Sony extends a7 line with new a7S

[mcraige]

I just saw this in the B&H newsletter:

http://www.bhphotovideo.com/explora...ment-NAB2014-_-140408-_-Body_Explora_Sony-a7S

full-frame
12.2MP CMOS sensor
record Ultra HD 4K (3840 x 2160) video with no pixel binning to an external recorder
an expandable sensitivity range to ISO 409600
an apt 25-point contrast-detection AF system
a top Speed Priority continuous shooting rate of 5 fps, as well as a 2.5 fps shooting rate with maintained autofocus

 

[havoc315]

Not sure who this is really made for.
It isn't exactly for 4k home movies -- As noted, you need an external recorder to actually make use of the 4k video.
At 2.5-5fps and contrast detect AF, it's not a sports shooter, like other "low" MP cameras.
It's fairly low resolution for 2014... While technically enough megapixels for 99% of needs, it's less than demanding users are used to now.

It is being promoted based on its extremely high ISO ability. But is it really "better" than other cameras in low light, or is the noise simply being hidden by the lower resolution? Or both? Slightly better noise performance, that looks much better since it's also smaller. Guess we will need to wait for objective testing on that one.

In terms of ISO range, it's similar to the Nikon 4ds. And I suspect that the price will be much cheaper than the Nikon D4s. But the Nikon is 16mp, 11fps, with a huge phase detection AF array.

So I can see the camera being useful for professional/semi professional videographers who typically use a whole rig for video.
In terms of regular consumers, maybe the low light performance will be the key selling point. Will need to wait for objective testing of just how good it is. It would have to be shooting super clean at 25,600 and higher in order to convince me that it is worth the lower resolution, etc. (I know it can go 4 stops higher than even that... but I doubt the highest stops are of any real usefulness)

 

[hakepb]

The Nikon D4s has a real ISO range up to 25,600 and is expandable to 400k

The Sony A7s has a "real" ISO range up to 102,400 and is expandable to 400k

It will be interesting to see if the Sony has any real advantage above 25600.

 

[havoc315]

The Nikon D4s has a real ISO range up to 25,600 and is expandable to 400k

The Sony A7s has a "real" ISO range up to 102,400 and is expandable to 400k

It will be interesting to see if the Sony has any real advantage above 25600.

From what I've seen online, the Nikon D4s is about 1/2 stop better in high ISO than your typical consumer grade full frame camera like the Nikon d800. (DXO scores the D4 about the same as the D800). At least when comparing raw. When comparing jpeg, the difference is bigger- suggesting the D4s has improved jpeg noise reduction.

For detailed work like portraits, the images start to fall apart above ISO 6400. (Eyes get fuzzed by noise, skin tones get disrupted) Can maybe stretch another stop to ISO 12800. For less detailed shots, downsized, you can stretch another 2 stops to 51200.

So when push comes to shove... Will the 7s give us clean sharp portraits at 25000+? Will we get clean usable shots over 100,000+?

I suspect it will only be an incremental improvement. Sure, push another fraction of a stop over the D4s. ( which is still impressive if offered for 1/3rd the price).

 

[fractal]

Looking pretty amazing to me.

http://www.youtube.com/watch?v=XgbUgNiHfXM

 

[havoc315]

Looking pretty amazing to me.

http://www.youtube.com/watch?v=XgbUgNiHfXM

Saw the video. Pretty impressive demonstration of high ISO *video.*

Which is still quite different than high ISO stills. Even 4k video is much lower resolution than photographs. And quality of video in a tiny youtube box does not reflect the type of quality you would want in even normal sized photographic prints.

Still, it is indeed very impressive for anyone who wants to do a lot of low light video. My A99 tops out at video ISO of 3200, so it doesn't exactly do low light video.

So far, I haven't yet seen any tests or demonstrations convincing me that the high ISO still images are significantly better than other cameras on the market. (If it was really 2-3 stops better than any other camera in existence, they would be doing a massive product release, not a footnote at NAB. It would be one of the biggest breakthroughs to ever hit digital photography. Sony promotes the A6000 as having the "fastest autofocus of any ILC" -- If the A7S high ISO performance was really such a leap forward, Sony would be blasting, "Best low light performance of any full frame camera in the world."

Differences in high ISO performance between sensors and brands tend to be very very incremental. There has been major improvement over the last 10 years, since dSLRs were first released, but those advancements have slowed. Even combining the progress over the last 10 years, you've really only seen around 1-1 1/2 stops of improvement over 10 years. (2004-2007 APS-C dSLRs were producing clean ISO up to around 400-600. Now they can hit clean ISO of 800-1200. Full frame dSLRs haven't been around as long, but current full frame dSLRs rate about 1 stop better than early generation full frame dSLRs in high ISO according to DXOMark.)

So with about 1-1 1/2 stops in high ISO improvement over the last decade, I doubt that this camera has suddenly advanced several more stops.

 

[SkaGoat]

My guess is it's low-light performance will be similar to Nikon D4s, the sensors, though different pixel counts, are probably very closely related.

It seems to me this camera is aimed at video, and they are achieving the low light performance by dropping the megapixel counts, meaning each pixel on the full frame sensor can collect more light. This makes sense, if you are designing it for video use, you don't need a whole lots of pixels, just need a lot of throughput from the sensor.

 

[Animagic!]

I'm in "wait and see" mode until the reviews come pouring in.

All I want to know is how will this camera be on dark rides? Obviously it can handle low light. I'm just wondering about the 12MP resolution. I understand why 12MP lends itself to low noise. I just wonder if the A7s is geared to much toward videography.

 

[havoc315]

I'm in "wait and see" mode until the reviews come pouring in.

All I want to know is how will this camera be on dark rides? Obviously it can handle low light. I'm just wondering about the 12MP resolution. I understand why 12MP lends itself to low noise. I just wonder if the A7s is geared to much toward videography.

I'd bet similar to the Nikon D4s which is also only 16mp.
Fact is, all modern full frame cameras, with a fast lens, can handle dark rides per well. If you are in fact getting a bit more ISO performance from the camera, it would allow you to possibly get away with a slightly slower lens (2.8 instead of 1.8/1.4), might allow you to shoot at slightly faster shutter speeds or smaller aperture, to get slightly sharper images. Of course, with lower noise.

With only 12mp, you are certainly getting "enough" detail, but you will be resolving less detail then the 20-36mp full frame cameras on the market.

But again, it's doubtful that it is substantially better than other full frame cameras. So if it's giving you an additional 1/3rd of a stop compared to a Canon 6d or Nikon D600, or even a whole stop over the Sony A99, that's not going to be enough to re-define dark ride photography.
Might come down to something like this:
With a APS-C camera and fast lens, you can get a good dark ride shot that will print between 3X5 and 4X6.
With a full frame camera and fast lens, you can get a good dark ride shot that will print between 4X6 and 5X7.

And IF the A7s is actually a "leap" forward, then maybe it will let you get a good dark ride print as large as 8X10.

(The Sigma 18-35 1.8 is a much bigger deal for dark ride photography as it may be the only zoom lens truly suitable for dark rides).

 

[Animagic!]

Honestly I'm looking at the A7r. Maybe the A7.

I like the idea of the Nikon D610, primarily because of the value and lens line up. However since a A7 is the same price(ish), I'd much rather have the EVF.

Anyone have any strong arguments for either the A7 or A7r? I'm leaning towards the A7r for the better low light performance and overall image quality. However, the A7 is much cheaper. And with B&H offering $200 of FE lenses with the purchase of a body I could get a A7r with the Sonnar T* FE 55mm f/1.8 ZA Lens for only $200 more than the A7r body.

 

[havoc315]

Honestly I'm looking at the A7r. Maybe the A7.

I like the idea of the Nikon D610, primarily because of the value and lens line up. However since a A7 is the same price(ish), I'd much rather have the EVF.

Anyone have any strong arguments for either the A7 or A7r? I'm leaning towards the A7r for the better low light performance and overall image quality. However, the A7 is much cheaper. And with B&H offering $200 of FE lenses with the purchase of a body I could get a A7r with the Sonnar T* FE 55mm f/1.8 ZA Lens for only $200 more than the A7r body.

I would not get the a7r. I've read in many reviews that the a7r is really meant as a studio camera. It's 36mp raw files are too large for a regular workflow for many. There have been issues with the shutter causing motion blur at moderate shutter speeds. It has a very loud shutter. (Bit quieter on the a7 which can use electronic first shutter).

I've seen several reviewers comment that consumers are making a mistake if they buy the a7r with the assumption that it's better just because it's more expensive.
The a7 gives you 24mp files which is plenty of resolution for almost everyone. Phase detection hybrid AF in theory gives you a tracking focus advantage. Can shoot at a higher burst rate. Electronic first shutter.
The a7 is cheaper -- and for most users it's probably better.

You might also want to consider the a99 which finally has a price break. It's a fantastic camera. Not quite as good in low light as the a7/r. But makes up for it with stabilization, a more ergonomic body (if you want a bigger body with more direct controls), and a better selection of native lenses. (Yes you can use those lenses on the a7, but you need an adapter and you lose stabilization, and you lose the low light advantage).

If I was buying from scratch, I'd be considering the a7 and a99. But my prefix sting lineup of a-mount glass pushed me to the a99.

Starting truly from scratch, I'd consider the Nikon d610. But I'd probably go towards the a99, for access to Minolta lenses, EVF, live view, and stabilization on all lenses.

 

[Animagic!]

I would not get the a7r. I've read in many reviews that the a7r is really meant as a studio camera. It's 36mp raw files are too large for a regular workflow for many. There have been issues with the shutter causing motion blur at moderate shutter speeds. It has a very loud shutter. (Bit quieter on the a7 which can use electronic first shutter).

I've seen several reviewers comment that consumers are making a mistake if they buy the a7r with the assumption that it's better just because it's more expensive.
The a7 gives you 24mp files which is plenty of resolution for almost everyone. Phase detection hybrid AF in theory gives you a tracking focus advantage. Can shoot at a higher burst rate. Electronic first shutter.
The a7 is cheaper -- and for most users it's probably better.

You might also want to consider the a99 which finally has a price break. It's afantastic camera. Not quite as good in low light as the a7/r. But makes up for it with stabilization, a more ergonomic body (if you want a bigger body with more direct controls), and a better selection of native lenses. (Yes you can use those lenses on the a7, but you need an adapter and you lose stabilization, and you lose the low light advantage).

If I was buying from scratch, I'd be considering the a7 and a99. But my prefix sting lineup of a-mount glass pushed me to the a99.

Starting truly from scratch, I'd consider the Nikon d610. But I'd probably go towards the a99, for access to Minolta lenses, EVF, live view, and stabilization on all lenses.

At the end of the day, low light performance is the most important factor.

The A7r scored so much higher than the A7 on DXOmark, with a 500 point lead in low light. To me, 500 points is pretty significant. I like everything about the Nikon D610 except the OVF. I've considered the a99 but for the money I think I'd rather have the A7 with it's better ISO performance over a99. But I'll admit its very hard to get excited about the FE mount.

However, since Sony is theoretically releasing new A Mount FF's I may just wait for one of those. An a99II may be the best possible option, especially since I already own a couple FF A mount lenses.

 

[havoc315]

At the end of the day, low light performance is the most important factor.

The A7r scored so much higher than the A7 on DXOmark, with a 500 point lead in low light. To me, 500 points is pretty significant. I like everything about the Nikon D610 except the OVF. I've considered the a99 but for the money I think I'd rather have the A7 with it's better ISO performance over a99. But I'll admit its very hard to get excited about the FE mount.

However, since Sony is theoretically releasing new A Mount FF's I may just wait for one of those. An a99II may be the best possible option, especially since I already own a couple FF A mount lenses.

Remember how ISO is measured. The difference between 2300 and 2800 is 1/3rd of a stop. So it's about the same as the difference between 400 and 500. Or about the same as 100 and 130.

In other words, not a huge difference. Pretty small actually.

DXO mark scores can we odd as well. You need to balance it with other reviews. For example, the Canon 6d rates below the Nikon d610 in low light according to DXO mark. But according to most other reviews, the Canon is better.

But getting back to ISO, A7, and low light performance...

Remember that aperture and shutter speed determine what ISO you can use. So wide apertures and stabilization are also critical for low light.

So according to reviews there are 2 similarly great primes-- the new Sigma 50/1.4 which you can get for the a-mount. And the 55/1.8 on the A7/7r.

So lets stick the Sigma on the a99. It becomes stabilized. So you can take a low light shot at ISO 1600, f 1.4, shutter speed of 1/25 (thanks to stabilization).

Now on the a7 or a7r-- your aperture is only 1.8, and you probably need to shoot at around 1/80. Now to get a proper exposure, you need to shoot at ISO 6400.

So which is likely to be better? The a99 at 1600, or the a7/7r at ISO 6400?

Of course, if you exclusively use stabilized zoom lenses with the a77, then the gap narrows. But lens apertures remain smaller than a-mount zooms. No 2.8 zooms for the A7/7r.

Yes you can adapt the a-mount lenses. But if you want autofocus, you need the lae4 adapter. And this adapter causes a loss of half stop of ISO. So it cancels out the ISO advantage, and you lose stabilization. A double loss.

 

[SkaGoat]

The A7r scored so much higher than the A7 on DXOmark, with a 500 point lead in low light.

According to dpreview.com, if you look at the noise chart the two cameras are almost identical, DXOmark means nothing. Their bias against Canon cameras is clear, and I don't know if it's intentional, or Nikon/Sony (a lot of Nikons have Sony sensors in them) have figured out how to game their tests.

I can tell, just from looking at test shots that for noise A7/A7r are almost the same, 6D is quirt a bit worse than the Sonys, and D610 is quite a bit worse than the 6D.

 

[fractal]

Remember how ISO is measured. The difference between 2300 and 2800 is 1/3rd of a stop. So it's about the same as the difference between 400 and 500. Or about the same as 100 and 130.

In other words, not a huge difference. Pretty small actually.

DXO mark scores can we odd as well. You need to balance it with other reviews. For example, the Canon 6d rates below the Nikon d610 in low light according to DXO mark. But according to most other reviews, the Canon is better.

But getting back to ISO, A7, and low light performance...

Remember that aperture and shutter speed determine what ISO you can use. So wide apertures and stabilization are also critical for low light.

So according to reviews there are 2 similarly great primes-- the new Sigma 50/1.4 which you can get for the a-mount. And the 55/1.8 on the A7/7r.

So lets stick the Sigma on the a99. It becomes stabilized. So you can take a low light shot at ISO 1600, f 1.4, shutter speed of 1/25 (thanks to stabilization).

Now on the a7 or a7r-- your aperture is only 1.8, and you probably need to shoot at around 1/80. Now to get a proper exposure, you need to shoot at ISO 6400.

So which is likely to be better? The a99 at 1600, or the a7/7r at ISO 6400?

Of course, if you exclusively use stabilized zoom lenses with the a77, then the gap narrows. But lens apertures remain smaller than a-mount zooms. No 2.8 zooms for the A7/7r.

Yes you can adapt the a-mount lenses. But if you want autofocus, you need the lae4 adapter. And this adapter causes a loss of half stop of ISO. So it cancels out the ISO advantage, and you lose stabilization. A double loss.

The FE 55 with OSS would have certainly been a more useful lens, however I think your comparison here is a bit exaggerated. I took this shot in motion (Space mountain) with my nex-7 (APS-c sensor) unstabilized lens ( Sigma 19mm ) ISO 3200, f/2.8 at 1/15 sec. It's not the greatest shot in the world, but with proper technique and practice you can shoot at much slower than 1/80 sec in low light.

 

[Animagic!]

Remember how ISO is measured. The difference between 2300 and 2800 is 1/3rd of a stop. So it's about the same as the difference between 400 and 500. Or about the same as 100 and 130.

In other words, not a huge difference. Pretty small actually.

DXO mark scores can we odd as well. You need to balance it with other reviews. For example, the Canon 6d rates below the Nikon d610 in low light according to DXO mark. But according to most other reviews, the Canon is better.

But getting back to ISO, A7, and low light performance...

Remember that aperture and shutter speed determine what ISO you can use. So wide apertures and stabilization are also critical for low light.

So according to reviews there are 2 similarly great primes-- the new Sigma 50/1.4 which you can get for the a-mount. And the 55/1.8 on the A7/7r.

So lets stick the Sigma on the a99. It becomes stabilized. So you can take a low light shot at ISO 1600, f 1.4, shutter speed of 1/25 (thanks to stabilization).

Now on the a7 or a7r-- your aperture is only 1.8, and you probably need to shoot at around 1/80. Now to get a proper exposure, you need to shoot at ISO 6400.

So which is likely to be better? The a99 at 1600, or the a7/7r at ISO 6400?

Of course, if you exclusively use stabilized zoom lenses with the a77, then the gap narrows. But lens apertures remain smaller than a-mount zooms. No 2.8 zooms for the A7/7r.

Yes you can adapt the a-mount lenses. But if you want autofocus, you need the lae4 adapter. And this adapter causes a loss of half stop of ISO. So it cancels out the ISO advantage, and you lose stabilization. A double loss.

I think from a physics point of view this is true. I think in practical terms it's all relative. I don't think its so black and white with cameras and lenses.

I can tell you that my Pentax K30 with 1600 ISO and f/2.4 at 35mm can DRAMATICALLY outperform my Sony a58 with ISO 3200 at f/1.4 at 30mm. The difference is huge. On the Little Mermaid dark ride my k30 was metering at 400 ISO and turning out great shots. At 3200 ISO my a58 was struggling. According to physics the A58 should have had really high shutter speeds and captured great shots. Instead the k30 blew it out of the water and did so with way less noise. My Sony Nex5 captured the same scene at ISO 1600, 50mm, 100 sec and did so nicely. I think a lot of it has to do with the camera and how the camera handles light and high ISO, hence why I reference DXOmark for their ISO performance.

 

[havoc315]

I think from a physics point of view this is true. I think in practical terms it's all relative. I don't think its so black and white with cameras and lenses.

I can tell you that my Pentax K30 with 1600 ISO and f/2.4 at 35mm can DRAMATICALLY outperform my Sony a58 with ISO 3200 at f/1.4 at 30mm. The difference is huge. On the Little Mermaid dark ride my k30 was metering at 400 ISO and turning out great shots. At 3200 ISO my a58 was struggling. According to physics the A58 should have had really high shutter speeds and captured great shots. Instead the k30 blew it out of the water and did so with way less noise. My Sony Nex5 captured the same scene at ISO 1600, 50mm, 100 sec and did so nicely. I think a lot of it has to do with the camera and how the camera handles light and high ISO, hence why I reference DXOmark for their ISO performance.

I'm a bit confused. Of course ISO 1600 will outperform ISO 3200. And certainly, of course ISO 400 will outperform 3200 massively.
Lower ISO is always better.

You need to compare shots of the same exposure.

If you take the exact same shot with the Pentax and the A58...
ISO 1600, f2.4, same shutter speed, same focal length. Then look at the pictures in the same size. How do they compare?
And comparing jpegs or raw? (To my knowledge, Pentax doesn't use a true RAW. It applies a bit of noise reduction even to RAW.)

Anyway, my expectation. If you compared to a58 to the Pentax. My best guess is that you wouldn't see any difference up to ISO 800 or so. At ISO 1600, the Pentax may look a little better if you pixel peep. At 3200, the Pentax will look noticeably better. Still not huge, but noticeable. And by ISO 6400, both cameras are struggling. Can still produce a small dark ride pic, but you don't want to produce any 8x10s. (Just based on my use of the a55 and comparing it to other cameras ).

Anyway, since you own the 3 cameras... If it's in your interest, run a low light test using the exact same settings on each camera. (Which should result in the same exposure on each.)

 

[havoc315]

Ran a test myself, comparing the A99 with the Rx100. According to DXO Mark, it scores the A99 at about 1600. It scores the RX100 around 400. So there should be a MASSIVE difference, right?

Let's see how big a difference of 2 stops... 1200 points really makes. I shot RAW+JPEG at ISO 3200, f4.5 and shutter speed of 1/60. So in raw, you can see how both cameras handled the picture with no additional post-processing. And you can see how each image looks after in-camera jpeg noise reduction.

RX100 raw:

rx100raw.jpg by Havoc315, on Flickr

A99 raw

a99raw.jpg by Havoc315, on Flickr

In raw, the A99 looks better. But how much better?
In regular viewing size, I can't really even see the difference. When pixel peeping, the A99 noise is clearly finer.

Now how do they both look when the camera applies jpeg settings:

rx100jpeg.jpg by Havoc315, on Flickr

a99jpeg.jpg by Havoc315, on Flickr

At this point, I see almost no difference. So this massive DXOMark is practically erased. The RX100 has slightly more detail loss, but it's very close.

Now you will also find differences in the metering systems of the cameras, autofocus quality, etc, which will affect the performance. Especially if you rely on automatic settings. But DXOMark scores don't tell you anything about that. It sounds like you simply prefer the Pentax metering system over the Sony Alpha metering system, in which case you wouldn't be happy with any of the A7 models either. (I've read a lot of reviews complaining about the shutter speeds selected at auto settings-- the camera is selecting shutter speeds that are too low leading to motion blur).

 

[havoc315]

This is the first real test I've seen of the A7s high ISO in still images.

http://www.photographybay.com/2014/04/10/sony-a7s-iso-comparison/

Not surprisingly, you can get away with very high ISO when you downsize the image.
But when looking at large size images, the performance looks pretty similar to other full frame cameras.
Everything looks nearly perfect up to around ISO1600-3200. 6400 looks okay... even 12800 doesn't look horrible. But above 12800, things start to turn ugly very fast in full size.
The test only goes up to 102000, which is pretty unusable unless totally downsized. (Which leaves to the imagination just how horrible 204000 and 406000 will be).

Does look better than my A99, which isn't surprising. (Larger pixels, no SLT mirror blocking light, and no loss of pixels for on sensor phase detection). It looks about one stop better than the A99. (The 25600 on the A7s looks a bit like the 12800 on the A99... The 6400 on the A7s looks like 3200 on the A99). I haven't seen enough of the A7 or A7r, but appears to only be marginally better than those two cameras.

While this camera may indeed be great in low light, it doesn't appear to be a massive technological breakthrough and low light miracle. At best, it seems like a small incremental evolutionary improvement over the last crop of full frame cameras.

Truthfully, for now, a "real" dSLR/SLT remains better for low light photography due to the availability of faster lenses (and more stabilized options as well). Unless you plan on adapting lenses for the A7s.
In low light,
A Canon 6d with a 1.2 prime lens will greatly outperform The A7s with a 1.8 or 2.8 prime lens.
A Canon 6d/Nikon d610/Sony A99 with a 24-70 2.8 lens will outperform the A7s with a 24-70 f4.
A Sony A99 with a stabilized 50/1.4 will outperform the A7s with an unstabilized 55/1.8.

The main reason to buy any of the A7 series, is really portability. Great full frame performance, very close to dSLR full frame, in a much more portable package. In exchange, you get slower native lenses. (Can apply faster lenses with an adapter but then you lose the size advantage). You get slower autofocus (though potentially more accurate). Slower burst rates. Less direct controls (a natural trade off to make the body smaller). Shorter battery life (the price of a small body and EVF).

To me, none of these trade offs are deal breakers. But likewise, a slightly smaller body isn't a major attraction to me, to lead me to the A7 series.

I suspect some day, almost all advanced cameras will be mirrorless. Western markets have been slow adopting the technology, but Western markets are no longer the only luxury consumers. Mirrorless is being rapidly adapted in the East, it's just a matter of time before it flows more to the West.
But I suspect Canon and Nikon, when they do take the dive into mirrorless more deeply, will make some dSLR-sized mirrorless cameras that work with their current lenses. Such a lineup may have less trade offs, but won't gain much portability.
Meanwhile, Sony had a small customer base of "big dSLR" users, so they have created a smaller mount system. Remains to be seen whether they will continue developing their dSLR-style mount, the A-mount.

 

[HPS3]

One bad thing with the A7 are the sensor reflections. Of you like to take photos with the sun or night lights you might want to go the D610 route. When I compared the A7 to my K5IIs at 6400 in raw it wasn't a huge difference. At normal viewing on a computer screen I don't think it's really noticable.