Samuel H
Major Contributor
I have MB Denoiser II, which doesn't have as many options but works great too. Even better, actually, depending on how much noise you want to take away. And when I got it it was cheaper too. Comparison here.
Clean Blacks
- Thread starter lyonfilms
- Start date
lyonfilms
Well-known member
If you Google Neat Video and "calibration chart" there is a link to Neat Video's website with links to different resolution versions of the calibration chart. Neat Video's instructions say to shoot the chart out of focus. Having done this and also watched a YouTube tutorial on how to work the "advanced" settings in Neat Video, it makes sense. I imagine some of it has to do with allowing the program to "see" the noise and not the grain of the paper you printed on, or the quality of the ink, etc. The calibration chart is a series of colored blocks that seem to provide the program a number of different areas to choose when sampling the noise profile of the particular scene.
I agree with the previous comments that Neat Video is a very, very good noise reduction/elimination program -- and its low cost makes it all the more attractive. Becoming familiar with the advanced settings will help you really dial in reduction for any given scene. However, I still firmly believe it is only one tool - and if it is overly relied up - it can't magically turn bad footage into good footage. It certainly has its limits and you will notice them on a big monitor if you push it too far. But overall, it can be fantastic.
Neat Video really does a phenomenal job for me. More often than not with my indoor gym shoots, I can use Neat Video to make the result both less noisy and sharper (in a good way) than when I started. Pretty amazing as I always felt cleaning up noise would necessarily result in a smoother image. So, I couldn't be happier, other than rendering is dog slow! Boy oh boy my MacBook Pro is seriously overmatched.
Love Neat Video! It crushes my system's performance so I try to save it towards the end.
However, someone mentioned a great tip about batching the job. So, you apply Neat Video to all your clips at once, then render and finally use the resulting files for your timeline.
However, someone mentioned a great tip about batching the job. So, you apply Neat Video to all your clips at once, then render and finally use the resulting files for your timeline.
The Irish TRex
Well-known member
So just to be clear...if I wanted to NOT use SLOG2 because I was hoping to limit the amount of grading that I'd have to get into in post due to a quick turnaround, and instead I shot a more baked in color, dark areas would be noisy on the A7s no matter the ISO? We shot something very similar to the original post where we lit someone with just one blue light at ISO800 in 4k and while they looked pretty good, the shadows and completely dark areas behind them and around them were incredibly noisy. I was shocked. Is this a characteristic of these types of cameras and their sensors? The A7s seems to have such good low light performance, but seemingly terrible no light performance. But then again, those first videos that came out when it was being first released which were shot in basically the dark, all looked really clean. Got me confused.
If you wanted to shoot a quick shot of someone at a desk with a single small light hitting his face only but keep all the background completely dark with no light, how would you do this without getting noisy in the areas with no light at all? Do you HAVE to use SLOG because I'm trying to shoot it and literally hand it over to someone without grading?
If you wanted to shoot a quick shot of someone at a desk with a single small light hitting his face only but keep all the background completely dark with no light, how would you do this without getting noisy in the areas with no light at all? Do you HAVE to use SLOG because I'm trying to shoot it and literally hand it over to someone without grading?
caseywilsondp
Well-known member
What picture profile did you shoot before? Also because there is no denoising going on on HDMI out, you'll get noise in shadow... It just happens. Cameras get around this by having internal noise reduction/mapping. HDMI out doesn't get this... It's the price you pay for the affordable cost of this system.
If you have to just give the footage over, could you supply a lut for them to drop on (it's really easy in most nles)
If you have to just give the footage over, could you supply a lut for them to drop on (it's really easy in most nles)
The Irish TRex
Well-known member
Thanks for the reply. And good point about the in camera noise reduction. Didn't think about that. Now it sorta makes more sense. I used the Shogun to record it in 4k. And I think I have it set to PP4 with some tweaks to really burn in a slightly more saturated look with deeper blacks. For this project I didn't have to turn it over to anyone but it will be shown at a local AMC movie theater for this festival. Down converting from 4k to 1080p seems to make the noise less obvious. But it's still fairly heavy. Hmmm. Wish there was a 4k route with the A7s that has noise reduction. My next shoot will be a quicker turnaround which is why I'm back on this thread trying to figure out things. With such great low light performance at high ISOs I guess I was just shocked to see it so noisy at really low ISOs on dark footage after going through the shogun.
lyonfilms
Well-known member
My friends Randolph Sellars and Michael Parisien spent some time with me at the studio last week continuing to poke around on this issue. We set up a human model in a few different situations to test a few of our own theories on how to best achieve "clean black shadows" or, in other words, underexposed areas that exhibit little to no noise. We also appreciate the effort that was put behind the 55 min video on the Wolfcrow system - there is a lot of great stuff in that video.
We experimented in SLOG2 + SGAMUT as well as CINE 4 + CINEMA out via HDMI to the Shogun in 4K.
I think we had the best results with CINE 4 + CINEMA, in terms of low noise in dark/shadowed areas of a frame (which included a person sitting on a chair lit in an extreme, dramatic side key light), when we pushed the exposure on the model's face to the edge so that we were raising the IRE floor of the shadows. Then, in post, we carefully graded the scene back down to proper exposure. The end result was that the scene where I "properly" exposed the skin had much noisier shadows/blacks than the same scene where I very carefully, using both the camera's zebras as well as the Shogun's waveform, exposed the skin to the edge of what we thought was recoverable in post (meaning no loss in texture, etc.,). This makes sense, but is tricky. Without "lighting" the shadows (and thus changing the nature of the shadow area) we just tried to give the sensor more light in those areas by opening up on the iris, since we were experimenting with a model in each shot we had to carefully balance how far we allowed ourselves to go in upping the amount of exposure in the shadows, because of course, it had the opposite effect of increasing exposure on the model's face.
We experimented in SLOG2 + SGAMUT as well as CINE 4 + CINEMA out via HDMI to the Shogun in 4K.
I think we had the best results with CINE 4 + CINEMA, in terms of low noise in dark/shadowed areas of a frame (which included a person sitting on a chair lit in an extreme, dramatic side key light), when we pushed the exposure on the model's face to the edge so that we were raising the IRE floor of the shadows. Then, in post, we carefully graded the scene back down to proper exposure. The end result was that the scene where I "properly" exposed the skin had much noisier shadows/blacks than the same scene where I very carefully, using both the camera's zebras as well as the Shogun's waveform, exposed the skin to the edge of what we thought was recoverable in post (meaning no loss in texture, etc.,). This makes sense, but is tricky. Without "lighting" the shadows (and thus changing the nature of the shadow area) we just tried to give the sensor more light in those areas by opening up on the iris, since we were experimenting with a model in each shot we had to carefully balance how far we allowed ourselves to go in upping the amount of exposure in the shadows, because of course, it had the opposite effect of increasing exposure on the model's face.
Last edited:
caseywilsondp
Well-known member
This makes sense as SLOG2/Sgamut is really stretching out the information for post work, so your shadows will have more information in them, and subsequently, more noise with slog2/sgamut. Glad you found your solution!