Question about "stop" values

[Gilthanass]

This question may be mostly irrelevant in real world terms, but the mathematical side of me is very curious, so I thought I would throw this out there.

When we're talking about "stops" of light, I understand we're essentially halving and doubling the amount of light (photons in real world terms) that reach the sensor. Okay, I get that. My problem is, the values we see in our camera are not real doubling and halving values. So, is the camera using the correct mathematical value, but showing us rounded values, or is it using the rounded values so we end with with a not quite perfect exposure value?

I'll try to explain. Let's start with shutter speed, and a speed of 1 second (for no real reason other than it's clean).

The real halving values should be: 1,1/2,1/4,1/8,1/16,1/32,1/64,1/128,1/256,1/512,1/1024,1/2048,1/4096,1/8192
(values rather familiar for anyone who has worked with computers)
But our cameras use:
1,1/2,1/4,1/8,1/15,1/30,1/60,1/125,1/250,1/500,1/1000,1/2000,1/4000,1/8000

Granted those are much cleaner numbers, but they aren't exact halves. So, when changing from 1/8th to 1/15th, you really aren't cutting the light by half, you're off by about 3%.

Same thing goes with aperature values. They are close, but rounded. f/stop 11 should be 11.31 (for example).

So, my question is, does the camera use the 11.31 value for 11 while telling you 11, or does it use 11? Granted this might not have a huge effect on your pictures, but doesn't that mean that every f/stop / shutter combination is not exactly equal?

Again, this is more for my "I need to understand how everything works" side than my "I want to take better pictures" side.

[Sorarse]

At the end of the day they are just numbers. Who cares whether they are real or rounded. They could be labelled apples, pears and bananas for all I care. The important thing is that if you move your shutter speed by one stop, and move the aperture by one stop to compensate, you still get a perfect exposure. That's all that maters to me.

[timmyeatchips]

The important thing is that if you move your shutter speed by one stop, and move the aperture by one stop to compensate, you still get a perfect exposure.

But with rounded numbers, then what you camera calls a stop isn't really a stop, so you don't. I'd imagine this could cause problems when multiplying things up for very long exposures.

[timmyeatchips]

According to Wikipedia (http://en.wikipedia.org/wiki/F-number):

'Modern conventions have rounded the numbers from f/5.66 to f/5.6, f/11.13 to f/11, and f/44.72 to f/45. This is only for ease of writing – the actual ratio of aperture size to focal length is still based on the doubling or halving of the amount of light getting through the lens.'

So it's only the speed values you really have to worry about, assuming the lens accurately achieves the aperture it reports. I'm not certain but I believe for the speeds you do get exactly what the camera says.

[PhotosGuy]

Your membership cards for "Measurebaters Anonymous" are in the mail. :D

[Roy Mathers]

Your membership cards for "Measurebaters Anonymous" are in the mail. :D


:D:D

[doctorgonzo]

For shutter speeds, you are right: they aren't exactly different by 50%. And the camera does use those exact numbers: it doesn't show 1/15 and use 1/16 internally. These speeds were standardized a long time ago, long before computers and binary. Film has enough latitude that being off by 3% doesn't really affect exposure, so for convenience they didn't use a strict power-of-two sequence. Even with digital, a few percentage points either way doesn't make much difference.

[Roy Mathers]

I think sorase had it right.

[Gilthanass]

At the end of the day they are just numbers. Who cares whether they are real or rounded. They could be labelled apples, pears and bananas for all I care. The important thing is that if you move your shutter speed by one stop, and move the aperture by one stop to compensate, you still get a perfect exposure. That's all that maters to me.

No, at the end of the day numbers mean something. f/stop values are the ratio of focal length to aperature size. Both of these things exist in the real world, so the value has real, physical implications. Shutter speed is measured in seconds, again, a very real measurement. You cannot substitute "seconds" for "apples" because apples means nothing. I'm sorry, but I take issue with your tone. Also, in this case, my point was that no, you do not get a "perfect" exposure. You get an exposure that is potentially off by 3%. Is that significant? Probably not, but it isn't exactly the same either, is it?

And to reiterate, twice in my post I stated that I was interested in it for interest's sake, not because I think it would have any impact on my exposure. Sorry, but that's the type of person I am, and that's why I work in the job I do. I understand the difference may seem inconsequential to some (most?) of you, but for me, I'd rather know, and I figured someone here would know.

For shutter speeds, you are right: they aren't exactly different by 50%. And the camera does use those exact numbers: it doesn't show 1/15 and use 1/16 internally. These speeds were standardized a long time ago, long before computers and binary. Film has enough latitude that being off by 3% doesn't really affect exposure, so for convenience they didn't use a strict power-of-two sequence. Even with digital, a few percentage points either way doesn't make much difference.

Thank you, that's what I was looking for. So, aperature values are rounded in the camera display, but not in reality; but shutter speed is rounded both in display and in reality. Good to know. I wonder why that choice was made? Was it easier for the photographer to understand the rounded numbers? Maybe it was easier for mechanical shutter designers?

And yes, I understand it would make VERY little difference to your exposure. Again, that's why I tried to make it clear that I was interested for interest's sake more than to use to better my photography skills.

[RDKirk]

Thank you, that's what I was looking for. So, aperature values are rounded in the camera display, but not in reality; but shutter speed is rounded both in display and in reality. Good to know. I wonder why that choice was made? Was it easier for the photographer to understand the rounded numbers? Maybe it was easier for mechanical shutter designers?

Up until the 70s, different manufacturers were still at variance on some shutter speeds, such as the choice between 1/25 or 1/30 and 1/50 or 1/60.

Remember, too, that the mechanical devices did not then--and still don't--have the ability to be as accurate as the numbers indicate. The real-world variation of shutter speeds, for instance, is greater than the difference between 1/25 and 1/30 even with electronically timed shutters.

[spkerer]

Also, in this case, my point was that no, you do not get a "perfect" exposure. You get an exposure that is potentially off by 3%. Is that significant? Probably not, but it isn't exactly the same either, is it?

If you go further down that path, you could probably argue that very, very few pictures have ever actually been taken with a "perfect" exposure. How often do you think the "perfect" exposure falls on an exact combination of the available settings for ISO, aperture and shutter speed? As was mentioned, film (and digital) has some latitude, or else the only useful camera would be one with continuous/infinite adjustments in at least a couple of the settings.

[timmyeatchips]

Up until the 70s, different manufacturers were still at variance on some shutter speeds, such as the choice between 1/25 or 1/30 and 1/50 or 1/60.

Remember, too, that the mechanical devices did not then--and still don't--have the ability to be as accurate as the numbers indicate. The real-world variation of shutter speeds, for instance, is greater than the difference between 1/25 and 1/30 even with electronically timed shutters.

Is this still the case? I would have thought the mechanism would be precise enough to distinguish between the fastest settings (say, 1/4000 and 1/8000) so that a few hundredths difference should be easy.

I agree that it's not really an issue in most real world photography, but it is useful to bear in mind, say, the difference between 1/50 and 1/60 (a fairly sizeable 20%) if you're multiplying up for long exposures or working at the limits of dynamic range.

[sandpiper]

Also, in this case, my point was that no, you do not get a "perfect" exposure. You get an exposure that is potentially off by 3%. Is that significant? Probably not, but it isn't exactly the same either, is it?



It doesn't have to be exactly the same. You cannot set a 'perfect' exposure as you can only use 1/3 stop increments on most cameras so you are rounding off to the nearest 1/3 of a stop anyway. The 3% is insignificant when rounding off that much.

We used to be able to manage to shoot transparency film (no latitude as you are using the actual film as your final product, so cannot correct exposure in printing or PP) on cameras with fixed ISO, 1 stop shutter adjustment and 1/2 stop on the aperture.

[Roy Mathers]

I don't understand this discussion - people have been practising photography for an extremely long time without worrying about it!;)

[timmyeatchips]

Well somebody has to. If I were an aperture it would comfort me to know people were out there worrying about me.

[toxic]

I can't say for shutter speeds, but for f/stops, they're truncated.

[Roy Mathers]

Well somebody has to. If I were an aperture it would comfort me to know people were out there worrying about me.
:D:D

[Azi]

The important thing is that if you move your shutter speed by one stop, and move the aperture by one stop to compensate, you still get a perfect exposure.

Wow, didn't know it. That makes it pretty easy getting the right exposure when you change shutter speed or f-stop.
What about changing ISO?

[sandpiper]

Wow, didn't know it. That makes it pretty easy getting the right exposure when you change shutter speed or f-stop.
What about changing ISO?

Yeah, still the same with ISO. A 'stop' is the same regardless of whether it is aperture, shutter speed or ISO - either halving or doubling (approximately ;)) the light reaching the sensor.

If you have everything set to 'one-third' stops, it still works the same way. So if you want two thirds of a stop more shutter speed (2 clicks), you open up the aperture by two clicks or add two clicks of ISO. OR, one click of aperture and one of ISO will do the same and compensate for two clicks of shutter speed. Note, I don't actually recommend using thirds ISOs as they can be a little noisier than going up to the next 'full' ISO due to the way they are done in software, but that's a different matter to understanding the link between the three values in exposure.

This is particularly useful when using manual. If you are on a bright sunny day and using he 'sunny 16' rule, you will know that you want around 1/100th sec at f/16, ISO 100. If you are shooting with a long telephoto and want to combat camera shake you may decide you need 1/800th sec (a 3-stop difference) so you need to open up the aperture by 3 stops (to f/5.6 ) to compensate. That may mean that you would be shooting wide open and you want to avoid that as the lens is slightly sharper stopped down a little, so you select f/8 instead (2 stops wider than f/16 ) and bump the ISO by 1 stop to 200. The cumulative exposure is still the same in all three cases.

1/100th, f/16, ISO 100 = 1/800th, f/5.6, ISO 100 = 1/800th, f/8, ISO 200 = 1/800th, f/16, ISO 800 etc., etc.

[rdenney]

Gee, whiz, guys, can we just answer the question? If not, then resist that "submit" button.

As was mentioned by one (1) person, the "f-stop" is actually the ratio of the focal length of the lens to the diameter of the aperture. So, a 50mm lens at f/2 has an effective aperture diameter of 25mm.

Now, to the question. The reason they follow the series they do is because the area of the aperture circle is what controls the light, but it's the diameter that goes into the f-stop. So, they vary by the square of half (or double) that diameter value.

Let's go back to that 50mm, f/2 lens. At f/2, the diameter is 25mm, and the area is 491mm^2. If we want to double the amount of light reaching the sensor, we double the area to 982mm^2. The diameter of a circle with that area is 35mm. Now, let's determine the f-stop value: 50mm / 35mm = 1.4. So, to double the amount of light, we need an f-stop of f/1.4.

Because the series moves exponentially, when we double the f-stop value, we increase the area of the aperture by a factor of 4. That's the same as doubling that area twice (two doublings = a factor of four). Thus, going from f/4 to f/8 (doubling the f-stop value) moves the exposure two stops, first to f/5.6, and then to f/8.

Yet another way to look at it: If we want to double the area of an aperture, we increase the diameter by 2 times the square root of 2. So, we can multiply each f-stop by 1.414 to get the next f-stop in the series, which is then reasonably rounded.

f/1 times 1.414 = 1.4
f/1.4 times 1.414 = 2
f/2 times 1.414 = 2.8
f/2.8 times 1.414 = 4
f/4 times 1.414 = 5.6
f/5.6 times 1.414 = 8

And so on. The standard series is not exact, but it is close enough.

There is a difference between theory and practice with complex lenses, and depending on where the aperture is located. Don't expect it to measure the way theory predicts. In practice, they adjust the aperture scale to provide the light transmission that matches theory, not by measuring the iris diameter. Only in mirror lenses will there be a significant difference between the f-stop and the light transmission.

Rick "fun with geometry" Denney

[Roy Mathers]

Yes, but does that help you to take better pictures?:D

[spkerer]

Gee whiz, Rick, why call us all out because we didn't "answer his question" and then go ahead and only answer one part of his question? Couldn't resist that submit button? He asked about BOTH shutter speed and aperture and you just answered about aperture. So there! :-) I hope this comes across sufficiently tongue-in-cheek.

I actually think that to discussion about why a 3% difference is tolerated is actually useful.

[CUclimber]

Yes, but does that help you to take better pictures?:D
Seriously-- why even bother with a response like this?

It's as if being ignorant of how things actually work is smiled upon these days.

No, of course it won't let me take a better picture of my cat so that I can post it on PotN. It will, however, teach me something that I didn't know before and that helps me better understand the science that lies underneath every single aspect of photography.

Fraknly, some of the "Why bother asking, since it just works out anyways" responses on the first page of this thread are rather depressing.

[Roy Mathers]

It was a joke for God's sake! Lighten up - or have you have a sense of humour bypass?

[AxxisPhoto]

Your membership cards for "Measurebaters Anonymous" are in the mail. :D

LOL!:lol::lol:

[rdenney]

Yes, but does that help you to take better pictures?:D

Who the hell takes pictures?

Rick "a camera collector" Denney

[rdenney]

For shutter speeds, you are right: they aren't exactly different by 50%. And the camera does use those exact numbers: it doesn't show 1/15 and use 1/16 internally. These speeds were standardized a long time ago, long before computers and binary. Film has enough latitude that being off by 3% doesn't really affect exposure, so for convenience they didn't use a strict power-of-two sequence. Even with digital, a few percentage points either way doesn't make much difference.

Actually, shutter speeds use a factor of two, and apertures use a power of two. And that gives you the right to punch me in the nose if we every meet, heh, heh.

The series in current use came into vogue not all that long ago--perhaps the 60's. I have shutters dating from the 50's (example: an Ilex #4, and also my 1953 Rolleiflex) that are calibrated a half a stop off from the current series: 1/2, 1/5, 1/10, 1/25, 1/50, 1/100. Going back much further, f-stops were often marked have a stop of from the current sequence.

But it didn't matter. The speeds were set by a cam that looks like it was cut from the top of a tin can, if it's within a stop, no big deal.

I tested shutters in a range of East German and Ukrainian cameras a while back, and the actual shutter speeds were grossly different from their markings. Yet all those cameras had made wonderful pictures. For cameras that don't have electronically timed shutters, my temptation is to always use negative film, and save the transparencies for shutters that are either electronically timed or recently tested.

Rick "who needs to hose that Ilex shutter down with a cleaner" Denney

[adam8080]

Have you ever thought they might have started from 1/8000 going to 1 instead of 1 going to 1/8000? It sure does look a lot closer that way.

[number six]

Doesn't appear to be any rhyme or reason to the sequence on this 19th century field camera lens:

366946

[bsaber]

It doesn't matter in practice but I found this very interesting learning.

[Pmolan]

Engineers think that way. They can't help it. I hope thats what you do for a living Gilthanass.

[yogestee]

.




f/1 times 1.414 = 1.4
f/1.4 times 1.414 = 2
f/2 times 1.414 = 2.8
f/2.8 times 1.414 = 4
f/4 times 1.414 = 5.6
f/5.6 times 1.414 = 8



Which is the square root of 2.. This takes me back to my first year of my Diploma of Photography..

[Ceyber]

As an engineer, thank you for asking the question :)

Interesting discussion, good explanation by rdenney.

[timmyeatchips]

Gee, whiz, guys, can we just answer the question? If not, then resist that "submit" button...

Very informative post, but I don't think you actually answered either of the questions in the OP =)

Tim "Just being facetious" Hawkins

[rdenney]

Very informative post, but I don't think you actually answered either of the questions in the OP =)

Tim "Just being facetious" Hawkins

Yes, probably right, as I go back and read it. But the ensuing discussion of what is relevant or not (to an unwritten standard) seemed to devalue any desire for rigor. There's nothing wrong with artists understanding really simple mathematics of exposure settings, it seems to me.

And it has value, too. The square root of 2 not only works for aperture diameters, but it also works for lighting distances. Do you want to reduce the brightness of a studio light that is 6 feet from the subject by one stop? Move it to 8 feet (similar to going from f/5.6 to f/8). Two stops? Move it to 11 feet. Want a 2:1 balance between a main and a fill light? Put one light 8 feet away, and the other light 11 feet away (or 2.8 and 4 feet).

Rick "fun with math" Denney

[spkerer]

And it has value, too. The square root of 2 not only works for aperture diameters, but it also works for lighting distances. Do you want to reduce the brightness of a studio light that is 6 feet from the subject by one stop? Move it to 8 feet (similar to going from f/5.6 to f/8). Two stops? Move it to 11 feet. Want a 2:1 balance between a main and a fill light? Put one light 8 feet away, and the other light 11 feet away (or 2.8 and 4 feet).

Gee whiz, Rick! Now in addition to answering the original question, you go and make some incredibly useful statement of yet another way this whole topic can be applied.

Seriously, good description there!

[JackLiu]

I understand the need to know the theoretical and practical aspects of exposure (ISO, shutter speed, aperture) but I agree with sorarse that the end result is a properly exposed image as seen and as photographed.

[Gilthanass]

Thanks to a few posters that jumped in. I was getting quite frustrated by most of the responses (those being "who cares? It doesn't affect how you shoot anyway") and might have responded in a way that is difficult to take back, so I took some time to cool off.


Yes, probably right, as I go back and read it. But the ensuing discussion of what is relevant or not (to an unwritten standard) seemed to devalue any desire for rigor. There's nothing wrong with artists understanding really simple mathematics of exposure settings, it seems to me.

And it has value, too. The square root of 2 not only works for aperture diameters, but it also works for lighting distances. Do you want to reduce the brightness of a studio light that is 6 feet from the subject by one stop? Move it to 8 feet (similar to going from f/5.6 to f/8). Two stops? Move it to 11 feet. Want a 2:1 balance between a main and a fill light? Put one light 8 feet away, and the other light 11 feet away (or 2.8 and 4 feet).

Rick "fun with math" Denney


My original questions had everything to do with rigor, and little to do with effect on your exposure. I wanted to know whether the camera used the real exact doubling values, or the approximations it was showing you. It matters to me what it uses, not because I can use the information directly to improve my picture (ok, so I'm setting the apperature at 11, which really means 11.31, so if the camera is using 11 my exposure will be 3% overexposed compared to what the light meter is telling me is optimal....) but because I like to get in the mind of the designer, and understand the science behind it all. The more I understand first principles, and why design decisions were made, the more I can extrapolate later. Call it what you will, but it's how I work. Sure, some people only care that "it works", but I care HOW it works, and WHY it works, and yes, I care to know if it's an approximation or an exact value.

Engineers think that way. They can't help it. I hope thats what you do for a living Gilthanass.

Hahaha, I almost went into engineering. I did get a BSc though (Biology, though that doesn't lend itself to exactness nearly as much as the others), have done my share of computer programing, and have a very mathematical mind if that helps at all. And yes, I can't help but think that way.

Seriously-- why even bother with a response like this?

It's as if being ignorant of how things actually work is smiled upon these days.

No, of course it won't let me take a better picture of my cat so that I can post it on PotN. It will, however, teach me something that I didn't know before and that helps me better understand the science that lies underneath every single aspect of photography.

Fraknly, some of the "Why bother asking, since it just works out anyways" responses on the first page of this thread are rather depressing.

That is what made me the most frustrated at the responses to this thread, and made me need to take a self imposed breather from it. I understand that some people don't care how things work. I get that. What i don't get is how some of those people will actually look down on you for WANTING to know how they work. How, in all the world, is it in any way negative to want to understand how things work at the most detailed level? Do some of you understand that it is people with that mindset that design your sensors? That it is people that need to know, to the smallest level, how light behaves and how it reacts to different media that designed every one of your precious lenses? People who weren't satisfied with "light just bends when it goes through glass, I don't know why, it just works!", those are the people that make your hobby possible.

I wanted to know an answer to a question I was curious on. If you didn't care about the answer, no skin off my back, just move on. Don't tell me something I already know (that it makes little difference to your exposure). Anyway, I think I got my answer. The camera uses the exact aperature values (example: 5.656854249) but for simplicity's sake tells you it is "5.6". On the other hand, the camera DOES use the exact shutter speed it tells you (or attempts to), which is not a true factor of 2 (in many places). I still don't have a definitive answer to why that is, but I assume it has to do with ease of photographer use (1/125 is a more round number than 1/128). I don't know why they didn't set cameras to use the real doubling value while telling you an approximation, but perhaps it had to do with accuracy of the shutters, and has since just become industry standard. Does everyone agree that this information is correct (or, more appropriately, does anyone know anything that would suggest any of this information is incorrect)?

Thanks to those who understood the question and seemed willing to contribute (as well as those who understood the question and didn't arbitrarily decide what I should and should not want to know).

[bsaber]

^I think you hit it on the spot with the accuracy of the shutters and then became industry standard.

EDIT: I'm just speculating though.

[number six]

I don't know why they didn't set cameras to use the real doubling value while telling you an approximation, but perhaps it had to do with accuracy of the shutters, and has since just become industry standard. Does everyone agree that this information is correct (or, more appropriately, does anyone know anything that would suggest any of this information is incorrect)?

Basically, yes. Bearing in mind that historically the "standard" wasn't always applied. My Nikon F has a mechanical shutter with the usual marked speeds, but shutter speeds tended to vary by about 30% - and the last time I took it in for service it was off a whole stop (slow) at higher speeds.

-js

[rdenney]

I don't know why they didn't set cameras to use the real doubling value while telling you an approximation, but perhaps it had to do with accuracy of the shutters, and has since just become industry standard.

Even the standard series has diminished as a standard, with newbies not seeing the traditional series mixed in with those one-third-stop values.

But mechanical shutters just aren't that accurate. I tested a series of mechanical cameras in 2003, finding that shutter speeds were routinely off by a quarter stop or more one way or the other, even expensive cameras that had been recently serviced at the factory. Some cameras were significantly worse. Yet even the cameras that were 1/2 to a whole stop off had proven themselves in the field. What does that suggest?

One thing is that for most film materials, the film latitude is much greater than the shutter accuracy. Transparency films are the exception, needing an accuracy of perhaps a quarter stop one way or the other.

The other thing it suggests is that good photographers who used such equipment tested the equipment, either directly or photographically, and learned their errors and worked with them.

Finally, it suggests that we don't make pictures of gray cards. For most scenes, the dynamic range of the scene is such that such exposure errors generally go unnoticed. We will get significantly different meter readings depending on what portion (and what mix of light and dark) of the scene is included in the meter's field of view.

And meters themselves are also not particularly accurate. In another test, I compared a range of in-camera meters with a "panel" of high-quality professional meters, making readings of blank surfaces both bright and dim. The high-quality meters include standards of professional use: A Pentax Spotmeter V, a Minolta Spotmeter F, a Sekonic L-718, and a Gossen Luna Pro. Their readings were spread over a range of plus or minus a half a stop. The in-camera meters I was comparing against were spread over a range of a quarter stop one way or the other. Again, we learn what works with the equipment we have through trial or test.

That brings up the engineering principle of the difference between accuracy and precision. Precision is the fineness of the measurement, accuracy is its correctness. They are different concepts. Your question related to precision, and many of the responses related to accuracy. That's part of what caused the mis-communication. The reason scales are rounded to conventional values (as in, values accepted by convention) is that the accuracy of the shutters and meters wasn't sufficient to warrant greater precision.

But even if we had that accuracy (which we probably do with electronically timed shutters), the medium isn't that sensitive to it, so the required standard doesn't need greater precision.

The manufacturers, the ASA (back when it was the ASA), and the user community would have applied the engineering principle of significant figures. 1/125 is more correct than 1/128, because 1/125 implies less precision (it's "8 thousandths of a second" and is thus one significant figure) than 1/128 (which implies to the non-binary thinker that the difference between 125 and 128, which requires three significant figures, might actually be important). Accuracy demands at most a couple of significant figures, and thus 1/1000 is essentially the same number, in terms of significance, as 1/1024. "A thousandth of a second" (represented scientifically as 1 x 10^-3) is a different number than "0.001 seconds" (represented scientifically as 1.00 x 10^-3) because of the representation of significant figures. Significant figures are a critically important topic in engineering measurement, but with the advent of calculators that take everything to ten decimal places I'm afraid we've lost our way on the topic. But remember--engineers used to calculate things with slide rules, which were incapable of carrying more than about three significant figures through a calculation. I'm sure this is why f-stops are only reported to two significant figures in the conventional series.

Finally, a point about the convention: The expectation of shutter speeds following a binary series (8, 16, 32, 64, 128, 255, 512, 1024, etc.) is mostly based on the familiarity we modern folk have with that series. When the conventional shutter series became the de facto standard, nobody thought in those terms.

Rick "finally answering the question" Denney

[RDKirk]

Finally, it suggests that we don't make pictures of gray cards. For most scenes, the dynamic range of the scene is such that such exposure errors generally go unnoticed. We will get significantly different meter readings depending on what portion (and what mix of light and dark) of the scene is included in the meter's field of view.

For that matter, "exposure error" becomes a matter of taste within half a stop or so of whatever is "correct." Any of us should be able to discern as little as a quarter stop exposure difference, but within a variation of as much as a stop on either side of "correct," we may disagree on which is actually the "correct" exposure.

[PhotosGuy]

the "standard" wasn't always applied. Even for film of the same type. Different speeds & different color balance for different batches assuming that storage from the manufacturer to you was exactly the same.
Only pro films could be relied on to be close, & I won't go into the "creative" processing from some of the pro labs we tested.

[Gilthanass]

Even the standard series has diminished as a standard, with newbies not seeing the traditional series mixed in with those one-third-stop values.

...

Rick "finally answering the question" Denney

Thank you!

I agree that the difference between accuracy and precision was probably lost in the beginning of this thread. Perhaps my initial post was to blame for not being clear enough.

Anyway, your point about significant digits and human perception is very interesting, and I must say it makes a lot of sense. In a culture used to using the decimal system, numbers that are easily divided (or multiplied) into tens are seen as "approximations" where numbers that are not are expected to be more exact. So if I told you to drive for 100 kilometers (or 50, or 25) you would expect it to be AROUND 100 km, but if I said "drive for 64 kilometers" you would expect to get there as soon as you got to 64 km. Therefor, I can see how using numbers more closely related to ten for shutter speeds would communicate that it is close to that value (but not necessarily exact).

And for the record, I didn't pick the binary series just because I was familiar with it (being in the computer age). If you are going to create a doubling series based on the SI unit for sime (s) you really have no choice but to use a binary series, and that should have nothing to do with how familiar you are with it (though again, back to the point above, people not used to seeing those numbers would assume a higher level of precision in that case).

Anyway, thanks for your answer.

[rdenney]

And for the record, I didn't pick the binary series just because I was familiar with it (being in the computer age). If you are going to create a doubling series based on the SI unit for sime (s) you really have no choice but to use a binary series, and that should have nothing to do with how familiar you are with it (though again, back to the point above, people not used to seeing those numbers would assume a higher level of precision in that case).

You didn't, but if the convention were emerging today instead of 50 years ago, it might have followed the binary series just because people are used to it from playing with their computers. Nobody thinks twice about supplying a compact flash card as an 8, 16, or 32-Gig card, even though the real number might not even be related to a binary value. To the modern person, the binary series might have no more implication of precision than the old shutter speed series. But the binary series didn't have popular meaning until the late 80's at the earliest.

Rick "thinking we ought to report shutter speeds in hexadecimal" Denney

[Gilthanass]

Rick "thinking we ought to report shutter speeds in hexadecimal" Denney

Haha, I think I'm going to do that from now on.

"yeah, I don't like to handhold below 1/3C shutter speed"

"What's the flash sync speed of the 40D? 1/FA"