Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: 2nd Take on the ISO Units,  (Read 2199 times)
ThomasH_normally
Newbie
*
Offline Offline

Posts: 22


WWW
« on: October 15, 2009, 03:24:54 PM »
ReplyReply

I just found the original topic "closed," no comment. I suppose that the forum manager got upset about writers sliding from camera sensitivity into the issues of world history and human relations.

Please folks, lets keep it on the units. Personally I suspect that such discussions are already running among organizations and departments controlling the issue. Maybe some of the ideas posted here could provide an inspiration for a new unit of measure? I liked the interesting take by panopeeper numbering stops, this is a quote from his post:
Quote
0 -> 3 (accurately: 3.125)
1 -> 6 (accurately: 6.25)
2 -> 12 (accurately: 12.5)
3 -> 25
4 -> 50
5 -> 100
6 -> 200
7 -> 400
8 -> 800
9 -> 1600
10 -> 3200
11 -> 6400
12 -> 12800
13 -> 25600
14 -> 51200
15 -> 102400
Regarding DIN, I think that the disadvantage was the arbitrary choice of 18 degree for what we know as ISO 50, and lack of simple relation between shutter time and sensitivity. With ISO its simple, the ratio in ISO sensitivities corresponds to the ratio in shutter time, assuming the same aperture. Many photographers would also feel at home with steps, isn't that how we deal with aperture?

With the shift of the main innovation and manufacturing to Japan in 1950-1960, ASA/ISO scale became the scale of choice because it was chosen by leading manufacturers, and it served us well for half a century. Back in a film time, I used to automatically calculate the new shutter time before putting in that ISO 800 or ISO 1600 film in. We are very accustom to these values, just like we like to have the "equivalent 35mm focal length" of a lens. Thus I think that simple pruning of the hundred and preserving the same numerical values would have the advantage of being very intuitive to many.

Thomas

Logged
Chris Sanderson
Administrator
Sr. Member
*****
Offline Offline

Posts: 1899



« Reply #1 on: October 15, 2009, 03:44:07 PM »
ReplyReply

Quote from: ThomasH_normally
... Please folks, lets keep it on the units. ...
Thomas

Precisely!
« Last Edit: October 15, 2009, 03:44:44 PM by ChrisSand » Logged

Christopher Sanderson
The Luminous-Landscape
Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2823



WWW
« Reply #2 on: October 15, 2009, 04:53:56 PM »
ReplyReply

I've always assumed that the use of "ISO" and the corresponding values was simply an adaptation from the film world to make things easier for those moving to digital.

The downside to this was failing to understand at that time where it was going to go (with 6 figure ISO,s).  Imagine where we'll be if they get another few stops out of the sensor.

It would have been easier to make the switch in those days.

To me the scale should be simply about stops.  ISO 0 is ISO 100.  ISO 200 becomes  IS0 +1, ISO 400 becomes ISO +2, etc. Or as suggested ISO just goes away and a new term, such as Gain or Sensitivity is used.

Now this setting easily corresponds to the other two exposure values of f/stop and shutter speed.
Logged

NikoJorj
Sr. Member
****
Offline Offline

Posts: 1063


WWW
« Reply #3 on: October 16, 2009, 07:00:41 AM »
ReplyReply

Quote from: ThomasH_normally
With ISO its simple, the ratio in ISO sensitivities corresponds to the ratio in shutter time, assuming the same aperture.
[...]
Thus I think that simple pruning of the hundred and preserving the same numerical values would have the advantage of being very intuitive to many.
Yes, but simply pruning 2 zeroes off the old ASA scale causes two problems :
- you can't describe sensitivities below ISO100/21 with an integer,
- it's only pushing the problem further away, and there still is the discrepancy that high sensitivities are represented with huge numbers. What about when ISO819.200/60 is reached - it's only 3 stops from now?

OTOH, I finally agree that DIN scale is obsolete even if it's incorporated to ISO standard - the log principle is good, but the 1 stop = 3 convention just doesn't make it usable (it requires a division before thinking of stops, and you can't represent half stops with integers though).

For me at least, a log scale where one stop = one unit is the most intuitive and practical. NB : You wouldn't ask for a scale with 1.414 increments as apertures, would you?  
The problem of where to set the 0 (is the reference the K25 or a modern DSLR?) is not a big one, it just may eventually ask for negative numbers (it's the other advantage of a log scale).


Side note to manufacturers : this will warrant you big guys are really into the digital age, but don't neglect a few other useful evolutions away from the film era as the ETTR matrix metering mode and the raw histogram (and don't neglect the DMD either, film era also had its goods).   And Yes, Canon, you do know what's particularly missing with you.
Logged

Nicolas from Grenoble
A small gallery
Jonathan Wienke
Sr. Member
****
Offline Offline

Posts: 5759



WWW
« Reply #4 on: October 17, 2009, 11:38:40 AM »
ReplyReply

Simply substitute unit abbreviations for zeroes, the same as anywhere else numbers are used. 1K = 1000, 1M = 1,000,000, etc. so that ISO 12500 becomes ISO 12.5K, and so on. Why reinvent the wheel all over again?
Logged

Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2823



WWW
« Reply #5 on: October 17, 2009, 02:26:47 PM »
ReplyReply

Quote from: Jonathan Wienke
Simply substitute unit abbreviations for zeroes, the same as anywhere else numbers are used. 1K = 1000, 1M = 1,000,000, etc. so that ISO 12500 becomes ISO 12.5K, and so on. Why reinvent the wheel all over again?

I guess to me just because the scale is representing a logarithmic sequence, the scale itself doesn't have to be logarithmic.  Simply truncating the 0's like we have done with data storage would be something that might happen if everyone started doing it, nothing the camera makers could do.  It still isn't as simple having the numbers each represent a single stop.  That's just me.

Of course, this is all assuming very high ISO's become very common place.  Even having them available I rarely use them.  Some photographers may use them frequently, but I'm guessing very few use higher than ISO 1600 very often.
Logged

Jack Flesher
Sr. Member
****
Offline Offline

Posts: 2595



WWW
« Reply #6 on: October 17, 2009, 03:45:32 PM »
ReplyReply

I like having actual ISO numbers.  Main reason is for estimating exposures if your meter bricks.  Some of us are old enough to remember the "Sunny 16" rule and how it uses ISO to determine an appropriate shutter speed...

,
Logged

imagico
Newbie
*
Offline Offline

Posts: 47


WWW
« Reply #7 on: October 18, 2009, 02:29:00 AM »
ReplyReply

Quote from: ThomasH_normally
Regarding DIN, I think that the disadvantage was the arbitrary choice of 18 degree for what we know as ISO 50, [...]

Actually it is not arbitrary:

DIN = 10 log10 ASA

This does not completely comply with the known relations (100 ASA would be 20 DIN, not 21) but this is probably due to the rounding applied in the ISO standard.

I am not so sure if a logarithmic scale will become the common measure for sensitivity.  First both shutter speed and aperture are not commonly set in a logarithmic measure - shutter speed is linear and aperture is a square root measure so sensitivity would be the only one done on a logarithmic scale.  And there will most likely not be a continuous inflationary growth in sensitivity of sensors.  Someone might want to do the math on how many photons/ADU a D3s sensor collects at ISO 12800 - i suppose it is already quite near the maximum of what makes sense.
Logged

Christoph Hormann
photolog / artificial images / other stuff
thierrylegros396
Sr. Member
****
Offline Offline

Posts: 647


« Reply #8 on: October 18, 2009, 11:46:32 AM »
ReplyReply

Quote from: Jonathan Wienke
Simply substitute unit abbreviations for zeroes, the same as anywhere else numbers are used. 1K = 1000, 1M = 1,000,000, etc. so that ISO 12500 becomes ISO 12.5K, and so on. Why reinvent the wheel all over again?

That's exactly what I thought !

In whole electronic we use that manner, for example 1mH Inductor is 0.001 Henry, 1F Capacitor is 0.000001Farad, 1MOhm Resistor is 1000000 Ohm.

More, everybody is now accustomized with ISO units, so why change ?!

Have a Nice Day.

Thierry
Logged
BJL
Sr. Member
****
Offline Offline

Posts: 5124


« Reply #9 on: October 19, 2009, 03:59:49 PM »
ReplyReply

I also had the thought of a logarithmic scale, like reviving the old German DIN scale. But maybe a simpler modern option would be to talk in stops of speed; say stops over (or under) 100. ISO 102,400 is just +10 over ISO 100.
An "ISO" scale that now reads ISO 50, 100, 200, 400, 800 ...
could be replaced by
a "sensitivity" or "exposure index" scale the reads -1, 0, +1, +2, +3 ...
This could simplify reckoning of reciprocal exposures: "I want to close the aperture down by two stops, so have to up the sensitivity by two stops".

But Jonathan W. might well be right: the only scale of powers we are likely to see are the slightly ill-fitting powers of 1000: ISO 12K, 24K, 50K, 100K ...
I severely doubt that ISO 102,400 is accurate to all digits anyway.


P. S. Whoops, I just saw Wayne Fox's post on the same idea. So .... Great suggestion, Wayne!

P. S. I would also be happy to dump the strange system of naming sensitivity by the acronym of the organization that publishes some standards for this, amongst thousands of others. For example,  "ISO 9000" is not a film speed, it is a quality management standard!
« Last Edit: October 19, 2009, 04:07:05 PM by BJL » Logged
Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2823



WWW
« Reply #10 on: October 19, 2009, 07:25:34 PM »
ReplyReply

Quote from: BJL
I also had the thought of a logarithmic scale, like reviving the old German DIN scale. But maybe a simpler modern option would be to talk in stops of speed; say stops over (or under) 100. ISO 102,400 is just +10 over ISO 100.
An "ISO" scale that now reads ISO 50, 100, 200, 400, 800 ...
could be replaced by
a "sensitivity" or "exposure index" scale the reads -1, 0, +1, +2, +3 ...
This could simplify reckoning of reciprocal exposures: "I want to close the aperture down by two stops, so have to up the sensitivity by two stops".

But Jonathan W. might well be right: the only scale of powers we are likely to see are the slightly ill-fitting powers of 1000: ISO 12K, 24K, 50K, 100K ...
I severely doubt that ISO 102,400 is accurate to all digits anyway.


P. S. Whoops, I just saw Wayne Fox's post on the same idea. So .... Great suggestion, Wayne!

P. S. I would also be happy to dump the strange system of naming sensitivity by the acronym of the organization that publishes some standards for this, amongst thousands of others. For example,  "ISO 9000" is not a film speed, it is a quality management standard!

lol ... nice to know someone else sees this the same way I did ... it all just seems so logical and simple.
Logged

Gordon Buck
Sr. Member
****
Offline Offline

Posts: 409



WWW
« Reply #11 on: October 19, 2009, 08:53:44 PM »
ReplyReply

A very interesting discussion.  As an engineer, the numerical value for sensitivity does not particularly matter to me as long as I can understand the basis.  As a photographer, the concept of "Sunny 16" is very useful; however, the meaning of extremely high ISO sensitivities begins to get out of the range of my own mental arithmetic.  Seems to me that some sort of Sunny 16 basis within "normal" camera limits would be useful.  At higher (or lower) ISO limits, the sensitivity scale could be adjusted by +/- f-stops.   Suppose, for example, that the basis was ISO 1600 (exposure of f16 at 1/1600 second in bright daylight); in that case, ISO 3200 would be "ISO 1600-1" meaning that the strong daylight exposure was one stop less exposure than that for ISO 1600 (for example, f16 at 1/3200 second).  Those knowlegible of the details would read this as "ISO 1600 less one stop" but most people would read this as "ISO 1600 dash 1".  ISO 6400 would be "ISO 1600-2", etc.  ISO 400 would still be ISO 400, etc.  


A variation would have ISO 3200 as "ISO 1600+1" meaning that the sensitivity would be ISO 1600 plus one stop.  This would probably be more acceptable to manufacturers as it would appear to be a more positive statement.  It's easy to imagine photographers then speaking of, for example, "plus three" sensitivity once the basis has been defined.  






Logged

Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad