DR, fiction or fact?

I have just read an article in the UK magazine "Amateur Photographer" that has explained (at least partially) why I seem to be totally at odds with Canon regarding Dynamic Range.

Prof Bob Newman explains that DR is borrowed from the electronics world and has a serious problem when used by us photographers. The problem is noise and the article has a little graphic demonstrating the problem. It shows on the top line a standard tonal scale from black at one end to white at the other, clearly broken into 17 distinct tones, the bottom line is the same but has noise artificially added and the whole thing becomes a total mess. There is not a single clear tonal change and all that can be said is that yes, one end is darker than the other! By how much is unclear.

To avoid this problem he claims that we should really be talking in terms of Tonal Range, not DR. Thats because as photographers, its TR thats important to us.

Theres a lot of science in the article which goes straight over my head, but at least I now know why Canon manage to annoy me whenever they quote fictional DR figures in their sales pitches, its not exactly fictional, but might as well be for all the use it is.

My question to Canon would therefore be "Why use DR when you could just as easily give us TR figures instead"?

That's why DR becomes lower at higher ISO: More noise. I'd say the same would go for TR.
I've never trusted any camera manufacturers sales pitch about anything

http://www.dxomark.com …or-scores/Use-Case-Scores

The biggest problem with DR is that there is no single clear and absolute definition of what the bottom point is. Some people define it as the point at which noise "obscures" optically derived detail. Obscuration seems highly subjective in the first place and relative to whether it is untreated Raw data or is after NR. (Sony, for instance, does NR to sensor data before writing it to a "Raw" file). Others have tried to objectify the bottom point by defining it as the point where the S/N reaches a certain value, but so far as I am aware there is no binding standard from either ISO or the Japanese camera makers' institute.

So does that mean that canon RAW files are more RAW than sonys?

I wish I understood more than the tiny fraction of the article that tried to explain why DR is a poor choice for a visual medium. I had not realised that DR was borrowed from the electronics engineers toolkit, not something ever designed for photography.

Maybe, if Canon can tackle the noise issues, we might get a better sense of the true tonal range of which our cameras are capable. I remain convinced that my 5DII achieves around 5, maybe 6 stops TR in good light, regardless of Canons inflated DR claims.

If its all about noise then why are HDR images always so noisy? It appears to me that they have got it wrong also.

Richard, it occurs to me that I have never seen any real numbers directly from Canon regarding DR, only numbers from third party web sites - DxOMark, DPR, Image Resources - who have examined and tested and reported their conclusions. These sites also report on cameras from other makers and I have never heard that they are affiliated with Canon or any other maker. Regarding the 5D2, for instance, official Canon publications which list the camera's specifications, the manual and the White Paper, do not include that parameter in the spec list. The White Paper, in a section comparing the 5D2 to the 5D, does make vague statements about the increase to 14 bit Raw providing "improved DR" and "enhanced highlight and shadow detail" but no definite numbers. Nor can I recall ever seeing any advertising from Canon or its agents that contained numerical descriptions of purported DR, only vague claims to be better although how much better was never said. So could you point me to the documentation from Canon of the camera's DR? I would be interested in seeing if the claimed DR is, in the light of my own experience, as inflated as you say.

The source of the term "Dynamic Range" is surely irrelevent as long as its meaning in the context of photography is understood. Borrowing terminology from other contexts is something that is done all the time; when I say an image is "noisy" I don't mean it hurts my ears nor when I describe a young lady as hot am I referring to her temperature. When we say "DR" we are talking about the difference in light intensity, expressed in stops of EV, between the saturation (white) point of the sensor and the darkest shadow that can be captured without an unacceptable amount of noise. Since we all know that that is the meaning of DR, there is no problem in using the term. The problem is in knowing what "unacceptable" means.

Elie,

You could well be right that the "inflated claims" that I object to so much are not of Canons own making. Now that I know that DR was never intended to be used in image terms, the term "Tonal Range" seems a far better method.

Its a shame I cannot post a link so that people like you with far more technical knowledge could read the article and maybe explain the concepts, but I can't and thats that.

As you know, I've commented in the past that I can meter the grass near me (as a mid-tone) and the sky, then use that difference to select a graduated neutral density filter which is most often just a two-stop filter or at the very most three. I've been severely critised for saying this, but it remains true regardless of others claims of Dynamic ranges exceeding this by massive margins, twice what I see if not more.