Sony Alpha 7S Review

I like IBIS better than OIS b/c lenses are small and cheaper.

The BAD part is that A7-2 does NOT have popular FL none OIS zoom lenses for example. not even in the lenses road map.

But be warned - you'll miss your point-and-shoot. You'll go through a period of saying, "I would have made a better shot with my old camera." You'll either get through it or trade it all in on an FZ200.

I've been using the Olympus OMD for a little over a year, switching from Nikon--and will not go back. There is a setting to turn off exposure preview so the EVF stays bright, even though the exposure is for the flash and the subject is lit with modeling lights or the room light many stops lower.

The first time is used the camera on assignment, I didn't know how to change this setting and the view finder was almost completely dark. I was guessing my compositions, but I pleased the client--and that's what really counts.

Sony A6000 has a setting for Liveview to just leave aperture wide open or not, I can't imagine the A7S doesn't have it.

Also Rishi I was wondering about effect of C-AF as well with 'simulated' liveview modes, but the behavior of the A6000 at least is not so simple...if you *want* a narrow aperture and the camera doesn't feel it has enough light it will open up the aperture further...the question then would obviously be in dynamic moving situations, having the camera take action to open up the aperture further will probably lose some keepers compared to just keeping it open all the time. In situations where there is enough light it probably doesn't matter, but in low light situations where you want a stopped down image, I imagine it would be to your advantage to not use simulated liveview (i forgot what they call it exactly).

Oh, actually, I think I misinterpreted your question. I thought you were interested in the aperture staying wide open, but you're more interested in a usable preview even when your exposure is set such that sans flash, you'd get a dark exposure.

In that case, yes, Sony does have a way of just keeping the preview at a reasonable brightness. You just set "Live View Display: Settings Effect OFF".

However, Timbukto - on Sony cameras, setting this to 'Off' does not actually mean the aperture stays wide open, since the aperture will close down if the camera 'thinks' the scene is bright. It'll then open/close the aperture to your selected aperture when acquiring focus, unless it thinks it needs to open it up b/c it's too dark, or close it down b/c it's too bright (to avoid saturation).

From there it gets really complicated really fast (as you say), as this behavior changes based on if the setting is 'on' or 'off'. I personally think the algorithms need some serious work.

Other cameras have been heavily penalized in the past for being specialized. I have not problem with giving full marks to a well designed specialized camera, I do have a problem when this is applied inconsistently. In this specific case it is even most obvious as this is more a video dedicated camera than a still camera, so again, it is even more of a problem in my eyes. For as still photographer it is really hard to justify to shell out this kind of money on a very limited still camera with limited lens selection. Even for a videographer I think it is hard to justify, because good 4K options will be readily available very soon, this camera will lose its marked value so fast. Poor investment unless somebody needs good 4K now and cannot wait.

For Armandino, A7s isn't only a very good video camera, is THE BEST camera for low light. Check the dxomark to convince your self. Also no shutter noisy, it makes it ideal for concerts shots.

"Best" is a qualitative expression. If the "Best" low light camera marginally edges other excellent cameras in the usable iso range holding you back in pixel count and AF performance it is a niche application, sorry. Especially when the system also suffers from limited lens choice. Electronic shutter is a nice touch indeed. Excellent iso performance and quite shutter do not make a gold and 86% to my taste if the rest is mediocre from a still point of view (we are talking of still cameras here right?)

We debated for a long time whether to score this as a video or stills camera, or both. And what the award refers to. It wasn't easy, and as always, there were compromises to both approaches.

Sorry you disagree with our final decision, but we were going to make some people unhappy, one way or the other.

If you want to check DXO, do it right. Look at the 18% SNR curve, it is virtually identical to the 6D from Canon. Look at the DR curve, you notice a huge and tell tale bump at ISO 6400. At ISO 3200 the curve would have dived under the 6D curve, but Sony starts to apply noise reduction above ISO 3200.
It is not the low light wonder Sony wants to make us believe or its 12mp low res. would suggest. Rather underwhelmed, to be frank.
http://www.dxomark.com/Cameras/Compare/Side-by-side/Sony-A7S-versus-Nikon-Df-versus-Canon-EOS-6D___949_925_836
Also check the dpreview comparison tool.

It does seem inconsistent. Great still cameras are often marked down for poor video, which is fair enough but the reverse should also apply

The lower the noise floor, the higher the DR. Sony Exmor sensors have higher DR due to lower read noise (because they have ADC on the sensor itself). Look at ISO 100, the 6D has lower DR caused by more read noise. The DR is "measured" by looking for a certain signal to noise ratio.

18% SNR is measuring signal to noise ratio on a midtone. Extended DR is in the dark blacks. As you will notice, the 6D/DF/A7s are all almost the same with the 18% SNR graph.

You're not going to get more DR out of the camera than what the Raw file provides at base ISO, which has been measured to be ~13EV by DXO. Unless the video signal pathway or output has something that bypasses the Raw signal pathway and the Raw signal pathway is somehow throwing away dynamic range.

Which I seriously doubt.

I wonder if Cinema 5D's results are from video that's had some level of noise reduction applied to it? If you apply NR to lower tones, you'd measure a deeper lower tone as having an acceptable SNR cutoff, thereby measuring a higher DR.

If that were the case, you wouldn't actually more DR than the 13 EV the Raw file provides, just more DR after NR - which is a non-standard way of measuring DR.

Either way, it still falls behind the Raw DR of the A7R, D810, D750, etc., b/c of increased downstream read noise compared to those cameras, which I personally think is due to quantization error b/c of the high FWC combined with the limited bit-depth of the ADC.

Where did you get that notion? Both cameras were shot at ISO 100.

Why do you think A7s has a base ISO of 3200?

If it had, the A7s would clip the image seriously at ISO 100.

The A7s has a native ISO of 3200, just as he says. It is published information. And that IS where it gets its full dynamic range. Not at ISO 100. Horses for courses. At IS3200, DR of this camera will beat the A7r's, at least in Video, others have tested this.

It turns out to be rather tricky to actually get the full DR this camera has on offer.

@Rishi. "Both cameras were shot at ISO 100 "

That is exactly my point. They shouldn't have been. Not if the goal was to find the true dynamic range potential. Base ISO is not always 100. It is for most cameras but not all. You seem to be making the wrong assumption that base ISO=100. Base ISO is the point for any given sensor where maximum dynamic range is possible. That is its' "native" ISO or where it is most happy. Anything outside of that up or down and image quality suffers. The A7s at 100ISO is outside of its' happy zone. Your review needs updated in that one section lest the credibility of the whole review be questioned. There is a reason Sony restricts the ISO in the higher dynamic range video settings S-log2 to 3200. That is because outside of this sweet spot the camera will not be able to capture enough dynamic range.

"conclusions moot" surely, or were the conclusions unable to speak?

The A7s has its maximum DR at ISO 80 according to DXOMark. There is also where it has its maximum TR and its maximum SNR. Where have you read that the native ISO is 3200? Sounds like nonsense to me.

@FineArtPhotographer - sure - let us take images at ISO 100 for A7r and at ISO 3200 for A7s and compare. That sounds more reasonable and fair - or?

We think we've worked out how you'd conclude that ISO 3200 is the native ISO (though we don't believe this to be the case). Could someone provide a source for this statement, so we can confirm or disprove our suspicions?

Agree, I also want to have some reference of the claim.

NOTE that exactly what native ISO is can be disputed, as e.g. the ISO definition do not talk of such a thing.

To the best of my knowledge native (or base) ISO of a sensor is the ISO you get when you measure clipping of the sensor.

No, that's incorrect. The base ISO of the camera is 100, not 3200, as clearly demonstrated by DXO. That's where the camera has maximum dynamic range, as also demonstrated by DXO.

Perhaps you'd like to define the term 'native ISO'... preferably without the word 'happy' in the definition?

wogg: "At IS3200, DR of this camera will beat the A7r's, at least in Video, others have tested this."

And at ISO 100, DR of this camera will beat DR of itself at ISO 3200... which is why we tested at ISO 100, where the camera DR is maximal.

If you have a high DR scene, you don't shoot at ISO 3200. On either camera. Not if there's anything you can do about it.

Big mistake for dpreview. It seems you guys were a little bit too anxious.

DP did it correct. A7r has better DR than A7s @ 100-400 ISO . A7s is better than A7r @ 3200 , but it's still nothing compared to what both cameras achieve @ 100.

Googled a bit.. seems that ISO 3200 is the native (lowest?) ISO if shooting video in S-log2 mode. Just a video thing. Nothing to do with still images.

@Steen Bay - absolutely right, ISO 3200 is the lowest available ISO when using the SLog2 mode (which appears to equate to the same sensor amplification as ISO 640 mode does with a standard tone curve, confusingly). But at the sensor level, 'ISO 100' appears to be base in the sense of offering maximum DR.

Big mistake for people who think it was a big mistake for DPReview.

Max DR is at base ISO, which is 100. SLog2 does raise the minimum ISO to 3200, as we pointed out here. On that same page, we also pointed out that hardware level ISO amplification appear to go up 4x (to be more precise, between ISO 400 & 640 levels), so dynamic range actually drops w/ SLog2 compared to ISO 100. In fact, it was b/c of this that we specifically asked Sony engineers why they raised the minimum hardware-level ISO amp as opposed to simply applying a more aggressive tone curve in SLog2 w/ ISO 100 levels of amp, as the latter would preserve the maximum DR the camera can record. You could use higher ISOs if needed b/c of exposure limitations, but the option for ISO 100 levels of DR would've been nice w/ SLog2.

As it is, though, you're always paying a DR cost by enabling SLog2, which we feel is unfortunate, & goes directly against OP's notion.

What'd be really nice would be if people actually read our reviews before making comments or making big assumptions that end up being completely wrong. A more appropriate way to address a topic you're clearly confused about is to merely ask us 'hey, as I understand it, it's like so-and-so... thoughts?'

I realize that's it's entirely unreasonable for me to expect civil behavior on the internet, but, still, it'd be nice.

Hope this clears up all the confusion.

So the SLOG base ISO of 3200 was chosen why? I get that its for video but there has to be a reason Sony chose it for this particular sensor.

Not the end of the world however and it doesn't change Dpreviews excellent findings - especially since Sony will have a mega-megapixel FF camera coming soon for landscape photographers.

The a7s isn't for those customers anyway.

Peace.

"So the SLOG base ISO of 3200 was chosen why? I get that its for video but there has to be a reason Sony chose it for this particular sensor."

Yes exactly, that's what we were wondering. And we never got an answer to it from any Sony rep, unfortunately.

I did make it clear though that in their next implementation, it'd be nice if they offered lower levels of hardware ISO amp.

In fact, we're confused about where the stated 'ISO 3200' even comes from- we measured ISO 640 levels of hardware amplification (by comparing clipping levels in Raw files for ISO 3200 SLog2 shots vs all other ISOs, finding that clipping occurred for the same tones in the scene at ISO 640 vs. a SLog2 ISO 3200 file).

I.e. SLog2 does affect the Raw- at ISO 3200, it's like a ISO 640 Raw in terms of actual Raw signal. But then the ISO 3200 SLog2 JPEG has a drastically non-standard tone curve applied. We figure it's called ISO 3200 b/c midtones are brightened to 3200-ish levels (but even, they're more like 1250).

I appreciate the insight. I know it can be tiresome to see all the critics in these forums but your answering questions is helpful.

Cheers

Lassoni makes a good point and another positive of the A7s is how much color information is held onto by the camera as the ISOs rise - dynamic range charts are only part of the battle when showcasing what a camera can do. Nikon and Sony are so incredible these days that photographers are truly spoiled.

I never shot film above ASA 800. Yes 800....and now even Micro4/3rds can do that clean without issue.

Wonderful time to take pictures and have fun!

Rishi, you know as well as the rest of us that it's not "free". If you're not the customer, you're the product - as is the case here:

"With respect to comments or other text-based content you submit or make available for inclusion, you grant us a nonexclusive, royalty-free, perpetual, irrevocable and fully sublicensable right to use, distribute, reproduce, modify, adapt, publicly perform, translate, create derivative works from and publicly display such content throughout the world in any medium."

"We reserve the right to make changes to the Web Site, policies, and these Conditions of Use at any time."

There is no opt-out specified, i.e. dpreview owns all content and can do whatever they want with it, including reproducing it in any form.

So I don't think you should be especially exempt from criticism - quite the opposite. You are on a payroll, not working for a charity.

@rishi:

In a market that moves very quickly. New products coming out every 9 months. Would you not agree that if a review is to be used as a benchmark for a purchase, then a review needs to surface relatively quickly?

I don’t think many are waiting for 12 months before a purchase... just to see what score DP has given a camera.

And if a review is not to be used as a benchmark for a purchase - what should it then be used for? Cosy Reading?

Today I use other sites then DP. DP is simple just too slow with new reviews. And personally I cannot wait 12 months before I make a purchase… and I’m just a private guy with a love for photography. Think about professionals.

DP you have to stop up – and release your reviews faster than today. I believe that most people would rather have a review come out fast and with fewer details – then wait a year and get a review with a lot of details. As said earlier… it’s a very fast moving marked.

Think about it.

We understand what you're saying, and we agree.

Any full frame lens on the planet will work on the A7, it's amazing how DSLR fan boys are such BS artists.

It is not a total game change for film stills photographers. One the camera maybe silent but the lenses are not. That is what the lens tubes are for on the Blimp. Two the silent shutter cannot be used for everything either. It is not good for subjects that are moving because of rolling shutter artefacts. Hence why most film stills photographers will not touch them.

As a counterpoint to the popular opinion out there on the a7 II thus far: after shooting with the a7 II this weekend, I really appreciated picking up an a7 at work today. And it was due to ergonomics. We'll tell you why in our full a7 II review. :)

Totally disagree. I think ergonomics are fundamental. Just about any digital camera produces great pictures and has loads of features now. For me, comfort and weight are crucial. I bought a D750 as it was the first Nikon I had ever held that was comfortable and for a DSLR it is relatively small and lightweight yet offers the advantages of lightening quick AF. I took back a Fuji X-T1 as it was not comfortable - the shutter button is too far back on the top of the body, like on the A7...

So are you going to upgrade the ergonomics score for the A7 now :D

Hey Rishi, can you let me know when next year to expect the A7II review? ;-)

there are only 3 variants in the first generation, and one version thus far in the second gen..if you can't keep track of that then you may have a missing toe or finger and not even realize it

They couldn't implement IBIS (in A7II) in one of 3 models before?! Before, is somwhere 12 months before.
They announced 4 models in 1 year span. With all the respect, it looks like deliberately money sucking.

@ Splendic

Not at ISO 25600 they don't!
I shoot jpegs with this camera as when shooting a festival or even some gigs, I will end up with up with hundreds of photos and I hate post processing.
The A7s Jpegs are good enough for me to give to a band, festival organisers or Newspaper at any ISO up to 51200.
The 12mp size also means I don't have to resize to email any as well....I just pick the ones I want and send as taken.

Splendic,
My friend is photo jurnalist. He shoot only JPG on his ff.

Thing is, one can go down a rabbit hole w/ what you're suggesting. B/c it's not even just about the number of stops pushed, it also matters which tones you're pushing. So you could just as easily ask us to try using different sets of exposures per ISO to place scene tones at different exposure levels in terms of sensor exposure...

So, ideally, we'd shoot a wedge that has more DR than the sensor, then do all the pushes, then quantitate SNR to look for where there are benefits to increasing your ISO amplification vs. not.

Or have a high DR scene (unchanging, mind you!) that extends beyond the DR of the sensor so you can choose which tones benefit, or don't, from ISO amplification vs. digital boosting.

If we did that, w/ a 'real scene' or wedge, yes, ideally we'd do it in the way you're suggesting. That'd be comprehensive.

We're exploring doing this sort of stuff, it just has to be balanced with time requirements, & the cost it has on other tests/content.

Yes, understand that you don't have infinite time & resources, but what I suggested would just require 2 extra shots at base ISO. 4 stops underexposed (and pushed) and 2 stops underexposed/pushed. Most people will probably never (or only very rarely) push the base ISO shadows more than 2-4 stops anyway, and if they don't, then maybe even a Canon DSLR would be doing just fine.

No... you'd need a comparison point, so you'd need ISO 1600, vs. ISO 100 pushed 4 EV, then ISO 400, vs. ISO 100 pushed 2 EV... and then we'll get someone asking for 3 EV, so then we'll need ISO 800, vs ISO 100 pushed 3 EV, then someone will ask for 1 EV... and soon you find yourself doing another whole series 'normalized' to ISO 1600 brightness. Then someone's going to ask for it normalized to ISO 3200 exposures. And so on and so forth.

I totally think it's worthwhile, yes, but every time a new camera comes out, you'd have to re-establish what that push amount is that leads to reasonable, but not unreasonable, results. Else you risk telling people nothing other than what that particular push (that you want to see) demonstrates. Our test undoubtedly shows you more than that, b/c it shows you what the results are comparatively across a number of cameras.

Can't quite see why a "comparison point" is needed. What I'm interested in here is base ISO DR. How many stops the base ISO shadows can be pushed with a reasonable result. To get an idea about that it's just necessary to underexpose for example 2, 4 and 6 stops and lift the shadows afterwards.

"ISO-invariance" is a different thing, and guess it's useful to know about that too if you want to shoot ISO-less, but don't think it's necessary to mix it up with the tests of base ISO DR.