May soon have PDR for 6d + Dual-ISO (photographic dynamic range)

Started by l_d_allan, August 04, 2015, 02:25:05 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

l_d_allan

I've always been curious just how much Dual-ISO improved DR on my Canon DSLR's, especially my 6D.

Recently, I submitted a series of test RAW's to Bill Claff at www.PhotonsToPhotos.net. He has developed software to calculate "Photographic Dynamic Range", which seems more or less related to DxoMark and SensorGen.info numbers:
http://www.photonstophotos.net/Charts/PDR.htm

I'm rather fuzzy on just how Dual-ISO works, and I'm not all that confident I provided Mr. Claff with appropriate test RAW's. The very preliminary results are less than expected ... maybe 1.5 stops at 100/800.

If someone has the interest and time to get involved in this, that would be appreciated.


l_d_allan

Here's a link to a DPR post with results:
http://www.dpreview.com/forums/post/56366592

> "The measured values aren't quite so high as those that Magic Lantern claimed."

a1ex

I've noticed this thread a bit late (from a recent discussion on dpreview).

Can you share the test files, and maybe ask Bill Claff to share some more intermediate results (not just the actual numbers) ?

A while ago I did a quick test (not an accurate measurement); on the SNR curve, you can see the DR on 5D3, at ISO 100/1600, is roughly about 14 stops. With the same method, after fitting a curve through the data points, I've got 10.9 stops on 5D3 at ISO 100 (link).

I should probably subtract 0.5 stops because of the half resolution in shadows, but other than that, I don't see any reason why the actual DR would be lower.

cmccullum

Can I just throw in that this guy may just be biased against canon in favor of Nikon...

Not really sure how relevant that may or may not be, but it's just an observation after looking at the webpage

Audionut

Simple maths really.

log2(white point/noise).  In this case, the white point is the value recorded at ISO 100 with the noise level being determined at ISO 400.

Lets use the values recorded from my 6D.
https://www.dropbox.com/sh/rrn7chasohg36hk/AAD7fngMRzvGNfrz6IzOgX3pa?dl=0



log2(77323/7.95) = 13.24.  The measured dynamic range of my 6D at ISO 100 is 11.41.  13.24 - 11.41 = 1.83.

In other words, the maximum dynamic range improvement on this 6D with dual ISO 100/400 is 1.83 EV (stops).

Now, photographic dynamic range is an entirely subjective measurement.  If it says some value below 1.83 EV, fine.

Sapporo

Quote from: Audionut on January 06, 2016, 06:17:12 AM
Simple maths really.

log2(white point/noise).  In this case, the white point is the value recorded at ISO 100 with the noise level being determined at ISO 400.

Lets use the values recorded from my 6D.
https://www.dropbox.com/sh/rrn7chasohg36hk/AAD7fngMRzvGNfrz6IzOgX3pa?dl=0



log2(77323/7.95) = 13.24.  The measured dynamic range of my 6D at ISO 100 is 11.41.  13.24 - 11.41 = 1.83.

In other words, the maximum dynamic range improvement on this 6D with dual ISO 100/400 is 1.83 EV (stops).

Now, photographic dynamic range is an entirely subjective measurement.  If it says some value below 1.83 EV, fine.
I tried your full well capacity at ISO 100 -noise level at ISO 6400
Log2(77323/1,76)=15,4. Dual ISO 100/6400 gives DR of 15,4 Ev? Seems I missed something.

Audionut

Quote from: Sapporo on December 14, 2017, 10:06:19 AM
Seems I missed something.

Only if you think you're going to get a full 15.4EV of good quality detail not drowned in noise, and you've forgotten that dual_iso has a con that revolves around the difference between the ISO's used.

Sapporo

Quote from: Audionut on December 15, 2017, 06:21:43 AM
Only if you think you're going to get a full 15.4EV of good quality detail not drowned in noise, and you've forgotten that dual_iso has a con that revolves around the difference between the ISO's used.
Thank you.

l_d_allan

Quote from: cmccullum on January 06, 2016, 04:00:35 AM
Can I just throw in that this guy may just be biased against canon in favor of Nikon...

Not really sure how relevant that may or may not be, but it's just an observation after looking at the webpage

Stale thread from this inactive ML forum user (sorry to have missed a1ex post on Jan 5, 2016), but is "this guy" ... l_d_allan (OP) or Bill Claff (from PhotonsToPhotos)?

Which webpage, por favor? DPR thread?
https://www.dpreview.com/forums/post/56366592

And here is an updated 6D "Special Edition" chart for ML Dual-ISO from Bill Claff ... I am unclear how to interpret ...

http://www.photonstophotos.net/Charts/PDR_MagicLantern.htm#Canon%20EOS%206D,Canon%20EOS%206D(ML)

This MTFMapper pre-newbie would find it interesting to see MtfMapper charts for the 6D at various settings for ISO and ML DUAL-ISO. My less-than-informed impression is that there is a trade-off of DR for resolution, but I'm fuzzy on how much that amounts to. IIRC, I asked about this is in a thread from about the time when Dual-ISO was originally implemented "way back when".