Jump to content

Fuji granularity ratings


Recommended Posts

Hello,

 

According to official Fuji product specs, the following stocks have a granularity rating of:

 

- Eterna 500 = 3.5

- Eterna 400 = 3.7

- Eterna 250 = 3.5

- Eterna Vivid 500 = 3.5

- Eterna Vivid 160 = 3.5

- Reala 500D = 4.0

- Eterna 250D = 3.5

- F-64D = 2.5

 

I understand that some stocks are older than others but these figures would indicate that (assuming that they accurately reflect grain levels for each stock) different film speeds are not directly linked to the level of grain with Fuji stocks. For instance, Eterna 400 has a higher grain rating than Eterna 250 / 250D. The same applies with the new Vivid range, I am perplexed as to how the 500 and 160 could possibly have the same amount of grain.

 

Any input or explanation into this would be greatly appreciated.

 

Thanks in advance,

Edward Goldner

Link to comment
Share on other sites

  • Premium Member

The RMS Granularity rating is such an old system, designed back when stocks were a lot grainier, that it is nearly useless today for gauging how grainy a stock is relative to another. I mean, Eterna 500T is grainier than Eterna 250D, just not enough for the RMS method to tell. The RMS system doesn't differentiate between different layers, it doesn't take into account how grain edge sharpness can affect the perception of graininess, etc.

 

You really have to shoot them and project them onto a big screen (or scan them and enlarge them) to see the grain structure and compare them.

 

Modern stocks are really close today in terms of graininess. Kodak Vision-2 100T, 200T, and 500T are all really close, but the old rule still applies that the slower stock within a series is less grainy.

Link to comment
Share on other sites

The RMS Granularity rating is such an old system, designed back when stocks were a lot grainier, that it is nearly useless today for gauging how grainy a stock is relative to another. I mean, Eterna 500T is grainier than Eterna 250D, just not enough for the RMS method to tell. The RMS system doesn't differentiate between different layers, it doesn't take into account how grain edge sharpness can affect the perception of graininess, etc.

 

You really have to shoot them and project them onto a big screen (or scan them and enlarge them) to see the grain structure and compare them.

 

Modern stocks are really close today in terms of graininess. Kodak Vision-2 100T, 200T, and 500T are all really close, but the old rule still applies that the slower stock within a series is less grainy.

 

David,

 

how did the RMS Granularity rating work? I think it would be interesting to have fixed, technical parameters to measure the amount of grain, are there any up-to-date methods?

Link to comment
Share on other sites

Here's how RMS granularity is measured and calculated.

Expose and process a frame of negative to a density of 1.00

Scan across a section of the negative with an aperture of 0.048mm, recording the density as it varies across the grain structure.

Plot the variation from the mean (average) density.

The RMS granularity is actually the root mean square of those deviations - in other words, the average amount the density varies.

 

The grain - or dye cloud - size in any emulsion varies a lot, but typically a dye cloud might be a tenth of the size of the aperture. Bigger ones will have more effect on the density as you scan across them. Smaller ones less so. So this is really a measure of the size of the grains or dye clouds, and the gaps in between them.

 

The standard aperture of 0.048mm was apparently settled on some time ago. It's roughly equivalent to a 0.5K scan. It's OK to compare different emulsions so long as you use the same aperture for all. Of course, a smaller aperture would discriminate more between different grain structures given that modern films are finer-grained - but if the aperture was too small, the measure would not really be representative of the appearance of the image.

 

Granularity is a completely objective measurement, and it leads to a single number (easy to quote and compare) but it doesn't say anything about different densities, or take into account the number of different coloured emulsion layers (which have quite different grain structures from each other).

 

Graininess, by contrast, is a subjective description of the appearance of grain in an image. But there isn't really a reliable way of measuring this (as it's subjective) except by showing images to judging panels and asking them to rate them in order of graininess.

Link to comment
Share on other sites

Hi,

 

Yeah, basically the lower the number the lower the granularity.

Here's an interesting comparison. I have an old spec sheet

of Kodak 5254 (given to me by dear John Pytlak).

That stock was 100 ASA. But the RMS granularity for it was 7.

5212, the current Kodak 100 ASA stock is around 2.5! So the current

500 ASA stocks at RMS 3.5 have much less granularity than a 100 ASA stock

of 35-40 years ago.

 

I've shot all of Kodak's stocks on 35mm and something like 5219 viewed on

a television has no apparent grain.

 

Regards,

Milo Sekulovich

Cinematographer

Link to comment
Share on other sites

While I can gauge the difference in a stop or resolution, the RMS reading (even knowing what root mean square is) proves difficult to wrap my head around.

 

What units are being used to plot the variance? I take it they are logarithms, just as in densitometry, but logs of the square root of the variance?

 

I have an old tech guide with a two-page pullout on RMS readings, but it makes no more sense after the first reading than the fiftieth.

 

 

If the graininess is measured by a 48nm aperture, Dominic, and that approximates to a 0.5K scan, wouldn't you need roughly, what, a sixteenth of that size to get an accurate reading of perceived granularity at 2K resolution?

 

SOT, it seems to me that LP/mm measurements are also completely outdated. The 1000:1 and 1.6:1 contrast ratios of fast films are often just as high as slow films, making them useless with current stocks also.

Link to comment
Share on other sites

I am also skeptical. . . there are too many 3.5s in that series to be coincidence, almost like Fuji has optimized their stocks to read favorably based on the RMS test.

 

Kodak is guilty of the same thing. They have stopped quoting LP/mm readings, and were using something called "PGI" perceived graininess index so they didn't have to directly compete against the competition.

 

I wish someone would come up with a concrete comparative index of resolution, granularity, and latitude.

 

Testing is the best option, but a lot of the testing could be conducted on a level devoid of an individual production if the standards were modernized, allowing for better fine-tuning in the actual production testing phase.

Link to comment
Share on other sites

I wish someone would come up with a concrete comparative index of resolution, granularity, and latitude.

Unfortunately, these are all in the muddy area between objective measurement and subjective impression, so a "concrete" measurement is a little problematical. And where you can devise a concrete measurement, it is usually more complex than people would like it to be.

 

So for resolution, it used to be measured in lppmm. (Line pairs per milllimetre). On a reso chart, how fine are the rulings of black and white lines that you can actually see? Too fine, and it's just a grey blur. There's a range where you can see the line structure, but it's down to light grey and dark grey. Instead of a density of 0.00 and 1.00 for example, the lines are 0.30 and 0.70. Or maybe 0.49 and 0.51. Where is the cut-off point? That's why MTF (Modulation Transfer Function) is used now: but it's a graph, not a single number.

 

And for granularity, no, a very small aperture wouldn't work. Think of it this way. A grain-structured but otherwise uniform grey patch is much the same as a random scattering of lumps of coal on a large white sheet. About half the sheet is covered, giving an average grey appearance.

 

Now, the granularity is a way of measuring the size of the grains or lumps of coal (which aren't uniform of course). In practice, it is a measure of how much the actual density in a small given area varies from that average. In a way, it is like taking a spotmeter reading, scanning steadily across the sheet. The smaller the spotmeter's field of view, the more chance you have that your reading will be all coal or all sheet. So you'll get a lot of 0% or 100% light readings, leading to a high RMS deviation from the average. If the lumps of coal are twice as big, you'll still get mostly 0%s or 100%s. So a larger aperture (or spotmeter angle) is actually more discriminating. Though if it's too big, it won't detect much variation at all.

 

I think the 0.045mm aperture is a legacy from much older and grainier film emulsions, and from days of black and white, when it was grain being measured, not colour dye clouds. It would be better if it were smaller. But something akin to a 2K scan (about a 0.01mm) would be ideal, not smaller still.

 

just while I'm here . . .

almost like Fuji has optimized their stocks to read favorably based on the RMS test.
Well, if they believe that RMS granularity is a good indicator of graininess, why wouldn't they?

 

As for latitude, that's a whole new thing. Another story for another time.

Link to comment
Share on other sites

just while I'm here . . .

Well, if they believe that RMS granularity is a good indicator of graininess, why wouldn't they?

 

As for latitude, that's a whole new thing. Another story for another time.

 

Thanks for the explanation.

 

 

I am not singling out Fuji. Kodak has done something far more marginal in abandoning lp/mm values altogether in favor of "PGI" which doesn't have any established correlation.

 

They've also established *their own* standards commission for evaluating dye fade on their products, using half the light and filtering out UV radiation compared to the otherwise-standardized evaluative standards used by other companies in evaluating color dye longevity characteristics.

 

 

 

Getting back to RMS, if the test is designed to evaluate perceived graininess, and it has instead morphed into something where perceived graininess is high but RMS is still deceptively low, there is clearly a need for revision of the evaluative standard to account for this discrepancy.

 

As for "another story for another time" when can we tune in for that episode, next week? :P

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...