This question comes up a lot, which gamma curve will give me the best results on a low key (dark) shoot. A common example would be “Should I use Gamma 2 which is brighter or should I use Gamma 5 which is darker but with a bit of gain”? or “Which is better to use in low light, normal gamma or log gamma”?

Log gamma (SLog, LogC etc) is great for capturing scenes with a large dynamic range, but it comes at the price of compressed highlights and overall less data per stop of exposure. Standard gammas have a lower dynamic range, but as a result the amount of data you are recording per stop of exposure is greater. It’s also vital to remember that if you do shoot in the dark with log you still need to expose correctly keeping middle grey and white at the correct levels in order that you keep mid tones and skin tones out of the heavily compressed part of the curve for the best results, so your images may look very dark and very flat. This can make focussing a big challenge (Top Tip – Use a large waveform monitor to check focus, as you go in and out of focus you should see the fine details in the waveform increase in amplitude as you become more focussed).

If you don’t need 13 or 14 stops of range, then use another gamma, don’t use log. Again with any of the extended range gammas like Cinegamma or Hypergamma watch where you place your skin tones, your exposure levels with these types of gamma curve are normally a little lower than with standard gammas (typically -0.5 to -1 stop).

A basic principle to understand with any video camera is that the actual sensitivity of the camera is purely a function of the sensor. Gain or raising the ISO is simply amplifying the signal off the sensor and results in more noise as the noise gets amplified too. Just like a music amplifier where you can make the music louder by turning up the volume (adding gain) you can make your pictures louder (brighter) by adding gain or raising the ISO. Listen to the music amplifier at high volume levels and what do you hear in the background? More hiss and noise, it’s the same with your video camera, more gain = more noise.

Gamma is also a form of gain providing different amounts of gain for different sensor output (brightness) levels to achieve specific recording and camera output levels, but it’s still gain (it can sometime be negative gain). The nearest you have to a gamma control on an audio amplifier would be an equaliser or bass and treble tone controls. Turn up the treble control on an audio amp and again you will here more hiss and noise, it’s nothing more than a selective gain control that shapes the sound you hear. Video camera gamma is very similar, it shapes the tonal range of the images you see.

So – and this is the bottom line: With any given gamma, combined with any given ISO or gain level, to achieve any particular brightness of output, for the same input level the total system gain (gain + gamma) is the same for that particular point in the scenes brightness range. So the noise level for any given brightness of output will be almost exactly the same whether you mix ISO “A” with Gamma “B”  or ISO “B” with Gamma “A” because the combined gain of A + B is the same as B + A.  At the end of the day the real sensitivity is governed by the sensor and anything else that makes the image brighter is gain, either regular “gain” or gamma gain.

So, choose a gamma curve that can deal with the dynamic range that you need and no more, don’t waste recording data on unused dynamic range. That way when you grade or do any post production noise reduction you have more data per stop to work from and that will help with the final image. If you do use log or a high range gamma (Hypergamma, Cinegamma etc) watch your exposure levels, the same rules apply in the dark as in daylight, you don’t want to over expose any skin tones as they won’t grade nicely.