What ever happened to signal to noise ratios the bar at which all video cameras were judged, we talk about native sensor ISO (International Standards Organization) these days, total confusion. Recent 4K/HD cameras give you a choice of either decibels (dBs) or ISO for gain, the problem with ISO is its far from standard.

Back in the days of film your film speed was measured as ASA (American Standards Association), as an example you had T-MAX 100 ASA and T-MAX 400 ASA film. The T-MAX 400 provides greater under-exposure latitude that results in more detail in the shadow areas. However, this is achieved at the expense of “smokey blacks”. Are we starting to see a similarity to filming in LOG which is where all this terminology originally stems from. We are so intent in making video footage look cinematic we have built a science around it.

During the 1960s news reports were usually filmed on 16mm film rated at 400 ASA allowing a faster shutter speed and footage in lower lighting conditions, it was slightly more grainy than 100ASA film.

Todays camera manufacturer sets the cameras native ISO (equivalent to film speed)  taking a few points into account…

1. Noise and grain

2. Optimum Dynamic Range (12-14 stops)

3. Detail

The sensor is at its best when all 3 are optimised and given its “Native ISO” but that may not correspond to 0dB.

In the world of ISO…if your camera has an ISO of 200 that should be the same on all cameras but in the real world that only works for the same model and make of camera.

Take the Panasonic DVX-200 its native ISO is rated at 500 which is the lowest ISO on the camera therefor it also equals 0dB. For every 6dBs you double the ISO, so 6dBs = 1000 ISO and 9dBs = 1500 ISO with 12dBs = 2000 ISO which is fine for the DVX-200 but if you have a Sony PXW-FS7…

The Sony PXW-FS7 has a native ISO of 2000 so you can see where the first problem lies there is 1500 ISO of a difference between the Panasonic and the Sony.

My gripe is that in the film days 400ASA was a set film speed that many makes of camera could use and exposing via a light meter would be standard throughout, today you can’t apply the same rule for ISO as the sensor sensitivity differs between each camera model.

Take the two cameras above, setting both cameras to 0dB at f4 will still give you vastly different exposures to each other. You would have to set the DVX-200 to 12dBs to match the Sonys native ISO of 2000, but the noise and grain would be a lot more visible on the Panasonic.

So is 0dB the same on all cameras ? 0dB should be the quietest point in a cameras sensor giving the best possible picture quality.

Why do we use ISO at all ? The camera’s ISO setting is its sensitivity to light. The higher the ISO, the more sensitive it is. This is measured according to international standards. Each ISO setting is double the one before, if you increase the ISO from 100 to 200, you double the camera’s sensitivity.

XDCAM User “Sony’s native ISO rating for the FS7 of 2000ISO has been chosen by Sony to give a good trade off between sensitivity, noise and over/under exposure latitude.” and “If you were to use SLog2 or SLog3 with the camera in custom mode and not use the native ISO by adding gain or changing the ISO away from 2000, you will not get the full 14 stop range that the camera is capable of delivering.”

Now for more confusion the Sony FS7s native ISO is 2000 but you have ISO 800, 1000, 1250 and 1600 before you hit the native 2000 ISO, most FS7 users I know prefer ISO 800 in Cine EI mode. You have to make up your own minds on ISO, some so called scholars will tell you never to go below the native ISO, thats just nonsense.

Use your eyes it all depends on the lighting conditions and what overall effect you want to achieve, the BBC prefer not to have noise levels above 6dB but how can they tell these days with cameras like the Sony FS7 !

To me the whole ISO system stinks with deliberate confusion and is far from standard, I spoke to 2 engineers and three competent cameramen today who were all confused by this topic so what chance does everyone else have.

I have some simple tips as follows…

1. Use your eyes, can you see grain in your picture, make sure you have a good monitor and a decent viewfinder.
2. Use -dBs if available to avoid using the ND filter as even ND filters can introduce their own noise. (Its down to how perfect they are made and are they dust free). TIP. When changing lenses make sure you have the ND filter out (clear position) to avoid dust getting onto the ND filter.
3. By using 0dB you may affect the full dynamic range of the camera i.e. 12 stops rather than 14 that in my opinion is a sacrifice I prefer to make to have noise free footage, remember the full Dynamic Range (DR) is only used in LOG mode once again something a lot of us do not bother with in day to day filming.
4. Most professional video cameras give you the option for dBs or ISO, personally I am happier to use dBs to control gain.
5. Remember dBs and ISO give you the same result its only useful to use ISO if you need to get all of your cameras dynamic range by sticking to its native ISO.

One of my readers David Heath has some interesting input…

“You have to go right back to basics – to sensor level – to really understand where the term “base ISO” comes from, and why it’s so important.

Start from total blackness, then gradually increase the light level – at some point the sensor will give a just discernible output. Carry on increasing light level and the output will correspondingly increase until it reaches a saturation point – no matter how much more light is input, the output won’t increase any more. It follows that the difference between these min/max points is the basic dynamic range of the camera, normally expressed logarithmically in stops.

Note this is AT SENSOR LEVEL, and is independent of such things as camera gain etc. Hence there is ONLY one grey scale range that can give this, so taking absolute brightness, f stop etc into account the camera can only have one ISO value where this is true. This defines the “base ISO”.

All “ISO value” really means is that for a given illumination, at a given shutter speed, it tells you what to set the f stop (more correctly, T-stop) to for correct exposure.

It’s not really true to say “0dB should be the quietest point in a cameras sensor giving the best possible picture quality”. It normally corresponds to MAX DR, which is not necessarily the quietest in terms of S/N. As example, on cameras before the complications of log etc it was typical to have a position for negative gain – which would give a better noise performance than 0dB. But at the expense of dynamic range, normally highlight handling, which is why the amount was typically restricted to only -3dB, or maybe -6dB on pro 2/3″.

So yes – the quote from XDCAM user is perfectly valid. By rating it at ISO 2000, you can then expect to set aperture etc to get THE BEST RESULTS THE CAMERA IS CAPABLE OF. (At least overall – I agree with the way the XDCAM user quote refers to best “trade off”.) It may not be a surprise that rating it at a higher ISO won’t give such good results – we’re all familiar that such implies gain and likely more noise – but the same is true if you try rating it at a LOWER ISO. In the case of the FS7, ISO 2000 produces the OPTIMUM result.

And if Panasonic say the base ISO for the DVX200 is ISO 500, then the same is true there. It then becomes worthwhile querying why the FS7 is then 2 stops better than the DVX200. Most obvious is sheer sensor size, but as s35 is only about twice as large as 16:9 4/3 (not 4x), that only accounts for about 1 stop in itself. I can only presume the other stop is due to Panasonic using a higher than 4K pixel count sensor (hence smaller than optimum photosites), and Sony exploiting their back-illumination technologies.

It does mean that assuming the same T-stop, the FS7 has a 2 stop advantage over the DVX200, and the FS5 (base ISO 3200) is nearer 3 stops!

And sorry, but your point 3 above is incorrect. You can expect 0dB to be where you get the BEST dynamic range. Using a log setting won’t automatically give you a higher dynamic range (contrary to a lot of belief) – what it WILL do is give better post flexibility as to how the basic camera DR is mapped to output. If you don’t use log, then the high DR gets mapped to a lower one (suitable for the eyes more limited range) in a set fashion, according to gamma curve, knee, slope and such factors.”