Is graininess proportional to speed?

BetterSense · Feb 9, 2014

This is not a question for casual speculation but a technical question for those who might have information deeper than that found in film datasheets.

Are there quantitative methods of describing grain? How do they work? What criteria are they based on?

What is the approximate functional relationship between graininess and film speed? Is graininess proportional to speed, proportional to the square root of speed, etc? I am very interested in this relationship for normal film illuminances, and including behavior at the extremes of speed.

What I'm getting at with this question:

Suppose we define a graininess scale such that a graininess of 0.5N enlarged 2X looks the same as a graininess of N at 1X.

If N is proportional to ISO speed for normal film illuminances, then it would mean that smaller formats would/could be less grainy than larger ones. If you cut the linear negative size in half, you double the effective graininess N, but you also boost the speed by 4x. If you then choose a film 2 stops slower, then you will end up with a result with the same speed as you started but half the graininess.

If N were proportional to the square root of ISO speed, then all formats could be similar in their graininess. This is why I'm interested in the functional relationship between ISO speed and graininess.

Tom1956 · Feb 9, 2014

Film technology is somewhat empirical of a set of rules to the end user. Consider T grain films as an example, compared to the older technology. But by and large, faster film means bigger grain, yes. I think I'm reasonably safe in my assessment, though certainly not a photgraphic engineer, or anybody who could have even ever have become one.

BetterSense · Feb 9, 2014

Of course faster film is grainier.

The whole point of the thread is the question: what is the functional relationship between speed and graininess?

Oren Grad · Feb 9, 2014

In Kodak and Fuji (but not Ilford) data sheets you will find a number for the "diffuse RMS granularity" of the film. If you inspect a variety of those you will find that the granularity value is not a simple function of rated film speed.

Tom1956 · Feb 9, 2014

If I were to speak further than I have on this topic, I would be outside my bounds to the point of just a know-it-all bigmouth. That said, Oren speaks for me.

Vilk · Feb 10, 2014

BetterSense said:
...what is the functional relationship between speed and graininess?

ah, this is deep... the darker it gets the less we comprehend... the horror! the horror... mistah kurtz--he dead

snapguy · Feb 10, 2014

experience

My experience in developing film and making prints since Eisenhower was president is that there are a multitude of reason why grain is big, little, attractive, nasty-looking or hardly there. This includes exposure, developer, developer time and temperature, constancy of temperature, purity of the chemicals, lens qualities, age of the chemicals, stray or backlighting, light leaks, type of film and if it is ripe or overripe, quality of manufacturing of the film, previously frozen or overheated film, over-or underexposure, humidity rates, film swelling due to too-long immersion, failure to use proper lenshoods, low quality lens filters, and just plain good or bad luck. To name a few.

Gerald C Koch · Feb 10, 2014

Yes, the graininess of a film can be measured by a function called the diffuse Root Mean Square granularity. When Kodak published the RMS granularity for their films you could see a definite correlation with the speed of the film. Faster films have a larger value. For example, Tech Pan 5-8, PX 10, TX 17. But there are other factors at play such as the choice of developer and the Gamma to which the film is developer. This can be seen by the range of values for Tech Pan.

BetterSense · Feb 10, 2014

RMS granularity may not be the best measure for this topic. What we need is something perceptual like Kodak's Print Grain Index only without the normalization for enlargement.

Yes there are many variables. That's all well and good, yet we know faster film is grainier than slower film. Despite the noisy data, it should be possible to make generalizations about the functional form. If ISO speed is proportional to grain^X, then we should be able to say at least something about X, namely, that X is positive, but is X greater than or less than 1?

TMAX100 is 2 stops slower than TMAX400. If I shoot grey cards with both, develop them to the same contrast and enlarge the TMAX100 to 4 times (linear measure) bigger than the TMAX400, will the grain be coarser or finer than the TMAX400?

NB23 · Feb 10, 2014

RMS ratings is probably what you're asking about. I would go with rms data to compare films, their speed and their grain.

Oren Grad · Feb 10, 2014

BetterSense said:
Despite the noisy data, it should be possible to make generalizations about the functional form.

This is not obvious.

BetterSense said:
If ISO speed is proportional to grain^X...

I don't think you can assume that. Even assuming development with the same developer to a common CI, whether you consider RMS granularity or some other perceptual measure, the difference between TMX and TMY will be neither the same as nor related by a simple multiplier to the difference between PX and TX, or FP4 Plus and HP5 Plus, or Delta 100 and Delta 400. If you try to make comparisons across "old technology" films and "new technology" films the picture will be even more muddied.

Anon Ymous · Feb 10, 2014

BetterSense said:
...TMAX100 is 2 stops slower than TMAX400. If I shoot grey cards with both, develop them to the same contrast and enlarge the TMAX100 to 4 times (linear measure) bigger than the TMAX400, will the grain be coarser or finer than the TMAX400?

You'd also need to compare grain from areas with the same density. Granularity increases as density increases in BW negative films. That's why you see diffuse RMS granularity measurements at a specific density (1.0). It is quite complicated...

BMbikerider · Feb 10, 2014

Generally and practically speaking, the answer to your original question is probably yes! However with different technologies being employed, some more modern films you will find the reverse is true to some extent. Take for example, Kodak Tmax 400, the grain on a correctly exposed and developed negative will be as fine as that on a similarly exposed and processed FP4+ negative. Additionally Ilford XP2 and Kodak Cn400 will be as fine as, or better than, almost any slower film down to about 100ISO. However they may not appear as sharp as 'normal' silver based film.

These are not technically accurate statements from tests carried out in controlled conditions in a laboratory, but from practical use in the camera and then in the darkroom which may be the answer you are looking for.

Xmas · Feb 10, 2014

Grain is like boketh - subjective - except for formal measurements like the data sheets has...

gone · Feb 10, 2014

What snapguy said.

I recently developed two rolls of film in the same tank that were shot w/ two different cameras. Each camera had a different lens than the other one, and both had been loaded w/ the same film from the same batch. Both had been electronically checked for shutter speed accuracy, and both had been shot on the same day using the same meter at the same time of day. After development they looked totally different. You would have sworn that I had used two different developers.

Is graininess proportional to speed?

Privacy & Transparency

Privacy & Transparency