Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
07-15-2009, 03:59 AM   #1
Senior Member




Join Date: May 2009
Posts: 124
ISO values such as 280, 1100, etc.

Probably a foolish thing to ask: does using ISO values like 280, 1100 (not typically 100, 200, 400, etc.) give any negative side effects?

07-15-2009, 04:26 AM   #2
Veteran Member




Join Date: Jan 2008
Location: Ste-Anne des Plaines, Qc., Canada
Posts: 2,014
No. It's just a convenience for when you need to boost the ISO, but don't need a full step.
07-15-2009, 04:30 AM   #3
Veteran Member




Join Date: Jun 2008
Location: Borås, Sweden
Photos: Gallery
Posts: 3,165
I don't know about Pentax, but at least on the Canon lower end cameras those steps are software generated and actually have worse noise than the higher "real" ISO. E.g. ISO 500 and 640 look worse than 800.
07-15-2009, 05:33 AM   #4
Veteran Member
RBellavance's Avatar

Join Date: Nov 2006
Location: Near Montréal, Canada
Photos: Gallery
Posts: 1,716
QuoteOriginally posted by pingflood Quote
I don't know about Pentax, but at least on the Canon lower end cameras those steps are software generated and actually have worse noise than the higher "real" ISO. E.g. ISO 500 and 640 look worse than 800.
I have read this several times, but I have never seen conclusive evidence of it. By curiosity, do you have a reference for this ?

07-15-2009, 05:39 AM   #5
Veteran Member




Join Date: Jun 2008
Location: Borås, Sweden
Photos: Gallery
Posts: 3,165
Here's some info on it..

Canon 40D Noise Level vs. ISO - Canon Photography Group

Interestingly on the 40D the "real" ISOs seem to be 160, 320, 640...
07-15-2009, 08:40 AM   #6
Senior Member




Join Date: May 2009
Posts: 124
Original Poster
Interesting.
07-15-2009, 08:51 AM   #7
Veteran Member
RBellavance's Avatar

Join Date: Nov 2006
Location: Near Montréal, Canada
Photos: Gallery
Posts: 1,716
QuoteOriginally posted by pingflood Quote
Here's some info on it..

Canon 40D Noise Level vs. ISO - Canon Photography Group

Interestingly on the 40D the "real" ISOs seem to be 160, 320, 640...
Very interesting. Now all we need is for someone to do the same experiment with Pentax cameras.
07-15-2009, 10:12 AM   #8
Veteran Member




Join Date: Jun 2008
Location: Borås, Sweden
Photos: Gallery
Posts: 3,165
Actually, the more I am reading about this (doing a little research) it looks like the reason ISO 160 is cleaner than 100 on the 40D is... exposing to the right. No, not joking. It seems that 160 is 200 overexposed a tad and then pulled back. If one is to believe the one guy who did VERY extensive research on sensor noise, that is just in line with what's to be expected.

ISO 125 is then generated by slight underexposure of ISO 100. So, more noise.

The down low: looks like overexposing ISO 200 by 1/3 stop and then reducing 1/3 in processing results in a better image than ISO 100 exposed normally.

07-15-2009, 05:54 PM   #9
Veteran Member
Duck Dodgers's Avatar

Join Date: Dec 2007
Location: in the 24½th Century!
Photos: Gallery
Posts: 439
Well, to my understanding, all sensor ISO values are just the base sensor ISO amplified. Seems odd that engineers would boost it past and then pull it back. At least you'd think that the software would only boost-pull where there was an advantage to do so. Strange.
07-16-2009, 12:54 AM   #10
Veteran Member
StarDust's Avatar

Join Date: Apr 2008
Location: Uppsala, Sweden
Posts: 400
QuoteOriginally posted by Duck Dodgers Quote
Well, to my understanding, all sensor ISO values are just the base sensor ISO amplified. Seems odd that engineers would boost it past and then pull it back. At least you'd think that the software would only boost-pull where there was an advantage to do so. Strange.
That's what I've always thought. A simple test would be to take expose at M mode at ISO 100 and then boost to 200 to then go back to 100 in PS. Might not be the most professional way but should give quite accurate result in noise level in my opinion.
07-16-2009, 01:11 AM   #11
Ash
Community Manager
Loyal Site Supporter
Ash's Avatar

Join Date: Mar 2007
Location: Toowoomba, Queensland
Photos: Gallery | Albums
Posts: 22,679
I've been hard-pressed to find a reliable source on this phenomenon in Pentax cameras, so if anyone out there has such info at hand it would be great to post it here. It would certainly change the way I would set ISO...
07-16-2009, 01:32 AM   #12
Damn Brit
Guest




This is more specific to DSLRs than photography technique, thread moved.
07-16-2009, 02:10 AM   #13
Veteran Member




Join Date: Dec 2007
Posts: 923
QuoteOriginally posted by Duck Dodgers Quote
Well, to my understanding, all sensor ISO values are just the base sensor ISO amplified. Seems odd that engineers would boost it past and then pull it back. At least you'd think that the software would only boost-pull where there was an advantage to do so. Strange.
The designed hardware sensitivity of a digital sensor is fixed. I think the term used is "Native ISO" and for the K-7 and K20D I believe its at ISO200.

All other ISO sensitivity settings are based on amplification of the analog signal before it gets converted to digital. That's why High-ISO photographs can get so noisy - the weaker signal from a dimly lit subject is inherently more noisy and the noise gets amplified as well.

To keep the circuitry simple, most (but not all) digital sensors have only a limited number of amplification points (like individual clicks on a control knob). Usually these follow full f-stop steps e.g. ISO 200, 400, 800, 1600, 3200.

Anything in between e.g. ISO 280, is an "artificial" creation done by further signal processing

Whether the design is based on boosting it to the next "click" past the intermediate sensitivity and pulling back, or boosting it just below the intermediate sensitivity and boosting it further digitally, I think only the internal techies know. Its not published anywhere, as far as I know.

There's a very good article in Wikipedia which explains both Film and Digital ISO.

Film speed - Wikipedia, the free encyclopedia
07-16-2009, 02:22 AM   #14
Veteran Member
ytterbium's Avatar

Join Date: Jan 2008
Posts: 1,076
If it was done in software you should see it in raw file, righ?
I mean it would fake shown ISO value, and be already boosted most likely, but DR would change.
I think it should be more precise measure than noise, since those graps could show some low level noise reduction kicking in at each stop.

If you shot a dark (low light, not black) scene at native iso100, where something barely shows up in least significant bits in raw files, by actually underexposing such shot at some equivalent iso160, you'd most likely decreased the light level under the ADC's range and couldn't boost that detail back in software, where ar "hardware" iso200 it would become visible again.

Similar could be happen with highlights, if the camera actually overexposes and compensates in software: Highlight detail would be less at iso125 than at iso200 (where it dosnt have to overexpose and compensate in software).

I guess it would be most easy to do with some kind of high contrast gradient shot, and then observe how the clipping at each end moves when changing ISO.

But at those bit depths (14..22..12) one must very carefully control the lighting to not make a random value error.

Here i found the PCB of k200d, can someone make out the labels on IC's?
http://picasaweb.google.com/lh/photo/0fV232RvYS34ByDCkDdhVQ
Then you could just have a look in the datasheet to see it there is some programmable gain amplifier and what values it can easily deliver.

Last edited by ytterbium; 07-16-2009 at 02:31 AM.
07-16-2009, 10:23 AM   #15
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
QuoteOriginally posted by ytterbium Quote
If it was done in software you should see it in raw file, righ?
Yes. In the most basic implementation, you'd see "holes" in the RAW data: certain pixel values that never occured. For instance, if the original digital output of the A/D converter were 0, 1, 2, 3, 4, 5, 6, etc., then the results of a software push by half a stop would be 0, 2, 3, 5, 6, 8, 9, etc. That assumes we are always rounding up: eg, 1 * 1.5 = 1.5, which I rounded up to 2; similarly, 3 * 1.5 = 4.5 which I rounded up to 5. The "holes" would be 1, 4, 7, etc. So if you look at the RAW data and don't see any 1's, 4's, 7's, etc., then you'll know it was done in software. but I don't think you could conclude from the "lack" of such holes that it *wasn't* done in software after all, it could simply have been a more clever algorithm that didn't simply round everything up.

Anyhow, I have always wondered about this too, and tend to avoid those settings. But really, that's kind of silly, as a software push of that amount could realistically be just as effective as using the analog amplifier, and of course, if I end up a little underexpose before I'm using half a stop too low of an ISO, I'm doing my software push anyhow in PP.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
camera, dslr, iso, iso values, photography, values
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Nature Осень 001. Братья. drmartin Post Your Photos! 7 09-12-2010 05:38 AM
13,280 x 9,184 pixel Canon Sensor bikecoboss Photographic Technique 7 08-25-2010 04:43 PM
pentax 280 flash fred412 Flashes, Lighting, and Studio 1 06-14-2010 07:06 AM
Values in the US bombo General Talk 151 05-17-2010 08:30 AM



All times are GMT -7. The time now is 02:53 AM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top