Home · Register · Join Upload & Sell

Moderated by: Fred Miranda
Username  

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

1
       2       3       end
  

Archive 2009 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)

skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #1 · p.1 #1 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


NOTES:

MAJOR EDIT: I forgot about the ISO rating differences between the cameras, this doesn't matter at all when talking about max DR but it does when you are comparing at all sorts of ISOs like this, the 5D2 and 40D probably need a bit of a penalty here compared to the 20D, the 50D too although not quite as much as the 5D2 and 40D, I don't think. Not sure about the 7D but it sounds like it might be rated similarly to the 50D. I suppose I could find the values from DxO and plug them in for each camera instead of the 100,200,400,etc. I used for each to make the plots fair.


Remember the 50D and 5D2 pattern band worse and the 7D the least so the realistic values might penalize the 50D and 5D2 a bit and boost the 7D a tad.

Remember SNR at 18% will give a pretty large boost to the 5D2 compared to the APS-C cameras. The APS-C should all be reasonably close with the 7D best then 50D then 40D then 20D just guessing by photon collection technology. But it is really the sensor size, at this point, that makes by far the largest difference.

Remember color depth was not tested here either.

The 40D tested was better than most copies. The 20D tested was worse than most copies. The 7D tested was pre-production using the mask area only.

COMMENTS:
The 7D appears to be the first DSLR where going above ISO1600 makes full sense
when shooting RAW (compared to pushing and pulling in post from ISO1600). The ISO3200 appears to truly be real in all senses of the word.

At ISO800 and up the 5D2 advantage might be realistically noticeable, for DR, compared to the others.

At ISO1600 and up the 7D advantage might be realistically noticeable, for DR, compared to the 20D and 40D (and probably compared to the 50D because of the pattern banding difference).

The 5D2 doesn't really lose any noticeable DR until you go above ISO800 but the APS-C start losing it above ISO400 (or one might say the 5D2 doesn't gain as much as it should at low ISO perhaps). That said, of course you are letting a lot less light hit at ISO800 than ISO100 so you get a lot more shot noise (but it's still not too bad on a 5D2 really).

In terms of 'pixel level' DR they are all essentially identical (aside from the 20D) until you go above ISO400 and then the 5D2 is a bit better (although considering the banding differences, realworld, you might need to get to ISO800 or more to seem better than the 7D).

Probably of most relevance is how the 7D will AF.

RESULTS:
OK, at left is 'pixel level' and at right is all normalized to 20D size:
http://skibum4.smugmug.com/photos/647484983_kzrMG-X3-1.jpg

and in plot form normalized, compressed Y-axis for easy viewing:
http://skibum4.smugmug.com/photos/647485653_nMqKN-X2-1.jpg
EDIT: here is the version where I tried to normalize for ISO (plotted the 7D three times depending whether it acts like the 50D,40D or 5D2):
http://skibum4.smugmug.com/photos/647621523_z8khz-X2-1.jpg

and same but full Y-axis to give a more realistic feel for what the differences actually imply (makes lots of it all seem more trivial as it is):
http://skibum4.smugmug.com/photos/647485313_NsQ2E-X2-1.jpg
EDIT: here is the version where I tried to normalize for ISO (plotted the 7D three times depending whether it acts like the 50D,40D or 5D2):
http://skibum4.smugmug.com/photos/647621906_7vqpQ-X2-1.jpg

EDIT: did NOT redo these below to account for ISO differences:
and compressed Y-axis plot at the 'pixel level':
http://skibum4.smugmug.com/photos/647484281_sBHeG-X2-1.jpg

and full Y-axis:
http://skibum4.smugmug.com/photos/647484531_y5kgz-X2-1.jpg

and full-details:
http://skibum4.smugmug.com/photos/647486147_XXPua-XL-1.jpg
http://skibum4.smugmug.com/photos/647486488_TKrwY-L.jpg


Edited on Sep 12, 2009 at 11:05 PM · View previous versions



Sep 12, 2009 at 08:56 PM
mfurman
Offline
• • • •
Upload & Sell: Off
p.1 #2 · p.1 #2 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Very interesting. Thank you. It would be nice to see 5D (mkI) her to see if 7D is getting close to it.


Sep 12, 2009 at 09:09 PM
thw2
Offline
• • • •
Upload & Sell: Off
p.1 #3 · p.1 #3 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Will Canon ever change their ADC (analog-digital converter) architecture so as to give their low ISO DR a boost?

How's the color sensitivity on the 7D?



Sep 12, 2009 at 09:18 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #4 · p.1 #4 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


mfurman wrote:
Very interesting. Thank you. It would be nice to see 5D (mkI) her to see if 7D is getting close to it.


I just re-did it for the normalized to 20D plots.

I haven't measured the gains myself and just tried to take how DxO rated the ISO sensitivities and use those for each camera.

There is no data at all on the 7D for this and I haven't bothered peaking at the IR RAW to try to make a good guess so I offered 3 plots: as if it behaves like the 40D,50D or 5D2 ratings (you can see it does make a bit of a difference between whether it works like the 5D2 or the 50D, anywhere from meaninglessly better than a 40D to modestly noticeably better).

With the new charts the 40D now falls a trace behind the 20D at the higher ISOs (but again all these minor differences are tough to see in the real world).

Although not show here, for kicks, I took the DxO 5D and 1D3 data and stuck it into my plot and it seems like the 1D3 is THE Canon DR king by a tiny margin at a normalized (NOT 'per pixel') level, but it's not a big deal only about a little less than 1/3 stop better than 5D2/7D (although with less pattern banding than the 5D2, the 1D3,1Ds3 and 7D should be the actual realistic winners).

Oddly the 1D3 seemed to pale above ISO400 for normalized DR, doing worse even than the 50D by a tiny bit....

The 5D seemed worse than the 5D2 and 7D at lowest ISO (about like 40D/50D) and similar to the optimistic trendline for the 7D at ISO400 and up in terms of DR.

Normalized to 20D scale (i.e. NOT per pixel), this is more like comparing prints from each camera printed at the same size and viewed from the same distance):
http://skibum4.smugmug.com/photos/647621523_z8khz-X2-1.jpg

Same but full y-axis:
http://skibum4.smugmug.com/photos/647621906_7vqpQ-X2-1.jpg



Sep 12, 2009 at 10:57 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #5 · p.1 #5 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


thw2 wrote:
Will Canon ever change their ADC (analog-digital converter) architecture so as to give their low ISO DR a boost?

How's the color sensitivity on the 7D?


Have to ask Canon about the first part (assuming DxO did handle the Nikon backpoint clipping properly and they really do have much better low ISO DR and that it really is reasonably possible).

No clue about the second. That and pattern banding are the two areas the 50D seemed to lag a bit. So far it appears as if the 7D has fixed the banding issue. I'm not sure anyone has the right files to test the color sensitivty yet.



Sep 12, 2009 at 11:02 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #6 · p.1 #6 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


It's late so I may be typing nonsense but:

Actually the main thing that jumps out at me is how the 5D2 does so well at ISO800 and above in therms of DR compared to the20D which has the same size photosites, it beats it quite handily per pixel so canon did make a pretty huge leap there, maybe 3/4 of a stop, with how they handle read noise/how many electrons they can stuff into a well of a given photosite size.

OTOH, it is interesting to note that at the print level the 7D, at best won't likely get 3/4 stops DR better than the 20D so small pixels may incur a slight high iso penalty (although perhaps giving a slight ISO boost). All the same the penalty for 8um vs 4um photosites only seems to be only 1/4-1/2 stop which may be a fair trade for so much reach and so much more landscape detail (and if they left it at 13MP as some suggested that might've only gotten you 1/8-1/3 or stops better DR which is meaningless).




Sep 13, 2009 at 12:00 AM
ExxWhy
Offline
• •
Upload & Sell: Off
p.1 #7 · p.1 #7 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Wish I had the slightest clue what this all means.


Sep 13, 2009 at 12:06 AM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #8 · p.1 #8 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


ExxWhy wrote:
Wish I had the slightest clue what this all means.



It means that the AF is the thing we should really be concentrating on with the 7D. Does the AF deliver or not.



Sep 13, 2009 at 12:27 AM
alundeb
Offline
• • • • •
Upload & Sell: Off
p.1 #9 · p.1 #9 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Nice work!

According to Clarkvision, the 5D2 full well is 0.2 -0.3 stops higher than the 20D. One thing I never understood about the Clark full well tables for 5D2: it is higher for ISO50 than ISO100. There seems to be a common wisdom that ISO50 is just a digital modification of ISO100. Do you have a chance to check the saturation levels at ISO50 and 100, or have you already done so?

Would you mind maintaining this kind of database with data sets from multiple copies as well?



Sep 13, 2009 at 03:08 AM
ExxWhy
Offline
• •
Upload & Sell: Off
p.1 #10 · p.1 #10 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


skibum5 wrote:
It means that the AF is the thing we should really be concentrating on with the 7D. Does the AF deliver or not.



That works for me. I didn't mean to sound disparaging, quite the contrary. Just that the technical stuff often causes a vacant look to appear on my face.



Sep 13, 2009 at 07:33 PM
RobertLynn
Offline
• • • • • •
Upload & Sell: Off
p.1 #11 · p.1 #11 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


I really wish I understood any of this.


Sep 13, 2009 at 07:50 PM
WillWeb
Offline
• •
Upload & Sell: Off
p.1 #12 · p.1 #12 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


The x-axis should be logarithmic.


Sep 13, 2009 at 09:35 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #13 · p.1 #13 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


WillWeb wrote:
The x-axis should be logarithmic.


yeah



Sep 13, 2009 at 10:28 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #14 · p.1 #14 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


alundeb wrote:
Nice work!

According to Clarkvision, the 5D2 full well is 0.2 -0.3 stops higher than the 20D. One thing I never understood about the Clark full well tables for 5D2: it is higher for ISO50 than ISO100. There seems to be a common wisdom that ISO50 is just a digital modification of ISO100. Do you have a chance to check the saturation levels at ISO50 and 100, or have you already done so?

Would you mind maintaining this kind of database with data sets from multiple copies as well?


If you look at the next to last image on the top post and check the second column (MSP) of data you will see that saturation point values, pretty much the same for ISO50 and ISO100. ISO50 isn't anything different than ISO100 in any way. All values are the same.




Sep 13, 2009 at 10:31 PM
lexvo
Offline
• • • •
Upload & Sell: Off
p.1 #15 · p.1 #15 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Does this mean that the 5D2 has the best DR, if you don't take the pattern noise into account?



Nov 07, 2009 at 01:38 PM
digitalbug30d
Offline
• • • •
Upload & Sell: Off
p.1 #16 · p.1 #16 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


major psychosis going on here...


Nov 07, 2009 at 01:44 PM
Rockies Photo
Offline
• • •
Upload & Sell: Off
p.1 #17 · p.1 #17 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


RobertLynn wrote:
I really wish I understood any of this.


+1. I do not comprehend these graphs. Some background, into understanding, what is being pointed out, would be appreciated.



Nov 07, 2009 at 02:24 PM
kakomu
Offline
• • • •
Upload & Sell: Off
p.1 #18 · p.1 #18 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Rockies Photo wrote:
+1. I do not comprehend these graphs. Some background, into understanding, what is being pointed out, would be appreciated.


It means that your photos are doomed if you don't use the best camera around!



...or not.



Nov 07, 2009 at 02:40 PM
Gochugogi
Offline
• • • • • •
Upload & Sell: Off
p.1 #19 · p.1 #19 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Rockies Photo wrote:
+1. I do not comprehend these graphs. Some background, into understanding, what is being pointed out, would be appreciated.


If your pictures are good, that's all that matters.



Nov 07, 2009 at 02:50 PM
RobertLynn
Offline
• • • • • •
Upload & Sell: Off
p.1 #20 · p.1 #20 · 7D,20D,40D,50D,5D2 DR at all main ISOs: (EDITED)


Gochugogi wrote:
If your pictures are good, that's all that matters.

Well...depends on who you talk to, haha.



Nov 07, 2009 at 03:00 PM
1
       2       3       end




FM Forums | Canon Forum | Join Upload & Sell

1
       2       3       end
 

You are not logged in. Login or Register

Username       Or Reset password



This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.