fredmiranda.com Login
This topic is in the 2015 archive. You can read it and like its posts here. It is read only.

Archive 2015Canon 5DS dynamic range: zero improvement (RAWs analyzed)

  
 
mttran
Offline
Details
p.19 #1Canon 5DS dynamic range: zero improvement (RAWs analyzed)
jctriguy wrote:
His posts on the DR/Canon topic tend to be written in a way to make it seem funny or clever, or some other unknown goal. Makes things nearly impossible to decipher. Retro throws in a like just to rally the troops on his agenda, but I doubt he knows what's written either.


ggreene wrote:
Both are very funny when they try to convince people that their opinion is fact. MTTRAN is especially good at making me laugh as he tries to convince us that the A6000 is a 1DX competitor. Retro is well, just retro. Still believes that Canon fired the CEO over low ISO DR. .


This special group seems so arrogant and nasty than I have ever thought.

Edited on Mar 29, 2015 at 01:46 AM · View previous versions
Mar 28, 2015 at 11:37 PM
mttran
Offline
Details
p.19 #2Canon 5DS dynamic range: zero improvement (RAWs analyzed)
RustyBug wrote:
mttran wroteObviously, sony sensor shall collect wider range data than canon sensor
Honest

How does a sony sensor collect a wider range of data?

I understand that it introduces less noise in the ISO 100-400 range, but how does it collect a wider range of data?

Are we not still on a 0-255 (really 0-1) scale.

BTW, last I checked (A7R) 14.07 - (6D) 12.11 = 1.96 ... not 3
(DXO results. If others have other data, presentation is welcome.)


Are we talking the same noise floor in both canon and sony platforms, Rus. What do we expect the differences from what we have now in 5Dsr while comparing 7D2 to other canon full frame data collection. I don't see anything better than 5D2 minus banding according to the below DXO reference and what canon has already said. It is clear to us that canon noise floor simply not as good as competition which has the clear win in usable (and wider) DR range. And we know, cleaner system always win in IQ at the end for both normal and bracketing exposure cases.
http://farm9.staticflickr.com/8658/16497853312_8cf3f538da_b.jpg
Imo, 11 stops DR and 11 usable stops DR are two different things in real world, it is relative to our IQ reference. Have we seen extremely differences in IQ between these FF bodies 5D/5D2/5D3/6D/1Ds2/1Ds3/1DX
Mar 28, 2015 at 11:40 PM
RustyBug
Offline
Details
p.19 #3Canon 5DS dynamic range: zero improvement (RAWs analyzed)
chez wrote:
Rusty, don't know if the post was directed to me...


Not directed @ anyone ... just a little "reflection" @ how we (self & people collectively in general) can see things sometimes. Just a moment of levity for a possible grin.

Your point @ no perfect camera (spouse, car or ) is one that I've advocated many times ... +1 @ horses for courses. Triple Crown winners are very rare to come by. Camera's are no different in that regard. Finding one that fit's the bill for all possible scenarios can be a "fool's game" in certain regard. Finding one that fits the vast majority of a person's prioritized needs / desires ... that's a different matter ... and one that is highly subjective, based on individual prioritization.

Nope, we're good ... no cheap shot from me on that one.


Mar 29, 2015 at 02:26 AM
RustyBug
Offline
Details
p.19 #4Canon 5DS dynamic range: zero improvement (RAWs analyzed)
mttran wrote:
Have we seen extremely differences in IQ between these FF bodies 5D/5D2/5D3/6D/1Ds2/1Ds3/1DX


Interesting graph, thanks.

I think the most telling tale is in the far right orange column. I've not previously seen that metric, but it shows a clear progression to me. I'd be curious to see how that compares to the non-Canon counterparts (not defensive, just curious @ relative).

Possibly, could you expand on the categories a bit. My read is that the last orange column is the collective output, but I may be not understanding it correctly ... so clarification would be welcome. That, and definition of Unity ISO.

My take away from this graph is that the 6D plays second fiddle to the 1DX (5Ds / 5DsR tbd). It also seems that the last orange column @ ISOless ISO is derived from the combination of full well capacity & noise characteristics ... which takes me back to my curiosity @ the relationships of non-Canon full well capacity & noise relationships.



Edited on Mar 29, 2015 at 08:37 AM · View previous versions
Mar 29, 2015 at 02:39 AM
Mescalamba
Offline
Details
p.19 #5Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Jeff Fenske wrote:
Mescalamba wrote
ggreene wrote
Mescalamba wroteTho as I said before, Im curious if Magic Lantern team and their dual-ISO will work with this. Should give some nice gains with bit of NR.
Canon has never told them that the 5D series is off limits so I'll bet you will see them work on it. I just wish that Canon would let them have a go at the 1DX. Would love to see what they could come up with.
Given noise floor on 1D-X I would say that pretty good DR as result.
Mescalamba,

Are you optimistic about the color potential of the 5Ds? I've *really* enjoyed your posts about color, and it seems that the 5Ds might finally have a stronger color filter array (CFA), defying the recent trend.


Well, it would be nice, wouldnt it? I hope it will have at least this right. We wont know till we see some regular samples from production models (and perhaps DxO mark results). Noise will be problem, DR too. But its 50 mpix, so with bit of SW magic, it might produce decent results (just not in full 50 mpix, but quite a bit less). For now, its just a lot of guessing..

And thanks.
Mar 29, 2015 at 07:52 AM
cgarcia
Offline
Details
p.19 #6Canon 5DS dynamic range: zero improvement (RAWs analyzed)
ben egbert wrote:
I also plan to use the 5Ds with the 11-24 in a way that will replace a TSE lens. IE shoot level and crop. This will solve the focus issues I have with manual focus.


I think that the 11-24mm may not always be a suitable replacement for the 17mm. If I have done well the maths in the table bellow, a 11mm lens can only emulate a 17 TSE shifted up by as much as 6mm, which simply is not enough for many shots. And at this point, due to cropping, it has lost a third of the sensor resolution, or up to 56% if the crop intends to keep a 3:2 ratio instead of the 2.25:1 cinema aspect ratio. In the worst case, 21.5 MP out of a 50 MP sensor could still be enough resolution for many purposes, but not only pixels are lost here: it is also sensor surface and light at the full image level (using 44% of a full frame sensor equals to shooting APS-C in this regard):

17 TSE vertical shift "equivalent" focal lengths (between brackets, such a equivalent lens unshifted and cropped):

shift 0 mm -> 17.00mm vertical FOV (emulating shift: 0% sensor surface lost at 3:2 ratio or 0% at 1.50:1 ratio)
shift 1 mm -> 15.69mm vertical FOV (emulating shift: 15% sensor surface lost at 3:2 ratio or 8% at 1.62:1 ratio)
shift 2 mm -> 14.57mm vertical FOV (emulating shift: 27% sensor surface lost at 3:2 ratio or 14% at 1.75:1 ratio)
shift 3 mm -> 13.60mm vertical FOV (emulating shift: 36% sensor surface lost at 3:2 ratio or 20% at 1.87:1 ratio)
shift 4 mm -> 12.75mm vertical FOV (emulating shift: 44% sensor surface lost at 3:2 ratio or 25% at 2.00:1 ratio)
shift 5 mm -> 12.00mm vertical FOV (emulating shift: 50% sensor surface lost at 3:2 ratio or 29% at 2.12:1 ratio)
shift 6 mm -> 11.33mm vertical FOV (emulating shift: 56% sensor surface lost at 3:2 ratio or 33% at 2.25:1 ratio)
shift 7 mm -> 10.74mm vertical FOV (emulating shift: 60% sensor surface lost at 3:2 ratio or 37% at 2.38:1 ratio)
shift 8 mm -> 10.20mm vertical FOV (emulating shift: 64% sensor surface lost at 3:2 ratio or 40% at 2.50:1 ratio)
shift 9 mm -> 9.71mm vertical FOV (emulating shift: 67% sensor surface lost at 3:2 ratio or 43% at 2.62:1 ratio)
shift 10 mm -> 9.27mm vertical FOV (emulating shift: 70% sensor surface lost at 3:2 ratio or 45% at 2.75:1 ratio)
shift 11 mm -> 8.87mm vertical FOV (emulating shift: 73% sensor surface lost at 3:2 ratio or 48% at 2.88:1 ratio)
shift 12 mm -> 8.50mm vertical FOV (emulating shift: 75% sensor surface lost at 3:2 ratio or 50% at 3.00:1 ratio)
Mar 29, 2015 at 08:25 AM
ben egbert
Offline
Details
p.19 #7Canon 5DS dynamic range: zero improvement (RAWs analyzed)
cgarcia wrote:
ben egbert wroteI also plan to use the 5Ds with the 11-24 in a way that will replace a TSE lens. IE shoot level and crop. This will solve the focus issues I have with manual focus.
I think that the 11-24mm may not always be a suitable replacement for the 17mm. If I have done well the maths in the table bellow, a 11mm lens can only emulate a 17 TSE shifted up by as much as 6mm, which simply is not enough for many shots. And at this point, due to cropping, it has lost a third of the sensor resolution, or up to 56% if the crop intends to keep a 3:2 ratio instead of the 2.25:1 cinema aspect ratio. In the worst case, 21.5 MP out of a 50 MP sensor could still
...Show more →

Not selling the 17TSE. But I have places where it will not cover even shifted, usually on the width in these cases. I usually only use shift to allow a level camera. The whole idea is that with 50mp, I can afford to crop some off the bottom and my track record with focusing the TSE is not so hot. Also, the 17TSE fully shifted starts to lose some IQ in the extremes.

Mar 29, 2015 at 09:09 AM
mttran
Offline
Details
p.19 #8Canon 5DS dynamic range: zero improvement (RAWs analyzed)
RustyBug wrote:
... which takes me back to my curiosity @ the relationships of non-Canon full well capacity & noise relationships.



Here ya go, Rus
http://farm9.staticflickr.com/8684/16942374076_1b77083440_o.jpg
Mar 29, 2015 at 11:14 AM
RustyBug
Offline
Details
p.19 #9Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Okay, thanks.

I could use some help with Unity ISO and ISOless ISO ... what do they mean / infer?
Mar 29, 2015 at 02:30 PM
Tom_W
Offline
Details
p.19 #10Canon 5DS dynamic range: zero improvement (RAWs analyzed)
I share RustyBug's questions. Interesting data, but I don't know what all of it means.
Mar 29, 2015 at 03:01 PM
Gunzorro
Offline
Details
p.19 #11Canon 5DS dynamic range: zero improvement (RAWs analyzed)
mmtran -- I can't understand this either. You can't just regurgitate someone else's study results without a full explanation of the terms and the conditions of the study. You have a habit of doing this to supposedly support your position in technical debates. If you can't or won't provide complete explanation, please don't use these examples.
Mar 29, 2015 at 03:39 PM
Mike Eckstein
Offline
Details
p.19 #12Canon 5DS dynamic range: zero improvement (RAWs analyzed)
I've never had anyone look at one of my images and say - that would really look nicer if it had more dynamic range.
Mar 29, 2015 at 04:23 PM
RustyBug
Offline
Details
p.19 #13Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Mike Eckstein wrote:
I've never had anyone look at one of my images and say - that would really look nicer if it had more dynamic range.


Reminds me of shooting Fujichrome ... just for kicks, took a look at the spec sheet.

Superb Push/Pull Processing

Minimum color balance varia-
tion over a range from –1/2 to
+1 stop to enable wider pho-
to opportunities. Also allows
fine adjustments of exposure
and density to be easily done
during processing


Superb @ - 1/2 EV to + 1 EV ... my how times have changed.
Mar 29, 2015 at 04:55 PM
AJSJones
Offline
Details
p.19 #14Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Gunzorro wrote:
mmtran -- I can't understand this either. You can't just regurgitate someone else's study results without a full explanation of the terms and the conditions of the study. You have a habit of doing this to supposedly support your position in technical debates. If you can't or won't provide complete explanation, please don't use these examples.


http://home.comcast.net/~NikonD70/index.htm is the source of the various plots
Mar 29, 2015 at 05:25 PM
chez
Offline
Details
p.19 #15Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Mike Eckstein wrote:
I've never had anyone look at one of my images and say - that would really look nicer if it had more dynamic range.


Maybe because the one's that were blown because of the lack of DR, either blown highlights or muddy shadows or banding in the shadows, get deleted before others can view them.

I've made plenty great photos with my Canon system...continue to do so, but I use techniques to "tame" the DR in the scene if need be. With a sensor that allows for greater dynamic range, I just need to rely less on these techniques and focus more on compositions and light.

Funny, not one person ever said my photos would look better with a camera capable of shooting 12fps.

Not one person every said any of my photos would look better with cleaner high ISO abilities.

In fact not one person ever said my photos would look better with "you fill in the blanks"

That's because photos that don't pass my strict criterias do not make it out of my computer. A camera with more DR just allows more of my photos to make it out of the culling stage which is a good thing.
Mar 29, 2015 at 06:31 PM
jcolwell
Offline
Details
p.19 #16Canon 5DS dynamic range: zero improvement (RAWs analyzed)
chez wrote:
...That's because photos that don't pass my strict criterias do not make it out of my computer. A camera with more DR just allows more of my photos to make it out of the culling stage which is a good thing.


Exactly the same reasoning explains why I use 12 fps and 10 fps 1D-series cameras for action photography. Having an advantage for a particular situation doesn't mean that you'll get the best possible results, it just increases your chances, a bit.

Mar 29, 2015 at 06:38 PM
Paul Mo
Offline
Details
p.19 #17Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Mike Eckstein wrote:
I've never had anyone look at one of my images and say - that would really look nicer if it had more dynamic range.


Viewers, if taking about Joe and Joanne Public, seldom know the sleight of hand behind anything they consume.

I am seeking the democracy of my eyes - I am in it for a fuller tonal range.
Mar 29, 2015 at 06:43 PM
EB-1
Offline
Details
p.19 #18Canon 5DS dynamic range: zero improvement (RAWs analyzed)
It's easy to photograph a mountain with a slow camera, but not so when handholding a 500/4 and tracking a subject at 10-12 FPS. Last time all the people that had the EVFs were cursing and missing the animal action shots or had to shoot wide and guess what the capture would be. Almost all of the good images were from the 1D IV, 1D X or D4, all with big teles.

If I had to prioritize the limitations that prevent me from capturing a good image they would be in order of importance:

1. AF/general camera speed
2. High ISO noise
3. Not enough reach
4. Lack of DR

In landscape images there are occasions when higher DR would be useful, but it's not the deal breaker very often and HDR offers a much higher increase in DR than a sensor bump does. I do also like dark shadows sometimes. In the old days we called that contrast - to delineate certain details.

EBH
Mar 29, 2015 at 06:55 PM
Monito
Offline
Details
p.19 #19Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Gunzorro wrote:
mmtran -- I can't understand this either. You can't just regurgitate someone else's study results without a full explanation of the terms and the conditions of the study. You have a habit of doing this to supposedly support your position in technical debates. If you can't or won't provide complete explanation, please don't use these examples.


+1

1. However, there are a lot of technical articles on the site but little or no explanation there, that I can find, of the codes for the charts that mttran posted. If someone can find it and post it or a link, I would be grateful.

2. mttran did not post the source, which is a definite failure: https://home.comcast.net/~NikonD70/ I think you have to have cookies enabled because the charts there are blank for me. Or maybe it is crappy microvirussoft SilverLight.

3. I'm not sure mttran understands what he is posting with the charts or why. He certainly doesn't try to explain them or interpret them. I'm not sure he could do so clearly even if he took the time he would need to do so.


Edited on Mar 29, 2015 at 07:14 PM · View previous versions
Mar 29, 2015 at 07:10 PM
chez
Offline
Details
p.19 #20Canon 5DS dynamic range: zero improvement (RAWs analyzed)
EB-1 wrote:
It's easy to photograph a mountain with a slow camera, but not so when handholding a 500/4 and tracking a subject at 10-12 FPS. Last time all the people that had the EVFs were cursing and missing the animal action shots or had to shoot wide and guess what the capture would be. Almost all of the good images were from the 1D IV, 1D X or D4, all with big teles.

If I had to prioritize the limitations that prevent me from capturing a good image they would be in order of importance:

1. AF/general camera speed
2. High ISO noise
3.
...Show more →

The beauty with expanded DR is that you can still have those deep dark shadows if you want...and if you don't, you can raise them...hopefully without ugly banding. More options is always better than less.

Using your list of importance, DR would the the highest on my list with the other 3 not even making my list. Maybe since we shoot such different images is leading to our views of the importance of DR. That is what really bothers me. If someone shoots images which don't challenge DR...why can't they just be happy with their current gear, but in doing so, also realize some people shoot different subjects which put priority on different aspects of the camera.

Many landscape photographers don't care about FPS, high ISO, AF tracking or in fact AF in general...you don't see them going into a bird in flight discussion putting down the need for better AF tracking.

EB, if you are happy with existing DR in your cameras, then you should be all set in that aspect. Can you not at least understand others just might not be as satisfied as you?
Mar 29, 2015 at 07:12 PM
FM ForumsCanon ForumJoin Upload & Sell
  
 

You are not logged inLog inRegister