Noisy S-Log and night shoots

berlinfool

berlinfool

Well-known member
I'm shooting on low light on a rave Saturday and I want to use the Q7 and record RAW 4K to ProResHD.
I'm scared of the noise in the dark areas. What's the solution? Crush the blacks in post?
 
yeah I've noticed a lot of noise in both s-log and rec709 4k2hd. The problem is you dont always have enough light to overexpose without cranking up the iso. Wish there was a way around it, but even properly exposed images I see quite a bit of noise in the shadows/dark areas. De noising does work well but would be nice to not have to do this step.
 
If you don't mind me asking, how do you overexpose a dark environment when he states that it is going to be shoot in low light?
To overexpose something you need light, no?

The only thing I can think of is using a fast lens and and try to get as many useful shoots as possible. Then as mentioned before use denoising plugins (denoiser II or neat video) when needed.

Correct me if I'm wrong but:
Also your recording internally at the same time on a SD-card, there should be less noise from the internal footage duo to the treatment it gets, so you do have a backup if some of the material from 7Q is to noisy.
 
Last edited:
freshTEA said:
If you don't mind me asking, how do you overexpose a dark environment when he states that it is going to be shoot in low light?
To overexpose something you need light, no?

The only thing I can think of is using a fast lens and and try to get as many useful shoots as possible. Then as mentioned before use denoising plugins (denoiser II or neat video) when needed.

Correct me if I'm wrong but:
Also your recording internally at the same time on a SD-card, there should be less noise from the internal footage duo to the treatment it gets, so you do have a backup if some of the material from 7Q is to noisy.
Click to expand...

Or get an A7s low light is one of the outstanding qualities in this minute cam, and soon we will have 4K (UHD IF YOU WANT TO BE TECHNICAL).
 
I found that 4k2hd or 4k24k is more noiser than raw file, from what i tested, especially when under expose it can be more worst.
I really think that fs700+7Q require lighting very much, without lighting well, the noise can be scary sometime. But with proper lighting it is just insane quality and i just love it.
Sometime not really overexpose then can avoid noise, it is really depend on situation, what mood/feel u want.
At outdoor night scene, most of the time i think only main subject need been seen more clearly, and if i need dark surrounding feel, i will just light on my subject only on normal exposure or sometime even underexpose and bring it up at post (as long as subject being light bring up in post is not a problem especially with raw) and noise in dark area i will just use denoise at shadow only.
 
olof said:
Or get an A7s low light is one of the outstanding qualities in this minute cam, and soon we will have 4K (UHD IF YOU WANT TO BE TECHNICAL).
Click to expand...

Hihi already have it in my arsenal :D

To be honest it was rhetorical question I guess..
When someone ask's a question like that it doesn't make any sense to answer with: Overexpose the image.
In a controlled situation, yes. But if the condition is going to be low light then what can he do to get best possible image with what he has.
 
I think you may be misunderstanding. By overexpose the image, what I mean is use as much gain as you can to avoid placing important detail in the shadow areas. Many people make the mistake of thinking that noise comes from increasing gain, but really noise comes from underexposing, and in most cases increasing gain, especially with SLOG2 where you can easily offset exposure in post, will actually help you reduce noise. With a camera like the FS700 that has adjustable analog gain amplifiers, in a fixed lighting level scene, it is likely better to run the camera at 18db gain and get a bright image than it is to shoot at 0db and get a dark image. This may seem obvious, but I'm amazed how often people misunderstand this. Obviously you may have to balance this if there are also bright highlights in the scene you want to retain detail in.

Higher gain modes are probably slightly noisier than 0db, but most people assume they are much noisier because they only use those higher gain modes in situations where they don't have enough light and using 0db would be very dark. And then when people use higher gain, they often are afraid to go too high because they think more gain equals more noise, and they don't use enough gain so the images are still somewhat underexposed and noisy but they would be better off increasing gain and bringing it down in post to minimiize noise.

Your question seems to imply that you might think SLOG2 is noisier than other gamma modes, but that is another common mistake. If you expose the same at the same gain it will have the same amount of noise as other gammas once you match them in post, only it may have a bit more DR preserved compared to some gammas if you want to use it, and SLOG2 is friendlier for LUTs and for log adjustments such as easy exposure offsets in post.
 
Last edited:
nyvz said:
I think you may be misunderstanding. By overexpose the image, what I mean is use as much gain as you can to avoid placing important detail in the shadow areas. Many people make the mistake of thinking that noise comes from increasing gain, but really noise comes from underexposing, and in most cases increasing gain, especially with SLOG2 where you can easily offset exposure in post, will actually help you reduce noise. With a camera like the FS700 that has adjustable analog gain amplifiers, in a fixed lighting level scene, it is likely better to run the camera at 18db gain and get a bright image than it is to shoot at 0db and get a dark image. This may seem obvious, but I'm amazed how often people misunderstand this. Obviously you may have to balance this if there are also bright highlights in the scene you want to retain detail in.

Higher gain modes are probably slightly noisier than 0db, but most people assume they are much noisier because they only use those higher gain modes in situations where they don't have enough light and using 0db would be very dark. And then when people use higher gain, they often are afraid to go too high because they think more gain equals more noise, and they don't use enough gain so the images are still somewhat underexposed and noisy but they would be better off increasing gain and bringing it down in post to minimiize noise.

Your question seems to imply that you might think SLOG2 is noisier than other gamma modes, but that is another common mistake. If you expose the same at the same gain it will have the same amount of noise as other gammas once you match them in post, only it may have a bit more DR preserved compared to some gammas if you want to use it, and SLOG2 is friendlier for LUTs and for log adjustments such as easy exposure offsets in post.
Click to expand...

Thank you nyvs for the post, appriciate the knowledge. I might have been a little vague in my post but I didnt mean so much that slog 2 is noisier. What I meant was that I was under the impression that the fs700 do treat the footage from the sensor before it reaches the SD card where it does reduce the noise. The signal the 7Q recives have not recieved the same treatment.
 
freshTEA said:
Thank you nyvs for the post, appriciate the knowledge. I might have been a little vague in my post but I didnt mean so much that slog 2 is noisier. What I meant was that I was under the impression that the fs700 do treat the footage from the sensor before it reaches the SD card where it does reduce the noise. The signal the 7Q recives have not recieved the same treatment.
Click to expand...

Ah ok, from my brief tests judging 4k2hd vs AVCHD, the noise is certainly more visible in 4k2hd, but I would say it that the noise is essentially the same except the shadows noise in 4k2hd is simply preserved due to the compression scheme and the AVCHD shadow noise is less visible but only because the high frequency detail in the shadows is muddier after compression quantizes away the least perceptually necessary detail, shadow noise. One consideration is what your final output is. Unless you are shooting for a theatrical release, if you are shooting for web, broadcast, bluray/dvd distribution there is a very good chance that your final product will at some point be recompressed even more heavily than 24Mbps AVCHD and much of that shadow noise that is preserved by prores will be lost, especially since a lot of that noise will be become less visible after grading since the shadow curve will likely be compressed in grading, reducing shadow range contrast and therefore noise contrast in those areas.
 
Last edited:
You must log in or register to reply here.
Back
Top