Canon M6 Mk II - Simulated ISO and "Real" ISO

I read in one Thread here about M6 MK II´s tips that there is something called ISO invariance. Some search later I think I got the point that real ISO invariance is when the Camera has almost no Backend Read Noise. Which is a good thing.

But here in the Forum the term seems to be used to describe when the camera is using simulated ISO. So the Frontend is not able to apply more voltage to the pixels and thus it raises the numbers just like a RAW Editor would do. Thus sometimes decreasing our Dynamic Range. Which is a bad thing.

Is my understanding so far correct?

Then I searched for our beloved M6 Mk II specifically. What are our real ISO values without losing Dynamic Range? Are our in between ISO stops real or also just simulated?

I read this quite deep article (https://photographylife.com/iso-invariance-explainedhttps://photographylife.com/iso-invariance-explained). In it the author linked to this Photons to Photo Graph.

https://photonstophotos.net/Charts/RN_e.htm

I would assume from this Graph that the in between ISO of the M6 II is simulated, because they are lying on a ladder with a full stop in ISO they are close to. Also it seems the last real ISO would be 6400, because after that the ladder effect stops. But I read here in the Forums and elsewhere that the M6 II could have real ISO only up to 400. That would be a major difference and I believe I just interpreted the data incorrect.

Then I found a link to this second Graph.

https://www.photonstophotos.net/Charts/PDR.htm

Here it would indeed show 400 or 800 as the last real ISO. But there is no way to tell what is happenig with the in between ISOs.

I would be very glad if someone with a deep understanding could enlighten me in this matter. Especially about these two questions regarding the M6 II:

What are our real ISO values without losing Dynamic Range? Are our in between ISO stops real or also just simulated?

Additionally I came across some other real ISO values for other cameras. And people were happy when their camera had only a real ISO of 100, saying somethin like 6400 or 3200 would be bad. I don´t understand this at all.

Shouldn´t be a real ISO of up to 6400 considered to be the best? As in this casse the Frontend has more flexibility to adjust for the incoming light (or lack) instead of pushing it like a editor. And a real ISO of only 100 would mean that everything else is simulated which is bad.

Thank you for this great forum and for any help that will come in this Thread.

To tell the truth, since I’ve started using DxO Photolab (5 and now 6) as my RAW converter I don’t pay much attention at all to the ISO setting, just so long as it’s not getting crazy high.

DxO is the Great Equalizer.

R2

Don't overthink this.  You can be sure that the Canon engineers know what they are doing, and chances are, they have done a good job on this.

R2D2 wrote:

To tell the truth, since I’ve started using DxO Photolab (5 and now 6) as my RAW converter I don’t pay much attention at all to the ISO setting, just so long as it’s not getting crazy high.

DxO is the Great Equalizer.

R2

Definitely. I took thousands of shots in a dimly lit gym on Saturday with my R7 (very similar sensor to the M6II). Quite a few were at ISO 25,600 and they looked great with DXO Deep Prime XD.

Alastair Norcross wrote:

R2D2 wrote:

To tell the truth, since I’ve started using DxO Photolab (5 and now 6) as my RAW converter I don’t pay much attention at all to the ISO setting, just so long as it’s not getting crazy high.

DxO is the Great Equalizer.

R2

Definitely. I took thousands of shots in a dimly lit gym on Saturday with my R7 (very similar sensor to the M6II). Quite a few were at ISO 25,600 and they looked great with DXO Deep Prime XD.

DXO also works really well with G7 X II.

-- hide signature --

“When I die, I want to go peacefully in my sleep like my grandfather. Not screaming in terror, like the passengers in his car.” Jack Handey
Alastair
http://anorcross.smugmug.com
Equipment in profile

Spherical wrote:

Thank you to all for replying. Sorry for not answering for so long, I kind of lost the train of thought in this topic. Still for everyone interested I found the graph which shows where the M6 Mk II is most accurate. And I made some test photos to see if it is really easy to recover the shadows when using such a low ISO as 400.

Maybe this is only really important for night sky shots, but I find it is good to understand the gear in and out. And on that basis of thourough understanding I want to build my routine of shooting without thinking.

For starters here is the link:

https://www.photonstophotos.net/Charts/RN_ADU.htm

This is the chart you should be looking at:

https://www.photonstophotos.net/Charts/RN_e.htm#Canon%20EOS%20M6%20Mark%20II_14

On the one you linked, each data point is scaled differently. That means it is multiplied by a different number, and it doesn't correspond directly to any physical unit.

On the chart I linked, the data points all correspond to an physical unit, namely the number of electrons, which in turn corresponds to the number of photons. For example, ISO 100 corresponds to 6 electrons, i.e., 6 photons. That tells you directly what level of subject luminance you can see.

ISpherical wrote:

edited them in DxO to kind of match the exposure....

This is a nice demonstration, but let's keep it clear.

In the first series the exposures were all the same. You determined that before you snapped the shutter. You used DxO to brighten the images to match them for lightness.

If you look at the chart I linked, you should see the biggest difference between ISO 100 and 400, and that actually corresponds very neatly to what you see. The large amount of noise at ISO 100 is dominated by read noise, which is shown in the chart.

In the second series the exposures are 2 1/3 stops higher, so the noise is dominated by shot noise (i.e., photon statistical noise). The read noise is much less visible, and the noise is similar in all images.

The best ISO setting for any exposure is always the highest value that does not result in clipping of any highlights.  Lower values, however, may give results that are nearly as good, and they may give you some needed margin of error.  With the M6 II, ISO 400 is pretty good for all of these photos.  If you could measure accurately, and I'm not sure you can from these images, I think you would find that the ISO values greater than 400 give slightly lower noise levels.  (Although I didn't actually check for clipped highlights, I don't see any.)

ThrillaMozilla wrote:

Spherical wrote:

The best ISO setting for any exposure is always the highest value that does not result in clipping of any [IMPORTANT] highlights. Lower values, however, may give results that are nearly as good, and they may give you some needed margin of error.

+1 The details of this tech talk can certainly seem complicated, but this is it in a nutshell!

When working within the sensor’s ISO Invariant Range, just do the above.

When working below the sensor’s ISO Invariant Range, just do the above.  Easy!  

The only time you would have to take any of this sensor-speak into consideration is when your output requirements exceed your end results.  (Just be sure to maximize your end results by shooting RAW and processing with something like DxO   ).

Happy Snapping!

R2