A fundamental aspect of electronic cameras is that the bulk of the noise comes from the sensor. So the amount of noise in the final image is mostly a function of the amount of light you put on to the sensor v the noise the sensor produces (which is more or less constant). This is known as the signal to noise ratio, often abbreviated to SNR.
Whether you use S-Log3 or S-Cinetone, even though the base ISO number the camera displays changes the sensitivity of the camera is actually the same, after all we are not changing the sensor when we change modes. In fact if you set the camera to dB you will see that in custom mode the base for both S-Cinetone and S-log3 (and every other gamma curve) is always 0dB.
All we are changing when we switch between S-Cinetone and S-Log3 is the gamma curve – which is a form of gain curve. The base ISO number changes between S-Log3 and S-Cinetone because if you were using an external light meter this would be the number to put into the meter to get the “correct” exposure, but the actual sensitivity of the camera remains the same.
First let’s think about what is happening at the base ISO of each if we were to use an external light meter to set the exposure…..
If we shoot at S-Cinetone and use the 320 ISO value in the light meter the aperture will be a little over a stop more open than if you shoot with S-Log3 and use 800 ISO for the light meter. So when using S-Cinetone at the base ISO there is a little over twice as much light going on to the sensor compared to S-Log3 at the base ISO and as a result the S-Cinetone will be much less noisy than the S-Log3. Not because of a sensitivity or noise performance difference but simply because you are exposing the sensor more brightly.
And if we use the SAME ISO value for S-Cinetone and S-Log3?
So now think about what might happen if you were to put 400 ISO into your light meter and use the values for shutter and aperture the meter gives and shoot with either S-Cinetone or S-Log3 using the very same aperture and shutter settings so that the same amount of light is hitting the sensor for both. The result will be that the amount of noise in the resulting image will be broadly similar for both and the same would happen if you were to use, let’s say, 4000 ISO (assuming you switch to high base for both).
There will tend to be a bit more noise in the S-Log and CineEI at the default settings, because by default NR is turned off in CineEI. But with the same in camera NR settings, again both the S-Log3 and S-Cinetone will have very, very similar noise levels when the sensor receives the same amount of light.
What about when there isn’t enough light?
So – when you are struggling for light, both will perform similarly from a noise point of view. BUT where there may be a difference is that with S-Cinetone all your image processing is done before it is compressed by the codec and what you see in the viewfinder is what you get. With S-Log3 the “underexposed” image gets compressed and then you will need to process that in post and when you add your post corrections this will be to the recorded image + compression artefacts so there will always be a lot of uncertainty as to how the final image will come out.
Personally I tend to favour S-Cinetone for under exposed situations. Generally if it’s under exposed dynamic range isn’t going to be an issue. S-Cinetone also spreads what image information you do have over a greater range of code values than S-Log3 and this may also help a little. But there is no right or wrong way and any differences will be small.
Tag Archives: S-Cinetone
Chroma Key and S-Cinetone
A question poped up today asking about how to expose S-Cinetone when shooting green screen.
The answer is really quite simple – no differently to how you would expose S-Cinetone anywhere else. But, having said that it is important to understand that S-Cinetone is a bit different to normal Rec-709 and this needs to be considered when shooting for chroma key or green screen.
S-Cinetone’s highlight roll off and shoulder starts much lower than most “normal” rec-709 curves. From around 73% the gamma curve changes and starts to compress the levels and reduce contrast. In the shdows there is a variable toe that increases contrast at lower brightness levels. The nominal “normal” brightness levels are also lower, all part of the contemporary film like look S-Cinetone is designed to give. A 90% reflectivity reference white card would be exposed at approx 83% instead of the more normal 90% (if you were using a light meter you should end up with a 90% white card at 83IRE). A white piece of paper will be a bit brighter than this as printer and copier paper etc is designed to look as bright as possible, typically printer paper comes out around 3 to 5% brighter than a proper white card.
The lower start to the highlight roll-off means that if you place skintones around 70% the brighter parts of a face will be affected by the rolloff and this will make them flatter. Expose skin tones at 60% and the face will be more contrasty and in my opinion look better. Although darker this would still be well with the “normal” exposure range for S-Cinetone so you will not have excessive noise and it will still key well.
S-Cinetone would be considered correctly exposed when a 90% white card is exposed between 78% and 88%. This is quite a wide window and is due to the way S-Cinetone is designed to give differening contrast levels simply by exposing a touch brighter or darker. The variable toe and shoulder mean that exposing brighter will make the image flatter and exposing darker more contrasty. Exposing as you would with normal Rec-709 levels with a white card at 90% will place skintones rather higher than “normal” and they will appear very flat. So either expose so a white card falls in the 78-88% window or use a calibrated monitor to observe how the skin tone look and be careful not to overexpose them.
Your greenscreen should be between 40IRE and 60IRE for a good clean key, I normally aim for 50IRE with S-Cinetone, but provided you don’t go below 40IRE or above 60IRE you should be good.
How Do You Expose S-Cinetone with the FX6/FX3/A7SIII
Lots of people have been asking about how to expose S-Cinetone, whether with the FX9, FX6, A7SIII or the FX3.
The short answers is: So that it looks nice!
S-Cinetone has a variable toe and knee. So exposing it brighter results in not only a brighter image but also an image with flatter skin tones and less shadow contrast, overall looking more video like.
Exposing a little bit darker results in a more contrasty film like image. Faces and skin tones have more texture. There is no one optimum exposure level. A white card could be anywhere between 78% and 88% depending on the look you want. Typical skin tones will vary from between anywhere between 55% and 75%.
Personally I like the way S-Cinetone looks when it’s exposed with Skin tones at around 63% and white at around 81%.
See the video I on S-Cinetone on the FX9 for more details as it all applies equally to the FX9 and FX6 as well as the A7SIII and FX3. The only small difference is that the base ISO’s are a little different between each camera.
Broadcast safe and direct to air using S-CinEtone and the FX9
S-Cinetone is a combination of both a gamma curve and colour matrix. So you can mix different gammas with the S-Cinetone matrix to get the same colour palette but with different contrast and brightness ranges.
The FX9’s default gamma curve is labelled “original” in the camera’s paint settings. This gamma curve is the closest to the look that the s709 “Venice” LUT provides. One important thing to note about this gamma curve is that it records using full range, so it can exceed 100% and goes all the way up to 109%. This means that it may not be suitable for some broadcast applications where the footage will go direct to air without any grading or post production adjustment.
If you do need a broadcast safe output then you should consider using Hypergamma 1 with the S-Cinetone matrix. The look is very similar to the original matrix, but Hypergamma 1 never exceeds 100% so it is “broadcast safe”.