FS700 - Ugly Noise

Simon Shasha

Well-known member
I don't know if anyone else has seen this video, but the noise in some of the shots using 3200ISO (1:58), 4000ISO (2:40) is really prominent and just downright ugly. To be honest, I'm really hoping they made a typo with those ISO numbers...

LINK: https://vimeo.com/43021945

(Also, I didn't watch via Vimeo, I downloaded the original uploaded file).
 
I watched the area you mention at 158 but don't see anything objectionable for that sensitivity level. There is a flicker in the background but I can't fault the camera for that- looks like a practical out of phase.

Ugly? For finding a little noise at 3200 iso? Dang, you're a tough customer!
 
I think you're talking about the compression artifacting. Looks kind of like shifting pixels. I noticed the same thing on the FS100.

I can't speak directly to the FS700 but I can tell you that we did a bunch of tests here in the studio on the FS100 and found that it happens most when you force the camera to process a lot of information in a short amount of time. For example, raising ISO and overcranking while filming something with a lot of contrast will give you the noise I think you're talking about. Then, the raised ISO highlights the compression even more in certain areas (usually mid tones).

Sony's prosumer line of camera is super sensitive and gives you a pretty clean image for the price--at the expense of a few quirks to the more discerning eye (like compression noise).
 
I watched the area you mention at 158 but don't see anything objectionable for that sensitivity level. There is a flicker in the background but I can't fault the camera for that- looks like a practical out of phase.

Ugly? For finding a little noise at 3200 iso? Dang, you're a tough customer!

A little? That's a lot of noise in my opinion. That 3200ISO looks as bad as the 3200ISO coming off my 60D. I'm not a tough-customer - it is simply aesthetically ugly - especially for $8000 (or $9900 here in Oz) - and the reason I brought it too attention, and the reason it bothers me, is because I am comparing it to Philip Blooms FS100 ISO tests from his "2011 Christmas Shootout":

https://vimeo.com/34637684 (FS100 ISO test starts at 22:30)

In Philip Bloom's recent FS700 review, he stated that the FS700 is virtually identical to the FS100's ISO performance (with the FS100 having a very slight lead due to lower mega-pixel sensor) - but the video I linked in my original post above makes the FS700's ISO performance look awful.
 
I think you're talking about the compression artifacting. Looks kind of like shifting pixels. I noticed the same thing on the FS100.

I can't speak directly to the FS700 but I can tell you that we did a bunch of tests here in the studio on the FS100 and found that it happens most when you force the camera to process a lot of information in a short amount of time. For example, raising ISO and overcranking while filming something with a lot of contrast will give you the noise I think you're talking about. Then, the raised ISO highlights the compression even more in certain areas (usually mid tones).

Sony's prosumer line of camera is super sensitive and gives you a pretty clean image for the price--at the expense of a few quirks to the more discerning eye (like compression noise).

It's both compression artefacts and high ISO noise. Looks like hell. I know for a fact, from using it time and time again, the FS100 gives a very clean image at 3200ISO. Given what Bloom said about the similarities of the FS700 and FS100's ISO performance, there is no way that shot above was at 3200ISO. It seriously looks like 3200ISO from a 60D.
 
It's both compression artefacts and high ISO noise. Looks like hell. I know for a fact, from using it time and time again, the FS100 gives a very clean image at 3200ISO. Given what Bloom said about the similarities of the FS700 and FS100's ISO performance, there is no way that shot above was at 3200ISO. It seriously looks like 3200ISO from a 60D.

It does look like hell. Its happening when shooting high speed shots without enough light. Obviously high speed (slo mo) needs loads of light then add the compression limitations involved in getting that slo mo with Avchd bitrates and you get a crappy image, so the person who shot the footage upped his iso in order to get exposure for some of those high speed shots. Just the nature of shooting high frame rate AVCHD slo mo: you need lots of light and a high bitrate to avoid the noise.
 
Last edited:
It's both compression artefacts and high ISO noise. Looks like hell. I know for a fact, from using it time and time again, the FS100 gives a very clean image at 3200ISO. Given what Bloom said about the similarities of the FS700 and FS100's ISO performance, there is no way that shot above was at 3200ISO. It seriously looks like 3200ISO from a 60D.

You must have a better FS100 than me because mine looks like poo at 3200. In fact, these samples are right on par with how mine performs when you don't 'spoon feed' it a perfect image.

A thought: rather than spending $10K on a prosumer camera, why don't you spend $13K on the Red Scarlet? The difference in image quality, peace of mind, flexibility and amount of time you won't spend worrying about weird image noise is worth far more than the extra $3K.
 
Never mind... forgot the Scarlet doesn't overcrank. And most people are buying the FS700 because they're enamored by slow motion.
 
It's the person behind the cam , how you light etc.. not the camera

No, it isn't the person behind the camera,it's the camera.

Because it's a PRE-PRODUCTION camera. The noise is not indicative of the final image, is what I was told when I had the FS700. Hence why you haven't really seen any lowlight footage from the camera yet.

This goes back to what I was saying about people keeping footage under wraps (or they usually do) because otherwise it'll be considered the final image, even though it isn't.
 
Scarlet at the end will cost you more than 13k with media etc. And dosent even produce a full super 35 censor 120fps image let alone 240 fps
 
Also, consider this:

I can't think of a major camera that's been released in years that didn't have people here complaining about SOMETHING they saw in early posted footage as (most often) noisey or extremely flawed in some other way.

Simultaneously, or a bit later, there is absolutely GORGEOUS footage posted from said camera - in other hands, and/or under other conditions or settings.

My advice is: Don't look at the worst footage you find from a particular camera (especially pre-release) and then worry about what "flaw" is making it look so bad.

Instead, look at the absolute best footage you can find from a particular camera - and work at it until you are able to achieve that same level of quality.
 
@ Kholi, It is the person IMO due to the fact that he or she did not light properly for the scenes or focus, and so on you cant just turn on a camera and expect it to look great you have to put work into any camera to get the best image out of it hence my comment not the camera but one who is behind it ..
 
But on another note ,I do think we need some more 24p footage! I have seen enough slow mo to last a lifetime LOL!
 
A little? That's a lot of noise in my opinion. That 3200ISO looks as bad as the 3200ISO coming off my 60D. I'm not a tough-customer - it is simply aesthetically ugly - especially for $8000 (or $9900 here in Oz) - and the reason I brought it too attention, and the reason it bothers me, is because I am comparing it to Philip Blooms FS100 ISO tests from his "2011 Christmas Shootout":

https://vimeo.com/34637684 (FS100 ISO test starts at 22:30)

In Philip Bloom's recent FS700 review, he stated that the FS700 is virtually identical to the FS100's ISO performance (with the FS100 having a very slight lead due to lower mega-pixel sensor) - but the video I linked in my original post above makes the FS700's ISO performance look awful.


Philip Bloom said that overall aliasing was the same, not that the total performance was the same. it was an ambiguous statement and so far he has avoided clarifying it in the comments section when people seem to be asking for more information.

Philip Bloom did not do his usual low light tests with people, a more telling subject that a moving booble head, but Andy Shipsides of AbelCine did. https://vimeo.com/42206660 In that test the FS700 AVOIDED many of the color banding and aliasing low light issues with the FS100 - exactly the opposite of what people seem to be saying on this thread.

When looking at compressed 24p footage with no download there were also examples in Andrew Young's https://vimeo.com/40369782
 
Last edited:
But on another note ,I do think we need some more 24p footage! I have seen enough slow mo to last a lifetime LOL!

+1

I have been extremely disappointed by all the preproduction reviews since they fouc on slo mo and do not do the usual normal speed tests. There was Frank's Iron Man which was great filmmaking but not a test, and was heavily graded.

.
 
Last edited:
I don't know who shot this clip being referenced here but I was struck that the liner notes say the scene was shot "by" an FS700.

thats about what we are looking at these days with all the pixel peeping. people credit the cam not the shooter.

hey- cameras can't shoot **** folks ;-)
 
The lighting in that scene looked like HELL. It was so bad. Not sure if any camera can make that light scenario look good. Plus with slow motion, you need more light because of the increased shutter speed. Seems like white balance was also way off as it was extremely green. Not exactly the best shot to judge a camera with.

Would the Scarlet really do a better job in that situation? I thought the REDs are not a low light camera and really need optimal lighting.
 
I remember a guy who was involved with the "VS" video said slowmo produced more noise than realtime speeds at same gain.
But the noise is really bad. I ofte used my FS100 with 18 dB of gain and it looked just great! Better than a GH2 at ISO 400!
The FS700 noise I´ve seen yet really has ugly blocky and heavy noise just like an old canon DSLR.

I already ordered the 700 so I have no interest to bash it.

Edit: Look at the noise at the slow-mo at about 1:30 and then the 24p shot right after that with the same ISO. Looks much much better!
 
Back
Top