On standard aperture values

[Doug Kerr]

Standard f-number values and their "labeling" is covered by standard ISO 517. I have only the first several pages (a "free sample"). In that is this list of "standard series of f-number marking":

05.-0.7-1-1.4-2-2.8-4-5.6-8-11-15-22-32-45-64-90-128

These are of course 1n 1-stop steps. There may be standard "rounded" values for 1/3- and 1/2-step series elsewhere in the standard, but not in the part I have. Or it my be that such lists have only been concocted by camera manufacturers. (There seems to be considerable consistency in that.)

Best regards,

Doug

 

[Ted Cousins]

Standard f-number values and their "labeling" is covered by standard ISO 517. I have only the first several pages (a "free sample").

The following site has all the pages for various ISO's

I posted this earlier in another thread:

https://cdn.standards.iteh.ai/samples/50089/796cb13d68a648b386176d3fe20c1323/ISO-517-2008.pdf

See 3.2.4 Subdivisions of the standard f-number

Each standard f-number, or “whole stop”, division of scale markings may be divided into two or three
subdivisions in 1/2 or 1/3 steps of a stop calculated, for 1/2 subdivision, in a geometric series with first term 0,5 and factor 4th root of 2 and for 1/3 subdivision in a geometric series with first term 0,5 and factor 6th root of 2 .[/1]

best,

Ted.

 

[Doug Kerr]

The following table gives various known series of aperture values:

The values for the various cameras mentioned are the marked" values of the available aperture settings (with the "1/3 step" option set, where applicable).

I have so far only found a list of preferred marked values in an ISO standard that has values at 1-stop intervals.

I note that the value series for the two Canon cameras are identical over their common listed range. (The ranges are actually wider than the portions I know or chose to list.)

There is considerable evidence that the series seemingly consistently used by Canon was derived from a Renard preferred value series, in particular the R20 series (20 values per decade), but in its "most rounded" variant (R''20). Note that a principle of all Renard series is that for two values separated by a certain number of steps (20 for the R20 series), the value increases by a factor of exactly 10

I note that there is complete consistency between the Canon series and the Renard R''20- series except for the values (Canon value stated) f/6.3 and f/7.1. This is no doubt a difference in rounding philosophy.

Thanks to Ted Cousins for bringing to our attention the Renard series and in particular the R''20 series.

Best regards,

Doug

 

[Doug Kerr]

This site has all the pages for various ISO's

No, these are only "free samples", with a limited number of pages, generally all in sequence from the beginning of the document..I think that for some standards with few pages, the "free sample" has all of the pages!

I posted this earlier in another thread:

https://cdn.standards.iteh.ai/samples/50089/796cb13d68a648b386176d3fe20c1323/ISO-517-2008.pdf

Here is the "watermark" from the cover of that document:

But thanks for calling attention to that standard - I missed the significance of it at the time.

Best regards,

Doug

 

[Ted Cousins]

No, these are only "free samples", with a limited number of pages, generally all in sequence from the beginning of the document..I think that for some standards with few pages, the "free sample" has all of the pages!

Oh ... Fortunately for us, my link to ISO 497:

https://cdn.standards.iteh.ai/samples/4548/eea948d9047b4a82b62a7ed459f7eaa4/ISO-497-1973.pdf

... shows all seven of it's pages for rounding preferred values including a table on page 3-4 with 'R' bases and degrees of rounding. Could we refer to that instead of making a new table for each discovery?

 

[Doug Kerr]

So overall, we seem to see a"1/3-stop" aperture plan that looks toward these principles:

• If we go forward 6 steps, the f-number increases by a factor of 2. [Note that for this series of f-numbers that is a difference of 2 stops.]

• If we go forward 20 steps, the f-number increases by a factor of 10.

Of course, both of those cannot be true overall. And indeed, in the "Canon plan", for example, there are several departures from those principles.

But we can hear the two metronomes ticking in the shadows!

Best regards,

Doug

 

[Doug Kerr]

This table shows a major part of what I call the "Canon series" of apertures (that isthe long list from the EOS 3 manual, but with the value f/13 replaced by f/12, as seen in some more modern cameras).

To the right I show the ratios between values 6 steps apart (2 stops) and 20 steps apart. There are 6 sets of the former (shown in separate columns) and 10 sets of the latter (only 8 of which are applicable in the range of this table. Dots show the predecessors of the first ratios of each set.

Ratios that are not exactly the "ideal" values are highlighted in pink (it is not possible for all ratios to be the "ideal" ones).

Best regards,

Doug

 

[Doug Kerr]

Oops! I was in error to make one of the values f/12. That should have been f/13. I will have a revised table shortly.

Best regards,

Doug

 

[Ted Cousins]

ISO 17 is interesting

Preferred numbers were first utilized in France at the end of the nineteenth century. From 1877 to 1879, Captain Charles Renard, an officer in the engineer corps, made a rational study of the elements necessary in the construction of lighter-than-air aircraft. He computed the specifications for cotton rope according to a grading system, such that this element could be produced in advance without prejudice to the installations where such rope was subsequently to be utilized.

Recognizing the advantage to be derived from the geometrical progression, he adopted, as a basis, a rope having a mass of 'a' grams per metre, and as a grading system, a rule that would yield a tenth multiple of the value 'a' after every fifth step of the series.

https://cdn.standards.iteh.ai/samples/3602/bdb2edef62674fb188bea630caf33012/ISO-17-1973.pdf

 

[Doug Kerr]

Hi, Ted,

Thanks. I had not been aware of that standard.

And especially valuable is ISO 497, cited there, which discusses the "more rounded" Renard values.

Best regards,

Doug

 

[Ted Cousins]

Oh ... Fortunately for us, my link to ISO 497:

https://cdn.standards.iteh.ai/samples/4548/eea948d9047b4a82b62a7ed459f7eaa4/ISO-497-1973.pdf

... shows all seven of it's pages for rounding preferred values including a table on page 3-4 with 'R' bases and degrees of rounding. Could we refer to that instead of making a new table for each discovery?

I compared my Weston Adams/Archer dial to the ISO 497:

It's f/44 finds no matches in the ISO 497 table. Other f/numbers e.g. f/0.7 only find exact matches in extended roundings.

Be that as it may, ISO 497 does mention departures from listed values at some length ...

... so maybe we can say "based on" instead of expecting full and perfect matching to the table??

 

[Doug Kerr]

Hi, Ted,

Yes, most likely the "free sample" of ISO 497 in fact is the whole thing.

Thanks for the pic of the Weston "AA" calculator. Its f-number scale has marks at 1/4 stop intervals with the 1/2 stop values marked.

If we limit ourselves to the marked values, the interval analysis looks like this:

We see that the series scrupulously follows a change of 1 stop for two steps, except for the "bump" at f/11. That is presumably to make the following values integers.

Since there are no values at 1/3-stop intervals, we cannot expect to see any "shadows of 10".

Best regards,

Doug

 

[Doug Kerr]

Hi, Ted,

Seemingly, most Weston meters, with their "stock" calculators, only have positions on the f-number scale up to f/32.

The Manual for the Weston Master V tells how to figger where the subsequent full stop steps would be. It says that those values would be f/45, f/64, f/90, and f128.

Best regards,

Doug

 

[Doug Kerr]

Of course the most important thing to remember is that, on exposure meters with circular slide rule calculator, if the major series of marks for one paramater is at a nominal spacing of 1 stop, then the intent is that it they are all at the "theoretical" positions for those 1 stop increments.

So, on the Sekonic L-398A, the aperture scale mark labeled "1.4" is intended to be at exactly the position for sqrt (2). It would not matter if it were labeled 1.3 or 1.41 or 1.5.

I expect the same can be said for the Weston meters. Them the mark that is in come cases labeled "44" and in other cases "45" is in both cases intended to be at the position {sqrt (2)}^11, or about 45.2548.

Now how cameras with digital settings deal with this I have no idea.

Best regards,

Doug

 

[Doug Kerr]

Further to that same point.

It seems likely that the intended "scaling" of the aperture and shutter speed dials on the L-398A is that 1 stop is exactly 12°.

But it is unlikely that on the drawing for the aperture scale the distance from "f/1" to "f/1.4" is shown (if angles are stated to 0.1°) as 5.8° (the ideal position for f/1.4). Almost surely that distance is shown (or implied) as 6.0° (which represents a value of about 1.41).

Best regards,

Doug

 

[Ted Cousins]

Doug,

I can think of no reason why the Weston Adams/Archer dial should not be marked "45" instead of where it says "44" thereby making that marking non-compliant (with the ISO 497 table).

In the table, preferred value 4.5 is rounded from the actual listed ratio of 4.4668; ergo preferred value 45 would be rounded from 44.668 (20th root of 10)^33 ... closer to 45 than 44.

 

[Ted Cousins]

Of course the most important thing to remember is that, on exposure meters with circular slide rule calculator, if the major series of marks for one paramater is at a nominal spacing of 1 stop, then the intent is that it they are all at the "theoretical" positions for those 1 stop increments.

So, on the Sekonic L-398A, the aperture scale mark labeled "1.4" is intended to be at exactly the position for sqrt (2). It would not matter if it were labeled 1.3 or 1.41 or 1.5.

I expect the same can be said for the Weston meters.

Sounds like we can ignore ISO 497 then ... which presumably lists what a mark should be rather than what it is ... LOL

Just kidding,

Ted.

 

[Doug Kerr]

The following table lists one decade of the Renard R20 and R40 series, along with their "more rounded" and (in the case of the R20 series) the "more rounded yet" variants.

The entries of the R20 series are spaced out vertically to match the values of the R40 series. The first entry of the next decade is included (shaded in gray) for the sake of continuity. For values beyond those of this first decade, all the values are multiplied by an integral power of 10.

The pink highlighting shows where the "rounded" variant values differ from the "base" values; the yellow highlighting shows where the "more rounded yet" variant values differ from the "more rounded" variant values.

Best regards,

Doug

 

[Ted Cousins]

I may have steered us off-track to an extent. Although I have been emphasizing ISO 497, that is not a photography standard, so there is no reason that photographic stuff should follow that other than aperture preferred values should reasonably follow series such as those by M. Renard.

As previously mentioned we have ISO 517:

https://cdn.standards.iteh.ai/samples/50089/796cb13d68a648b386176d3fe20c1323/ISO-517-2008.pdf

Firstly, it specifies aperture markings for all types of lenses used in still cameras, and gives tolerances for the stop numbers ... so not for light-meters ... but possibly for digital camera lens aperture settings by implication; after all, it would odd if one could not set a lens on a digital camera to 𝑓/8. In other words, the camera is just a fancy substitute for an aperture ring and it's markings.

Again, as already stated by Doug:

The standard series of f-number marking shall be as follows:
0.5 – 0.7 – 1 (or 1,0) – 1.4 – 2 – 2.8 – 4 – 5.6 – 8 – 11 – 16 – 22 – 32 – 45 – 64 – 90 – 128

Secondly, it gives tolerances:

e.g. "For 𝑓/5.6 and above, + 19 % − 16 % (± 1/2 stop).

Bearing in mind the above tolerances, I withdraw my red herring about Weston dial markings.

 

[Ted Cousins]

Standard f-number values and their "labeling" is covered by standard ISO 517. I have only the first several pages (a "free sample"). In that is this list of "standard series of f-number marking":

05.-0.7-1-1.4-2-2.8-4-5.6-8-11-15-22-32-45-64-90-128

These are of course 1n 1-stop steps. There may be standard "rounded" values for 1/3- and 1/2-step series elsewhere in the standard, but not in the part I have.

See table 1, page 3:

https://cdn.standards.iteh.ai/samples/50089/796cb13d68a648b386176d3fe20c1323/ISO-517-2008.pdf

Slightly off-topic, does anyone have a full copy of ISO-2720 ?

 

[Doug Kerr]

Hi, Ted,

See table 1, page 3:

https://cdn.standards.iteh.ai/samples/50089/796cb13d68a648b386176d3fe20c1323/ISO-517-2008.pdf

Those are the "theoretical" values, rounded to 3 or 4 significant figures, but not the further-rounded values used in marking cameras and such, to which I alluded (but was perhaps not clear about that).

Best regards,

Doug

 

[Doug Kerr]

Now what I don't know is, for example, in my Canon EOS M50, when I set the aperture (manually) to f/5.6, do the tables in the camera try to set the aperture to f/5.600 or to f/5.657 (working here arbitrarily to 3 decimal places) as best it can?

I know others have studied at length the aperture stop setting systems of Canon cameras, but not I. And the details may well differ between various lens models (for interchable lens cameras)

Best regards,

Doug

 

[Ted Cousins]

Now what I don't know is, for example, in my Canon EOS M50, when I set the aperture (manually) to f/5.6, do the tables in the camera try to set the aperture to f/5.600 or to f/5.657 (working here arbitrarily to 3 decimal places) as best it can?

I know others have studied at length the aperture stop setting systems of Canon cameras, but not I. And the details may well differ between various lens models (for interchable lens cameras)

I've never owned a Canon.

In my old Sigma DSLR's, the raw files contain decimal parameter values for some variables and rounded values separately (what you see on the LCD) in ASCII string format for those same variables.

... ... BEGIN: CAMF matrix meta data (CaptureAperture)
float [1]
x: D0
4
... ...
[20] "APERTURE" = "4.00000"
[21] "AP_DESC" = "4"
... ...

One floating point number perhaps sent to the lens and two strings - AP_DESC is what I see in the camera LCD.

Probably shot in Manual or Aperture Priority, posted FWIW.

 

[Doug Kerr]

Hi, Ted,

I've never owned a Canon.

In my old Sigma DSLR's, the raw files contain decimal parameter values for some variables and rounded values separately (what you see on the LCD) in ASCII string format for those same variables.

Ah! A good way to look at that.

Thanks for the hint.

Best regards,

Doug

 

[Ted Cousins]

Hi, Ted,

Those are the "theoretical" values, rounded to 3 or 4 significant figures, but not the further-rounded values used in marking cameras and such, to which I alluded (but was perhaps not clear about that).

Best regards,

Doug

Then it looks like we are stuck with ISO 497 and this highly intelligible diagram from page 3-4:

One can only assume that rounding in this case is done per ISO 8000-1 or whatever ISO.

 

[Ted Cousins]

Then it looks like we are stuck with ISO 497 and this highly intelligible diagram from page 3-4:

View attachment 14465

A pity because there are no perfect intervals of 1 EV in that table's calculated values other than by rounding - and why should there be when the 'R series' true value ratios are roots of ten ... as opposed to ISO 517 where values obligingly double every so often, which is really nice for lenses and light-meters.

 

[Ted Cousins]

Hi, Ted,

Those are the "theoretical" values, rounded to 3 or 4 significant figures, but not the further-rounded values used in marking cameras and such, to which I alluded (but was perhaps not clear about that).

Yes, the value ratios in the table are:

Full-stop: 2nd root of 2.
Half-stop: 4th root of 2.
Third-stop: 6th root of 2.

ISO 517 also shows a "fully" rounded full-stop sequence, as seen on our Sekonics, but with no mention of the rounding method employed.

0,5 – 0,7 – 1 (or 1,0) – 1,4 – 2 – 2,8 – 4 – 5,6 – 8 – 11 – 16 – 22 – 32 – 45 – 64 – 90 – 128

AI agrees that the rounding is all over the shop, as does the super-knowledgeable Jim Kasson:

https://blog.kasson.com/the-last-word/why-photographic-numbers-are-strange/

I checked my Sigma SD10 1/3-stop f/numbers from 𝑓/2.8 to 𝑓/22 on the LCD display and the rounding method is unclear: sometimes compliant with ISO 517's listed values, sometimes not ...

best,

Ted.

 

[Doug Kerr]

Hi, Ted,

ISO 516 (free sample available from Iteh) gives this list of "preferred" values of the marking of exposure time (shutter speed):

. . .8, 4, 2, 1, 2, 4, 8, 15, 30, 60, 125, 250, 500, 1000, 2000. . .

(Black numbers are the denominators of fractions, with the numerator always 1, in seconds; red numbers are in seconds

The ellipses seem to imply that the list should continue in the same fashion, which probably means that after 2000 we would have:

4000, 8000, 15000, 30000, 60000

and in the "longer" direction we might have

125, 60, 30, 15, 8

The standard also says, after giving that list:

The highest marking, however, need not necessarily be selected from this series, but the series
beginning with the next lower number should be selected from this series, whenever practicable, and
progressing as far as is required in the particular application.

I also note that the standard specified that the actual times should ideally be from the theoretical series (interval ratios of 1/2), not the values in the list of "preferred markings". (There is of course a tolerance provided for.)

Very interesting.

Best regards,

Doug

 

[Ted Cousins]

Hi, Ted,

ISO 516 (free sample available from Iteh) gives this list of "preferred" values of exposure time (shutter speed):

. . .8, 4, 2, 1, 2, 4, 8, 15, 30, 60, 125, 250, 500, 1000, 2000. . .

(Black numbers are the denominators of fractions, with the numerator always 1, in seconds; red numbers are in seconds

The ellipses seem to imply that the list should continue in the same fashion, which probably means that after 2000 we would have:

4000, 8000, 15000, 30000, 60000

and in the "longer" direction we might have

125, 60, 30, 15, 8

The standard also says, after giving that list:



Very interesting.

Best regards,

Doug

Thanks, Doug. My camera complies except top speed is 6000 ...

Like f/numbers, the raw meta-data also has floating-point and string shutter speeds ...

 

[Doug Kerr]

Hi, Ted,

Thanks, Doug. My camera complies except top speed is 6000 ...

In a later edit to my post I quoted the passage from ISO 516 that recognizes just that!

Best regards,

Doug