7D users, noise question...

Started Mar 1, 2014 | Discussions thread
Jerry-astro Veteran Member • Posts: 9,719
Different eyeballs - different result

Dale, I did test this myself a while back and got the opposite result, with the pulled ISOs being ever so subtly less noisy than their native counterparts. However, as you stated, the differences are so subtle that they are all but inconsequential and (obviously based on our different results) very open to interpretation. I'm typically used the "pulled" multiples these days based on the limited testing I did, but I don't see it as a game changer in any real way.  BTW, all the test images I took were with the lens cap off. :-). Figured that this was slightly more typical of what I shoot in the real world -- at least intentionally.

You're also quite correct that the "native" ISOs are most certainly 100/200/400/etc. and I have never seen Canon refer to the 160/320 multiples using that terminology.

Dale Buhanan wrote:

TTMartin wrote:

AmbleYonder wrote:

native ISOs 160,320,640......=less noise

The native ISOs are actually 100, 200, 400, etc.

160, 320, 640 are actually digital pulls from the next higher ISO. They have less noise because they provide an automatic 1/3 stop ETTR.

What you really want to avoid are the ISO 125, 250, 500 etc. as these are digital pushes from the next lower ISO. This digital push is written into the RAW file and impacts even RAW shooters.

What TTMartin has said is true. The native ISO are 100, 200, etc. And the 160 multiples are digital pulls from the next higher native ISO, give an ETTR effect.

But that is not the end of the story. Hence the reason for my post.

There are two major component that make up the noise: 1) The first is the electronics noise, or read noise. 2) The second is shot noise from the incident light.

The electronics noise is caused by the readout electronics and subsequent digital and analog multiplications. This is the component that is measured with the lens cap on ( as the test data was) and is reduced by shooting at the 160 multiples. However the other component is not.

When I first got my own 7D I did my own testing to see if my data matched the reported natice ISO and 160 multiple pulled ISO. With the lens cap on, I got the same results. But -- and this is a big but -- when shooting real world photos where both noise components were present I did not.

What I saw was the the combination noise was best at the native ISO's. ie., 100 gave less noise than 160. and 200 gave less noise that 320, even though with the lens cap on tests, the 160 gave less noise than 100.. The shot noise added enough more noise that for real world shooting, native ISO's were still best. So I got a sort of sawtooth noise diagram where noise always increased either more or less with increasing ISO -- in real world shooting. Each succeeding ISO increase also gave more noise. But with the pushed ISO values, the amount was more than with the pulled ISO values, even though both were more than the native ISO values.

Try the test yourself and be convinced. However, bottom line, the difference in noise between ISO 100 and ISO 160 isn't very much, 100 is still better for real pictures, and 160 is best with the lens cap on.

-- hide signature --

kind regards

 Jerry-astro's gear list:Jerry-astro's gear list
Fujifilm X-T2 Fujifilm X-H1 Fujifilm XF 18-55mm F2.8-4 R LM OIS Carl Zeiss Touit 2.8/12 Fujifilm XF 56mm F1.2 R +3 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow