and noise levels (possibly superior on these last two but still too early to call).
See my other reply to you below.
i think part of of the reason the a7s is better is theres no line skipping, down resing or anything for 4k atleast, guess it scales well for 1080p too 🙂
Is it the time to say ‘Thanks, Captain Obvious’? Well no, this is just a comparison for fun. There is no such camera as the A7s in that price range, so there is nothing to compare (except you’re willing to compare it to the D4S).
The test hasn’t been thought through before being released because even at ISO51200, it’s not the differences in noise that we can see.
Night would turn in to day if the 5D3 was shot in raw video…
genotypewriter- have you actually done some tests? i shoot raw all the time with the canon 5dmk3 (I’m a stills photographer). the iso performance is very lame. i also shoot alongside TVC’s and the canon is miles behind the arri alexa in sensitivity (and lattertude). its not hard for the sony to leave it in the dust. the only way canon internal recorded video is considered ok with iso is that the video out is so soft that it hides a lot of the noise.
If you think the 5d3 has lame high iso performance in stills you have very unrealistic expectations and or you are not comparing things properly. I’m not saying it’s the best at high ISOs but it’s one of the best and with careful processing it can be brought up to the levels of what’s considered best.
My comment was about the sharpness of the canon video in this particular comparison… which can be vastly improved by shooting raw. It’ll also help if applying NR.
Not really trying to say whether the a7s is better than the 5d3 or vice versa. Way too early to jump in to that conclusion.
Is that a supprise?
When the RX1 came out there was a Sony rep talking about Sony’s disruption scheme. The peak of this disruption scheme is when the niche takes over and then it’s too late for the competitors to respond. We are now at that point…….. It’s “All In” time for Canon!
“Kills” ha! I can breath easy now that all the 5DIII are destroyed by the not so palm sized and “affordable” 7s. 😉
Den Lennie, another internet clown. No need to go into details why this video is useless, right Den?
a7s vs gh4 .
https://www.youtube.com/watch?v=yl0FHQuBpvI
gh4 looks better
And who is this Den Lennie? Already the name sounds stupid, the video is even worse.
Through simply exclaiming your ignorance as to who Den Lennie is you have identified yourself as an ignoramus. Before making silly statements, why didn’t you use the machine in front of you to find out who he is?
You may then understand the context and purpose as to why he created such a video. One thing I am certain of and that is his goal was not to knock the 5DMkIII (as some seem to suggest), but simply to use it as a benchmark as it is a very popular camera amongst filmmakers and videographers.
I suggest you visit the FStop Academy website and learn a thing or two. You may find Den Lennie’s expertise to be of value.
Hello, I am Den Lennie, I am an utter failure but I managed to get 10 minutes of fame. Envy me!
At least grow a pair and register your alias, “Tony”, before you go personally attacking people.