Reading mode: Light Dark

5D3 Photographic Dynamic Range etc.

Go to last post
from the links i posted before:

5d iso 50 7.3ev
5d iso 100 8.2ev

5d2 iso 50 7.7
5d2 iso 100 8.4

1ds3 iso 50 7.7
1ds3 iso 100 8.6

1d4 iso 50 7.8
1d4 iso 100 8.6

So Dpreview's DR numbers are wrong?
 
Do you mean photo DR or what? Straight out from the camera?

The useful dynamic range in the Raw is much greater in Nikon D7000, D800 due to less read noise than the Canon cameras , which has higher read noise, pattern noise and banding. (5dmk3 appear to have better values)

In a contrast-rich motif in such a street and half the street is in sunlit and the other half is in shadow,you can with a little knowledge to pick out much more information from the Nikon file and from the shadows than you can from Canon.

One of mine 5dmk2 compared to Nikon D7000
https://picasaweb.google.com/106266083120070292876/DR5dmk2VsD7000

Video

http://www.youtube.com/watch?annotation_id=annotation_737987&hd=1&feature=iv&src_vid=Dp2OeIcB1ak&v=Dwp7jfcsRSo
bclaff said:
This is my first post in any dpreview Canon forum.

I think the information will be of interest to certain people but certainly not to everyone.
I consider my testing methodology to be objective.
So please don't "shoot the messenger" if you don't like the results.

That said, I have far more experience testing Nikon cameras than Canon so errors with the Canon measurements are slightly more likely than for Nikon.

The testing methodologies are pretty standard and are outlined at my site, primarily in the Sensor Analysis Primer.( http://home.comcast.net/~NikonD70/GeneralTopics/Sensors_&_Raw/Sensor_Analysis_Primer/Sensor_Analysis_Primer.htm )

If you're not familiar with Photographic Dynamic Range (PDR) it is analogous to DxOMark Landscape DR but with different normalization criteria.

I have just processed the Imaging Resource CR2 files for the EOS 5D Mark III.
Most people find my Photographic Dynamic Range Chart the best overview:
( http://home.comcast.net/~NikonD70/Charts/PDR.htm )
A good cross section for comparing cameras is this bar chart:
( http://home.comcast.net/~NikonD70/Charts/PDR_65.htm )
More technical information would include read noise:
( http://home.comcast.net/~NikonD70/Charts/RN_ADU.htm )
And a restatement of PDR that shows how shadow DR improves with ISO:
( http://home.comcast.net/~NikonD70/Charts/PDR_Shadow.htm )

Click on the model name in the chart legend to select and deselect the cameras to display.

You can see that the 5D Mark III compares favorably with the top Nikon models in PDR.

I will be adding 5D Mark II data that I already have shortly. The Mark III is similar to the Mark II at low ISO and shows slight improvement at high ISO.

Finally, if you would like to contribute Canon data for analysis then please contact me by email.

Regards
--
Bill (visit me at http://home.comcast.net/~NikonD70/ )
Click to expand...
 
Personally I think that the 'noise floor' you have chosen (SNR 20:1) is pretty high, and that the output size you have chosen (something like 1mp) is pretty low. That's OK, of course, since you're doing all the work and therefore can define your PDR just like you want, but how about making it possible for the users/photographers to choose the noise floor and output size themselves?
bclaff said:
This is my first post in any dpreview Canon forum.

I think the information will be of interest to certain people but certainly not to everyone.
I consider my testing methodology to be objective.
So please don't "shoot the messenger" if you don't like the results.

That said, I have far more experience testing Nikon cameras than Canon so errors with the Canon measurements are slightly more likely than for Nikon.

The testing methodologies are pretty standard and are outlined at my site, primarily in the Sensor Analysis Primer.( http://home.comcast.net/~NikonD70/GeneralTopics/Sensors_&_Raw/Sensor_Analysis_Primer/Sensor_Analysis_Primer.htm )

If you're not familiar with Photographic Dynamic Range (PDR) it is analogous to DxOMark Landscape DR but with different normalization criteria.

I have just processed the Imaging Resource CR2 files for the EOS 5D Mark III.
Most people find my Photographic Dynamic Range Chart the best overview:
( http://home.comcast.net/~NikonD70/Charts/PDR.htm )
A good cross section for comparing cameras is this bar chart:
( http://home.comcast.net/~NikonD70/Charts/PDR_65.htm )
More technical information would include read noise:
( http://home.comcast.net/~NikonD70/Charts/RN_ADU.htm )
And a restatement of PDR that shows how shadow DR improves with ISO:
( http://home.comcast.net/~NikonD70/Charts/PDR_Shadow.htm )

Click on the model name in the chart legend to select and deselect the cameras to display.

You can see that the 5D Mark III compares favorably with the top Nikon models in PDR.

I will be adding 5D Mark II data that I already have shortly. The Mark III is similar to the Mark II at low ISO and shows slight improvement at high ISO.

Finally, if you would like to contribute Canon data for analysis then please contact me by email.

Regards
--
Bill (visit me at http://home.comcast.net/~NikonD70/ )
Click to expand...
 
I feel that displaying PDR as a function of ISO is shortchanging a lot of cameras (the 'ISOless' cameras to be specific). Take the 5DIII and D800 at ISO 1600, the 5DIII has a higher PDR in your graph but if we were to keep the exposure the same but change the ISO from 1600 to 400 on the D800, we gain two stops of highlight DR with only a little bit of increase in the noise in the shadows.

There is no neat way of solving this. One extreme way would be to say that PDR is a constant value, ie, the base ISO value and that one can trade PDR against shadow noise. Or, instead of plotting PDR over the ISO settings value, one could plot it over EV100 values that map into a specific noise value. Or over the noise value a specific EV100 value would map into. Both of these two approaches would mean as many graphs as ISO settings for a single given camera.
 
Your comparison is impressive!

Canon shadow recovery is quite bad indeed even at base ISO.

My Leica M9 is absolutely perfect in comparison to my 5DII. Don't know if the new mkIII will be much better... sadly it seems a Canon footprint.
:(

--
You can't depend on your eyes when your imagination is out of focus.

Marc Twain
 
Steen,
Steen Bay said:
Personally I think that the 'noise floor' you have chosen (SNR 20:1) is pretty high, and that the output size you have chosen (something like 1mp) is pretty low.
Click to expand...
The SNR was choosen because ISO standard for acceptable is SNR 10 and for excellent is SNR 40 so SNR 20 seemed a like it might be reasonable for "good".

The criteria affects the PDR reported but I've found that the differences between cameras remain pretty much the same regardless.
So as a tool to compare cameras I think it's a fine choice.

As you point out no one criteria fits everyone.

But allowing the user to adjust the criteria is not feasible. I would require collecting much more data per camera.

Regards
--
Bill (visit me at http://home.comcast.net/~NikonD70/ )
 
These are measured values, they don't shortchange any camera.

An "isoless" camera simply doesn't have shadow improvement when ISO is raised so by raising ISO don't get you much.
"isoless" cameras have PDRs that are straight lines with a slope of -1.

Even if you gain nothing in the shadows by raising ISO, you still lose one stop of highlights.
This chart is a good way to quantify the effect:
( http://home.comcast.net/~NikonD70/Charts/PDR_Shadow.htm )
For example, try comparing the 5D Mark III to the D7000.

BTW, the dip at ISO 6400 for the 5D Mark II on that chart is a slight measurement error.

Regards
--
Bill (visit me at http://home.comcast.net/~NikonD70/ )
 
bclaff said:
These are measured values, they don't shortchange any camera.
Click to expand...
They shortchange a camera if somebody compares, eg, the D800 with the 5DIII and concludes that the 5DIII has a higher DR in low light situations . I know, people should not do this because your graphs shows the DR as a function of the ISO setting and not as a function of the available light .
bclaff said:
An "isoless" camera simply doesn't have shadow improvement when ISO is raised so by raising ISO don't get you much.
"isoless" cameras have PDRs that are straight lines with a slope of -1.

Even if you gain nothing in the shadows by raising ISO, you still lose one stop of highlights.
This chart is a good way to quantify the effect:
( http://home.comcast.net/~NikonD70/Charts/PDR_Shadow.htm )
For example, try comparing the 5D Mark III to the D7000.
Click to expand...
I fully understand that. I am just suggesting that it might be instructive to see the DR as a function of noise in, eg, 18% grey for given exposure. Or noise in 18% grey (or any other output brightness) for a fixed DR.
 
DPR uses jpegs and in-camera settings, and some visual grayscales, quite poor quality stuff.
bclaff said:
Marcus,

I don't know how the dpreview numbers are measured but I'm sure they use a different criteria.
I would guess that those are highlight numbers and not total dynamic range.

Regards
--
Bill (visit me at http://home.comcast.net/~NikonD70/ )
Click to expand...
--
Renato.
http://www.flickr.com/photos/rhlpedrosa/
OnExposure member
http://www.onexposure.net/

Good shooting and good luck
(after Ed Murrow)
 
Bill, Thanks a lot for the great work. Its like an alternative standard to Dxo. :)

My understanding is the plots here speak about SNR/Noise/DR at various settings for cameras when normalized to some "size"... right?

But one integral component missing in the comparison is differences in resolution/detail when normalized to the same "size" as above, which can sometimes be traded for noise reduction... Is there a scientific way to incorporate that to define a figure of merit which makes more sense? (Please correct me if something missing in my understanding...) Are the 5DIII/D800 going to have an advantage over D3s/D4 due to their resolution even at high ISOs when all are normalized in size?

Similar to DR analaysis from RAWs is there also a way to extract the color depth/tonal range plots as well?
bclaff said:
This is my first post in any dpreview Canon forum.

I think the information will be of interest to certain people but certainly not to everyone.
I consider my testing methodology to be objective.
So please don't "shoot the messenger" if you don't like the results.

That said, I have far more experience testing Nikon cameras than Canon so errors with the Canon measurements are slightly more likely than for Nikon.

The testing methodologies are pretty standard and are outlined at my site, primarily in the Sensor Analysis Primer.( http://home.comcast.net/~NikonD70/GeneralTopics/Sensors_&_Raw/Sensor_Analysis_Primer/Sensor_Analysis_Primer.htm )

If you're not familiar with Photographic Dynamic Range (PDR) it is analogous to DxOMark Landscape DR but with different normalization criteria.

I have just processed the Imaging Resource CR2 files for the EOS 5D Mark III.
Most people find my Photographic Dynamic Range Chart the best overview:
( http://home.comcast.net/~NikonD70/Charts/PDR.htm )
A good cross section for comparing cameras is this bar chart:
( http://home.comcast.net/~NikonD70/Charts/PDR_65.htm )
More technical information would include read noise:
( http://home.comcast.net/~NikonD70/Charts/RN_ADU.htm )
And a restatement of PDR that shows how shadow DR improves with ISO:
( http://home.comcast.net/~NikonD70/Charts/PDR_Shadow.htm )

Click on the model name in the chart legend to select and deselect the cameras to display.

You can see that the 5D Mark III compares favorably with the top Nikon models in PDR.

I will be adding 5D Mark II data that I already have shortly. The Mark III is similar to the Mark II at low ISO and shows slight improvement at high ISO.

Finally, if you would like to contribute Canon data for analysis then please contact me by email.

Regards
--
Bill (visit me at http://home.comcast.net/~NikonD70/ )
Click to expand...
 
marcus2137 said:
from the links i posted before:

5d iso 50 7.3ev
5d iso 100 8.2ev

5d2 iso 50 7.7
5d2 iso 100 8.4

1ds3 iso 50 7.7
1ds3 iso 100 8.6

1d4 iso 50 7.8
1d4 iso 100 8.6

So Dpreview's DR numbers are wrong?
Click to expand...
DPR gives numbers based off of jpgs. If there is additional information in the shadows with the ISO50, it's quite probable that the jpg files would not be able to properly represent it as they are just 8bit (256 steps).

What I find interesting on my nikon cameras, is that even though the "base" ISO is something like 200, I find that there is still color noise at that setting, perhaps indicating that there is some amplification going on. Going below base to 100, the color noise goes away, at the expense of highlight range. I still haven't seen if the tradeoff of shadow detail versus loss of highlight range gives a net gain or loss. Not sure if it's at all similar with canon sensors.
--
Cloverdale, B.C., Canada
Currently shooting: Nikon D3S, D700
http://www.joesiv.com
 
rhlpetrus said:
Iliah suspects some NR trick is going on with 5D3's RAWs in high ISO. Given base ISO DR seems about same, not totally unfounded.
Click to expand...
But with Canon cameras the low ISO DR and high ISO DR or noise are unrelated. If the pixel noise was reduced you wouldn't see that effect at low ISO because there is not enough gain at base ISO for pixel noise of even the 5D2 to exceed later stage noise.
 
You must log in or register to reply here.

Similar threads

3Percent
Canon EOS 5D Mark III: Quiet and Excellent at High ISO
Replies
11
Views
25K
redhatcode29
R
Ephemeris
Canon EOS R5 Mark II Lab Test – Rolling Shutter, Dynamic Range and Exposure Latitude
Replies
14
Views
57K
Ephemeris
Ephemeris
H
Panasonic did something special with the Z6 III sensor
4 5 6
Replies
118
Views
86K
Leonard Shepherd
L
B
Canon EOS R5 Mark II Sensor Measurements at PhotonsToPhotos
3 4 5
Replies
81
Views
31K
SvO2
S
G
EOS R6 Mark II in camera settings to look like 5D Mark III?
Replies
19
Views
9K
Zoifio
Zoifio

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top