Login or Join Now

Upload your photos, chat, win prizes and much more

Username:
Password:
Remember Me

Can't Access your Account?

New to ePHOTOzine? Join ePHOTOzine for free!

Like 0

Smart ISO - when a higher ISO rating is LESS noisier than a lower one

Join Now

Join ePHOTOzine, the friendliest photography community.

Upload photos, chat with photographers, win prizes and much more for free!

New PortraitPro 12 SALE + 10% OFF code EPZROS814
adyj
adyj  530 forum posts United Kingdom
13 Dec 2012 - 5:14 PM

I was looking into this debate about so-called 'native' ISO settings on DSLRs and why multiples of 160 supposedly offer the lowest noise levels. Maybe I'm late to this debate - there are pages of (some contradictory) discussion out there. Wanted to find some clarity over this.

First off, selecting 160 multiples does seem to produce less noise, going by this test from Andrew Schär. But there had to be a downside to this of course and it seems loss of dynamic range, specifically highlight detail, is the compromise. This doesn't show in aforementioned test but is best explained here by film maker Josh Silfen.

In a nutshell, the real native ISOs (as most of us always assumed) are the 100 multiples and 160, 320, 640, 1250 etc are all 1/3rd stop pulls of the next and greater ISO rating - ie. 160 is 200 pulled back or underexposed 1/3rd of a stop. Settings like ISO 125, 250.. are a result of digital 1/3rd stop pushing and will give more noise. ISO choice used to be so simple!

OK, this is less important to stills shooters with the option of RAW processing though it could affect your choice over shutter/ISO combination in low light situations. Likewise, when highlight detail is paramount, you might decide to avoid those 160 multiples.

Clearly situation dependant but choosing the right ISO was never so critical.

Last Modified By adyj at 13 Dec 2012 - 5:17 PM
Sponsored Links
Sponsored Links 
13 Dec 2012 - 5:14 PM

Join ePHOTOzine for free and remove these adverts.

User_Removed
13 Dec 2012 - 7:02 PM

The "160 multiples" myth is pure garbage, with no scientific foundation whatsoever.

The "native ISO" argument is better founded although, in any meaningful context, a little outmoded.

Most modern dSLR sensors perform well at anything in the ISO range 100 - 400. What really matters is the degree of quality drop-off above that and (an often overlooked point) the significance and nature of any such drop-off. Theoretically there may be a drop-off in quality below the "native ISO" but, to be frank, it is totally insignificant for all serious purposes.

If I can use my Nikon D800 as an example. The first step up the ISO scale at which I will notice any IQ degradation at all is 1600. But, for virtually all purposes, that is inconsequential. Going up to 6400, the effect is more noticeable at 100% (but who cares?) and the more important factor is that, at 6400, the degradation looks like "grain" rather than "noise" (a sharp difference from the D700, D300 or D3s, for example - or any Canon I have examined). Above 6400, the D800 IQ does deteriorate noticeably (and more seriously than the D3s).

With a camera like the D800, with its 14.3 EV dynamic range - when coupled with the amazing highlights and shadows controls in Lightroom 4.3 (and, I assume, current ACR version), ISO actually becomes of very marginal importance for 90% (or more) of all serious photography. That, I suggest, is one of the relatively few real benefits from dSLR design in the past 12 months.


.

Last Modified By User_Removed at 13 Dec 2012 - 7:05 PM
lemmy
lemmy  71831 forum posts United Kingdom
13 Dec 2012 - 8:20 PM


Quote: there may be a drop-off in quality below the "native ISO" but, to be frank, it is totally insignificant for all serious purposes

I was happy to read that, Leftforum. Sometimes think I'm the only person who thinks that way.

I earned my living with film, the best of it being Kodachrome with a dynamic range of about 8 stops. I recently went over to M43 from DSLR for weight reasons (camera, not mine!). My current Panasonic GH2 has a dynamic range of about 11 stops, Leftforum's D800 has 14 stops....

The performance of modern cameras has gone far beyond the needs of any non-scientific photographer. adyj says choosing the right ISO was never so critical. I'd say it never mattered less.

There are few occasions now where sticking the camera on auto everything yields anything less than optimum results.

User_Removed
13 Dec 2012 - 9:25 PM


Quote: adyj says choosing the right ISO was never so critical. I'd say it never mattered less.



That's what I was trying to say a wee bit more diplomatically, Lemmy.

Wink

lemmy
lemmy  71831 forum posts United Kingdom
14 Dec 2012 - 11:03 AM


Quote: That's what I was trying to say a wee bit more diplomatically, Lemmy.

I did gather that and very nicely done too. But that remark is undiplomatic about my being undiplomatic and I am now traumatized Smile

Seriously, one of the downsides of modern cameras is the immediacy of results, giving photographers the chance to test, retest and obsess over technical things. Using film, doing tests was a tedious and expensive business and only the real nerds could be bothered. Nowadays we can all be nerds.

Nothing wrong with that, I love modern technology with a passion but I do accept that it has little to do with photography as a means of making pictures.

JohnParminter
14 Dec 2012 - 1:48 PM


Quote: There are few occasions now where sticking the camera on auto everything yields anything less than optimum results.

This isn't true in my case and my D300 for example. I don't know about anyone else's camera but mine is quite a half-wit, especially in Auto.

Just the other day I was thinking how I wanted to take a shot and what I wanted to achieve, I gave my camera the chance to shine so momentarily switched it to auto and blow me it buggered things up. For some unknown reason it can't have understood what my creative intention was, I mean, in this day and age you would think that all this modern technology would be able to determine how I wanted my DOF, where I wanted focus, what the colour balance should be, how I wanted the flowing water to look but no, it completely lost the plot. I quickly bunged it back in full manual I can tell you, back to dumb mode where it shall remain forever under my guiding eye.

Wink

strawman
strawman  1022006 forum posts United Kingdom16 Constructive Critique Points
14 Dec 2012 - 2:00 PM

Perhaps it is because I am using older kit, but on my 40D the noise is less if I keep it in whole ISO steps (i,e 100, 200, 400 etc). The plausible reason I was given was that for the whole steps it has amplifier gain steps, but for the middle ones it under exposes then corrects in software. i.e ISO600 would be taken at ISO800 then SW corrected so you may as well have gone to ISO800 as it will have lower noise. Perhaps the more modern sensors do not suffer from this.

Having said all that on my compact I have it set to AUTO ISO with a limit on the maximum and shoot it in aperture priority mode and it does just fine in the vast majority of cases.

Is it the biggest issue, well no and as every 3 years or so they knock sensor noise down another level it is becoming much less of an issue.

User_Removed
14 Dec 2012 - 4:09 PM


Quote: Perhaps it is because I am using older kit, but on my 40D the noise is less if I keep it in whole ISO steps (i,e 100, 200, 400 etc). The plausible reason I was given was that for the whole steps it has amplifier gain steps, but for the middle ones it under exposes then corrects in software. i.e ISO600 would be taken at ISO800 then SW corrected so you may as well have gone to ISO800 as it will have lower noise..

That may have sounded plausible at the time, Strawman - but if you think about it, it can't be correct.

If it were the case, then you would see no change in shutter speed (if using Aperture Priority) or in aperture (if using shutter priority) when changing the ISO from one intermediate step to the next or to the next full step. And that's not the case. In every dSLR I have had (D80, D300, D3s, D800), if you moved the ISO setting by one-third of a stop, then the actual exposure moved one-third of a stop as well. The D80 was probably a similar vintage to your 40D.

.

Last Modified By User_Removed at 14 Dec 2012 - 4:10 PM
mikehit
mikehit  56312 forum posts United Kingdom9 Constructive Critique Points
14 Dec 2012 - 4:36 PM

I have heard the same explanation. If you go from ISO 100 to ISO 160, you will go f5.6 to f6.3 to keep the same shutter speed. But if the ISO has actually not changed the image is underexposedand the camera software will 'boost' it with in-camera processing.

The same explanatoiun is given for ISO 50 when the native ISO is actually 100 (in other words, the ISO 50 is a sham, doing in camera what you would otherwise do in post processing).

adyj
adyj  530 forum posts United Kingdom
14 Dec 2012 - 8:09 PM

Hey, you got me. I'm clearly not nerdy enough for these subjects Grin Like I said I'm late to this debate.

Never heard about this 160 'rule' until I watched a Philip Bloom interview recently on shooting video with the 5DMKII. I suppose this whole issue only came to light because of the concerns of film makers. Perhaps Bloom was the original source of this 'myth'. Though to be fair, his advice related specifically to the 5DMKII and in video mode. But the discussion elsewhere was not brand or model specific.

The multiple thing did sound spurious and if I wasn't interested in shooting video I wouldn't have dug any further. Like stills shooting in JPEG, Canon DSLR video doesn't allow a lot of headroom in post so anything that advantaged me in this area interested me.

By critical I didn't mean essential but in the sense that, where ISO was a consideration for the photographer, more care may be required in the light of this information. This clearly may only be relevant, of course, to those of us using vintage Canons like the 5DMKII Wink And may not be relevant to anyone if much of what I've read about this on the web is bollocks!

ISO probably won't be a deal breaker when shooting RAW but thought the info worth posting - if just to stir the nerds amongst us Smile

Last Modified By adyj at 14 Dec 2012 - 8:11 PM
lemmy
lemmy  71831 forum posts United Kingdom
14 Dec 2012 - 8:33 PM


Quote: if just to stir the nerds amongst us

The anti-nerds need the nerds and vice versa, all part of life's rich pattern.

I'm an anti-nerd. Though nerdy things interest me, so much of the technical twaddle you read is the equivalent of modern psycho-babble, just meaningless outpourings with no relationship to the real world.

I used to get fed up with all this technical stuff from staff in a pro shop I used in London. I'd go in to buy something and the assistants (nice guys, by the way) would assault you with technical stuff of why this was better than what you wanted to buy. I used to wonder why, if they knew so much more than me, I was the one travelling all over the world with my cameras and not them.

I still don't know why but I did learn that good technicians are not necessarily good photographers and vice-versa. The faux technicians and obsessive pixel peepers you read today are probably neither. They still enjoy themselves, though, which is the ultimate point for any amateur (ie someone who does something for the love of it).

I just don't think they should be taken seriously.

adyj
adyj  530 forum posts United Kingdom
14 Dec 2012 - 8:43 PM


Quote: I still don't know why but I did learn that good technicians are not necessarily good photographers and vice-versa. The faux technicians and obsessive pixel peepers you read today are probably neither. They still enjoy themselves, though, which is the ultimate point for any amateur (ie someone who does something for the love of it).

Absolutely and nicely put.

Paul Morgan
Paul Morgan e2 Member 1315182 forum postsPaul Morgan vcard England6 Constructive Critique Points
14 Dec 2012 - 9:01 PM


Quote: I still don't know why but I did learn that good technicians are not necessarily good photographers and vice-versa

Good printers were worth there weight in gold Smile

http://www.bjp-online.com/british-journal-of-photography/news/2128428/celebrated...

Last Modified By Paul Morgan at 14 Dec 2012 - 9:03 PM
lemmy
lemmy  71831 forum posts United Kingdom
14 Dec 2012 - 9:12 PM


Quote: Good printers were worth there weight in gold

They sure were.

strawman
strawman  1022006 forum posts United Kingdom16 Constructive Critique Points
14 Dec 2012 - 11:23 PM

These things tend to only be of interest if you are at the edge of the performance you are trying to achieve. so for me up to ISO1600 is usually acceptable, beyond that I tend not to like the results with my current camera. No doubt a new one is a lot better, and for my compact anything up to ISO800 is usable. that is why on the compact I let it run on auto ISO limited to 800 and worry about choice of aperture or composition etc.

As for how the "simulation" of in between ISO stops would work, the camera would take the shot at the lower ISO and calculate the shutter speed to be at the "artificial ISO". in that case the shot would be under exposed so in software it would apply an exposure compensation, not unlike applying a fixed exposure adjustment in a photo editor piece of SW. You end up with the correct exposure and because you have amplified it more you will have pushed up the noise, especially ADC converter steps.

Its not worth getting hung up on, I did some shots to test and found that on the 40D it was better at ISO1600 than at the ISO steps between 800 and 1600. Perversely on my compact it looks to make sod all difference. It could even be a foible/feature of the 40D.

Add a Comment

You must be a member to leave a comment

Username:
Password:
Remember me:
Un-tick this box if you want to login each time you visit.