Sony Fs700 Vs A7s

Demistate

Demistate

Veteran
I just got my a7s in today and i'm running some tests to see how it differs from the Sony FS700. As I find more things i'll update this first post:
  • Rolling Shutter is worse on the A7s in Full Frame mode.
  • Slog2 on the A7s records in TV safe levels (16-235) while slog2 on the Fs700 records to the entire RGB range (0-255) I'm still investigating this.
  • Cine1 and Cine 4 will record to the entire RGB range on both cameras. (0-255)
  • If the viewfinder shows Zebras (even at 100+), there is still detail behind there in a Raw photo.
  • Don't shoot still photos with the Slog Profile. It puts the camera in ISO3200, and way underexposes the image requiring you to bring it up in Adobe Camera Raw by about 2.5 stops. This Looks worse than the same picture auto exposed at ISO 100, no picture profile
  • I bought a 67mm Tiffen Variable ND for shooting outdoors and maintaining a 1/50th of a second shutter speed (you can't do 1/48th like the FS700 :~() It is not enough ND to keep clouds from peaking on a sunny day (100+ zebra) even with max ND in slog 2 mode.
  • Knee and Black Level do not affect the image in Slog2 on the A7s. This is also true for the FS700.
  • Black Gamma does however affect Slog2 Profile in the A7s. Mostly the lower 40% of the signal
  • 24P and 30P modes are downsampling from the full 4K sensor, 60P and 120P do not, and will have aliasing.
  • I'm seeing a green color cast in shadows on the A7s, while I see a magenta color cast in the shadows on the FS700. This explains why Film emulation luts that contain both Slog2, and Slog2 -- FS700 profiles are almost the same but have more/less magenta in them.
I have my own hosting space where i'll post raw videos for you to critique my work. I welcome it.
 
Last edited:
Rolling Shutter Test:

I mounted the A7s on top of the Fs700 with a ball shoe mount and whipped the camera around on a tripod left and right.

fs700_vs_a7s_rs.gif


My verdict: a7s has a little more rolling shutter than the Fs700, but it's nothing to really get worked up about. I think people notice it more because the unit is so light so it's more prone to happening. Also as a nice side effect you can see the 28mbps codec falling apart at the frame I chose, while the 50mpbs XAVC s codec holding up a lot nicer.


Download my samples

Test setup: Sony Fs700 V3.0 firmware with Metabones II adapter and 24-105 Canon F4/L lens -- IS on -- Cine 4 profile -- AVCHD FX (28mbps)
Sony A7s 1.0 Firmware with 24-70 F4 OSS lens -- OSS On -- Cine 4 profile -- XAVC S 50mbps -- Full Frame mode
 
Last edited:
Codec information:

I took this mediainfo from one of the XAVC S files it recorded. its a .mp4 container, with h264 level 4.1, high profile, video at average 49.2Mbps. It can peak up to 60mbps in the clip I have. It may go higher but I'd have to look at other clips.

Audio is PCM, and by the definition of the ISO .mp4 standard not supposed to be in a .mp4 container. I could see some hardware players barfing on these files if you tried to play this file back on them.



Code: 
Format                         : XAVC
Codec ID                       : XAVC
File size                      : 286 MiB
Duration                       : 47s 47ms
Overall bit rate mode          : Variable
Overall bit rate               : 50.9 Mbps
Encoded date                   : UTC 2014-07-15 21:29:12
Tagged date                    : UTC 2014-07-15 21:29:12


Video
ID                             : 1
Format                         : AVC
Format/Info                    : Advanced Video Codec
Format profile                 : High@L4.1
Format settings, CABAC         : Yes
Format settings, ReFrames      : 2 frames
Codec ID                       : avc1
Codec ID/Info                  : Advanced Video Coding
Duration                       : 47s 47ms
Bit rate mode                  : Variable
Bit rate                       : 49.2 Mbps
Maximum bit rate               : 60.0 Mbps
Width                          : 1 920 pixels
Height                         : 1 080 pixels
Display aspect ratio           : 16:9
Frame rate mode                : Constant
Frame rate                     : 23.976 fps
Standard                       : NTSC
Color space                    : YUV
Chroma subsampling             : 4:2:0
Bit depth                      : 8 bits
Scan type                      : Progressive
Bits/(Pixel*Frame)             : 0.990
Stream size                    : 276 MiB (97%)
Encoded date                   : UTC 2014-07-15 21:29:12
Tagged date                    : UTC 2014-07-15 21:29:12


Audio
ID                             : 2
Format                         : PCM
Format settings, Endianness    : Big
Format settings, Sign          : Signed
Codec ID                       : twos
Duration                       : 47s 42ms
Bit rate mode                  : Constant
Bit rate                       : 1 536 Kbps
Channel(s)                     : 2 channels
Sampling rate                  : 48.0 KHz
Bit depth                      : 16 bits
Stream size                    : 8.61 MiB (3%)
Encoded date                   : UTC 2014-07-15 21:29:12
Tagged date                    : UTC 2014-07-15 21:29:12


Other
Type                           : meta
Duration                       : 47s 47ms
 
Last edited:
While you can't shoot at 1/48 shutter on the a7s, you can shoot 1/45. Which to my eye is better than shooting 1/50 on these DSLRs.
 
1/45 is available on ntsc model. And I agree ssvet, to my eye 1/45 is the sweet spot for nice motion blur on these dslrs.
 
Comparing the A7s to the FS700 without the 7Q is like watching Gravity without 3D glasses.

While the A7s might best a stock FS700, with the 7Q it's just in a different league.
 
Thanks for the detailed report

* To set 1/45s, you have to tell the a7s to use half-stop-increments instead of third-of-a-stop-increments
* Slog2 definitely goes beyond 235, but it starts at 23. It's not TV-safe levels, it's a high black point (which still sucks when your codec is 8-bit, but not that much: 12.6M colors instead of 10.6M, whereas full-range would be the usual 16.7M)
 
Demistate said:
I just got my a7s in today and i'm running some tests to see how it differs from the Sony FS700. As I find more things i'll update this first post:
  • Rolling Shutter is worse on the A7s in Full Frame mode.
  • Slog2 on the A7s records in TV safe levels (16-235) while slog2 on the Fs700 records to the entire RGB range (0-255) I'm still investigating this.
  • Cine1 and Cine 4 will record to the entire RGB range on both cameras. (0-255)
  • If the viewfinder shows Zebras (even at 100+), there is still detail behind there in a Raw photo.
  • Don't shoot still photos with the Slog Profile. It puts the camera in ISO3200, and way underexposes the image requiring you to bring it up in Adobe Camera Raw by about 2.5 stops. This Looks worse than the same picture auto exposed at ISO 100, no picture profile
  • I bought a 67mm Tiffen Variable ND for shooting outdoors and maintaining a 1/50th of a second shutter speed (you can't do 1/48th like the FS700 :~() It is not enough ND to keep clouds from peaking on a sunny day (100+ zebra) even with max ND in slog 2 mode.
  • Knee and Black Level do not affect the image in Slog2 on the A7s. This is also true for the FS700.
  • Black Gamma does however affect Slog2 Profile in the A7s. Mostly the lower 40% of the signal
I have my own hosting space where i'll post raw videos for you to critique my work. I welcome it.
Click to expand...

The output levels in XAVC are difficult to pin down, there are two issues here. One is that Slog2 by its very nature has a raised black level (25-30), on the FS700 it's still raised but lower (16ish). Secondly there are superbrights in there too, over 235 but you have to make sure the app you're using works with them. In AE (which imports them) only if you're running in float will you see these overbrights, if you are running in 8 bit then it is clipped.

Now internally the data is stored YUV, so my question about the XAVC is perhaps there is some mapping with the 16-235 not working properly, in that there are overbrights but the higher raised blacks compared to the FS700 are because the 16 isn't being normalised down to 0. If it was then we'd see the same range as the 700.

This is compounded because there isn't much support for XAVCs, and transcoding to something else might change the data.

So important to remember these RGB levels we are looking at can be created by either Quicktime or a supporting App depending how the data is read. So it can vary from platform to platform. The underlying source data is YUV.

I get the same with cine4, in that i see overbrights just like Slog2.

The waveform range is usually 0 to 109, so it would make sense that 100+ would still have detail. This way of monitoring exposure is the dark ages really.

If you're shooting RAW then putting Slog2 on the profile didn't make that difference for me when i looked, perhaps i got it wrong. 3200 appears to be the same as 800 in other profiles. I shot some stuff yesterday that i need to work though to check this.

I thought i saw the knee affecting Slog2 if set to manual?

Good work though, look forward to more

cheers
Paul
 
gene_can_sing said:
Comparing the A7s to the FS700 without the 7Q is like watching Gravity without 3D glasses.

While the A7s might best a stock FS700, with the 7Q it's just in a different league.
Click to expand...

Two reasons why i'm doing this:

1) I dont have a 7Q so I can't test against that.

2) I currently use a 5D Mark II, and an FS700 and wanted to see if this could be a tool I use instead of bringing a lot of gear to more remote locations.

I'll say this, I'm liking it more than the GH4 for the image, but the grippability isn't quite so good.
 
The Remote:

Sony sells a seperate IR remote for the A7s, but if you already own an FS700, you can use the supplied RMT-845 remote on your A7s. You just need to enable remote control in the menu. The remote works from the back and front of the camera.
_DSC0188.jpg



The only big downside to the FS700 remote is that it doesn't have a photo button on it, so you cant trigger stills, but you can trigger recording, or use it to navigate pictures/videos when you plug it up to a TV.

Raw photo here
 
DRO While taking Raw pictures:

DRO, or Dynamic Range Optimization does not have a tonal effect on the raw data recorded, However it DOES matter what DRO is set to when shooting RAW only.

Setting DRO to OFF gives you a result like this when you create a JPEG from Adobe Camera Raw:

DRO_OFF.jpg



Setting the DRO from 1-5 gives you a result like this when you create a JPEG from Adobe Camera Raw:

DRO_LV1.jpg



It appears that setting DRO ON, removes the Lens falloff correction. That correction is actually being burned into the RAW file itself when DRO is set to OFF. The default is AUTO and I see the lens profile correction being burned into most of my RAW files. I've also noticed if you try to use ACR's lens profile correction you'll see the corners actually become brighter than the center of the image because you are correcting for lens falloff twice.

I could not see a difference in tonality or dynamic range when DRO was enabled to various levels. It does not effect the processing of Raws, only the appearance of JPEG files created directly in camera.

Here is a .Gif showing the image in the same place so you can see the shading Comp turning on and off. You'll see more banding becuase Gif's are only 256 colors, and dithered:

DRO_animation.gif


You can check my Raw files and verify my results here

Test setup: Sony A7s, 24-70 Zeiss F/4 OSS, OSS On, M mode, manual focus, Picture Profile OFF, Creative Style OFF, Adobe RGB color space, Creative Style NEUTRAL, Shading Comp AUTO, Chro Aber. Comp. AUTO
 
Last edited:
Ailasing At different framerates:

I took the A7s and the FS700 and recorded them at various framerates to check for ailasing in the Slog 2 profile. Sony has been notorious for having "Highlight ailasing" when a really bright spot hits next to a very dark spot.

Take a look at the video:


My conclusions:

24P, look very good on both cameras, Highlight aliasing apparent on both, but the A7s has a small edge over the Fs700 on the downsampling from the 4K sensor.

30P, same as 24P on both cameras.

60P: A7s goes into a mode that looks worse than the 240P 2k pixel binning that the FS700 does. It's still very usable. FS700 stays the same quality as the 24P mode

120P (720P resolution): A7s only -- Is a zoomed in version of the 60P so the aliasing is even more pronounced. I'm not sure if this is APS-C crop, but it is NOT full frame.

240P: Fs700 Only -- Is still really good, but the noise is even more apparent as the internal electronics are not denoising the signal.


I've uploaded all the Raw clips on my site. I welcome your second opinion.
 
Last edited:
On a side note, I used the same shutter speed, Lens, and Fstop across both cameras. I kept ISO at native (2000 on FS700 and 3200 on A7s). The a7s seems to see more into the dark areas, this is why Sony rates the camera for ISO 3200.

The A7s is definitely more green at first glance, and I used no ND on either camera.
 
Thanks for the tests. Hopefully 4k out will remove the highlight outlining as there is no downscaling happening.
 
I have a Decklink 4K here at work and will try to test it if I have time tomrorrow. My PC was acting up today and just got replaced. The good news is that shiny part of the PS3 is in my office so I can capture that same exact spot.
 
Demistate said:
Two reasons why i'm doing this:

1) I dont have a 7Q so I can't test against that.

2) I currently use a 5D Mark II, and an FS700 and wanted to see if this could be a tool I use instead of bringing a lot of gear to more remote locations.

I'll say this, I'm liking it more than the GH4 for the image, but the grippability isn't quite so good.
Click to expand...

If you're looking for less equipment in remote locations and want a focus solution along with remote trigger for your sony FS700 then check out Cinegears. They have a wireless Follow Focus with a remote trigger for the FS700

Protector_preview.jpg
 
You must log in or register to reply here.
Back
Top