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

Archive 20155Ds production samples posted (RAWs and JPEGs)

  
 
ronno
Offline
Details
p.2 #15Ds production samples posted (RAWs and JPEGs)
jcolwell wrote:
Well, at least we all agree on one thing. Trouble is, we haven't yet identified what that one thing is.


jcolwell, I think we can agree that there is more detail and seemingly more D.R. here in the 5Dr, but shadow noise could be better. (but yeah, I'm pretty sure someone will come argue with that statement in 3...2...1...)


Jun 06, 2015 at 10:38 AM
mttran
Offline
Details
p.2 #25Ds production samples posted (RAWs and JPEGs)
ronno wrote:
gdan, I am often shooting business owners, etc. who have things to do other than having their photo taken at *optimum light times*.


gdanmitchell wrote:
Just stand them on the shady (or sunny) side of the trees. :-)


ronno wrote:
You've still got blown light areas and muddy shadows, fine (barely) with the D750, not fine with many other cameras.
Anyway I am pretty sure that you are not arguing that more D.R. is useless, so I am going to move on from this debate.

-ron


I am with you on this one, Ron. I rather have less but healthy pixels than more with colors distortion and noisy pixels. The later often defeats the purpose of bigger print whenever our subject was in strong & uneven lights...I guessed, I am a little picky on my tools, so better DR should be a safety bet for me.

http://farm1.staticflickr.com/397/18537125251_d58875e1f1_h.jpg
Jun 06, 2015 at 05:45 PM
Ferrophot
Offline
Details
p.2 #35Ds production samples posted (RAWs and JPEGs)
Pretty well every DSLR camera that Canon has introduced after the 40D has had a group of naysayers who make up reasons why it will be no good and won't sell. As each camera hits the market and real photographers try it out we are all pleasantly surprised. Every camera is not for every photographer but all seem to find an enthusiastic following. The 5Ds is not for me but I think many will love it.
Jun 06, 2015 at 10:31 PM
FLSTCSAM
Offline
Details
p.2 #45Ds production samples posted (RAWs and JPEGs)
Just glancing through this thread and wonder if I am reading the EXIF data on the posted examples correctly.

I see horrible noise, but the EXIF says they were shot at ISO 100?

Sam
Jun 07, 2015 at 10:46 AM
gdanmitchell
Offline
Details
p.2 #55Ds production samples posted (RAWs and JPEGs)
FLSTCSAM wrote:
Just glancing through this thread and wonder if I am reading the EXIF data on the posted examples correctly.

I see horrible noise, but the EXIF says they were shot at ISO 100?

Sam


The 5Ds and 5Ds R sample images are actually very good. A while back I downloaded one of the sample images, resized to 30" x 45" and then to 60" x 90" and printed sections of the resulting files with very good results.

Since then I have taken some of the recently-available raw files and played with them, too. In one (a photograph of a bridge in London) there are boats in the foreground that are partially in shadow and which seem to have luminosity values of about 0. I used the exposure slider to push these by something like four stops and could produce visible luminosity and chroma noise — by turning black into something just north of neutral gray.*

I'm pretty certain that most (all?) of the files supposedly demonstrating awful noise from the 5Ds and 5Ds R have been pushed in similar ways, since that's the only way I can produce such images.

It would be helpful if people who post photographs designed to demonstrate the technical capabilities of cameras would clearly disclose all of the processes that led to the images. To do otherwise does not contribute to getting closer to truth, has the potential to mislead people who see the images and do not know wha they actually represent, and may even smack of attempts on the part of the poster to mislead in order to push some agenda. (Disclosure can also reveal when artifacts were not caused by the camera, but rather by incautious post-processing — as was the case in one recent post that showed high levels of noise in 100% magnification crops at ISO 100, but where it turned out that the poster had accidentally introduced these artifacts by means of an add and ill-advised sharpening procedure.)

In general (and with only very rare exceptions) be very skeptical of people who try to make differences between equipment look hyperbolically huge or people who propose that one option is near perfection and the other is nearly unusable or at least grossly bad.

Take care,

Dan,

who is not intending to imply that the 5Ds or 5Ds R will be perfect or even God's Own Perfect Camera, and who has not yet been able to work with files he created himself using his own copy of one of the cameras.

*I'm busy printing and mounting today, so I can't take time to find and reprocess that file, but I may try later on when I have time — I'll share the results and explain what you are being. Meanwhile, you can download sample files yourself from Canon and from a variety of third party sources.

Edited on Jun 07, 2015 at 11:58 AM · View previous versions
Jun 07, 2015 at 11:40 AM
atodzia
Offline
Details
p.2 #65Ds production samples posted (RAWs and JPEGs)
I agree with Dan. Download some Raw images and then take them through your normal work flow. I used Capture one and then Photoshop. The results were very good. The only major processing slow down was using Topaz Denoise which seemed to take about twice as long as I am used to with 5D3 and 7D2 files.
Jun 07, 2015 at 11:57 AM
Shutterbug2006
Offline
Details
p.2 #75Ds production samples posted (RAWs and JPEGs)
About the only thing wrong with the 5DS that I can see, is the five frames per second limitation of camera.

I'd like more megapixels, but I think I might wait until the 2nd generation shows up.

Jun 07, 2015 at 01:45 PM
mttran
Offline
Details
p.2 #85Ds production samples posted (RAWs and JPEGs)
gdanmitchell wrote:
FLSTCSAM wroteJust glancing through this thread and wonder if I am reading the EXIF data on the posted examples correctly.

I see horrible noise, but the EXIF says they were shot at ISO 100?

Sam
The 5Ds and 5Ds R sample images are actually very good. A while back I downloaded one of the sample images, resized to 30" x 45" and then to 60" x 90" and printed sections of the resulting files with very good results.

Since then I have taken some of the recently-available raw files and played with them, too. In one (a photograph of a bridge in London) there are boats in the foreground that are partially in shadow and which seem to have luminosity values of about 0. I used the exposure slider to push these by something like four
...Show more →

Gdan,

You do need a mouth wash. Last two posts of mine is nothing more than a reverse engineering to illustrate the (IQ.vs.DR) that has been shown by dpreview: http://tinyurl.com/qh7urzy

There is no such agenda like your trolling post but my simple process to show the canon IQ is somewhat depending on the DR of the scenes. Matter of fact you do know that when you said “Just stand them on the shady (or sunny) side of the trees.” on your reply to Ronno but too bad you've always solved the problem halfway since the canon low iso banding ages.

Here is another two samples by the same 5Dsr at iso 100. Both raw were pushed hard 2 stops to lift up the noise. One look much better IQ than the other because one has softer DR lights than the other.

http://farm1.staticflickr.com/378/18385707039_907ffe59fa_h.jpg
http://farm9.staticflickr.com/8846/18384228560_61c4d916ac_h.jpg

Got it now , Dan? Please, think twice before saying naughty things. You have been used the same tactic to comment people post like this regarding the same subject whenever you have disagreed upon to and that need to be stopped.

Edited on Jun 08, 2015 at 12:35 PM · View previous versions
Jun 07, 2015 at 02:51 PM
Cliff L.
Online
Details
p.2 #95Ds production samples posted (RAWs and JPEGs)
ronno wrote:
And about "too many pixels": in all my 20 or so years shooting professionally, I'm pretty sure that I have never had a client ask for more detail in a finished image.


I realize you're just starting out, so you haven't figured out that the number of pixels on the sensor has little bearing on the amount of noise it produces - especially when you down-sample the images to small file sizes like the D750 has. Of course, exposing the image properly makes a huge difference, too, but we can save that for another thread.
Jun 07, 2015 at 03:01 PM
gdanmitchell
Offline
Details
p.2 #105Ds production samples posted (RAWs and JPEGs)
(For anyone interested in discussing the sample images, I just reposted them in their own thread.)

I just took a break and had time to play with this file again.

The file is a raw format test image from http://www.photographyblog.com/previews/canon_eos_5ds_photos/. This was photographed with the 5Ds at ISO 100, f/8, 1/400 second. It isn't clear what lens was used, but it appears that it could have been either the 50mm f/1.8 STM lens or the 24-70mm f/2.8 II.

Opened the file in ACR. Made no adjustments to curves, color, etc. Let ACR automatically correct for CA. Default ACR sharpening used at 15 with masking at 50.

Brought file into Photoshop as a smart object to allow non-destructive re-editing in ACR. Confirmed that areas along the waterline of the boats in shadow have luminosity values of 0 by checking the Lab color representation and watching the L value which hits 0 in several spots. The general area is shown by the rectangle in the following image:

http://www.gdanmitchell.com/images/5DsNoise/5DsPhoto.jpg

Here are five versions of a 600 x 400 pixel crop including this area of the scene. All are 100% magnification crops. They are as follows:

1. original exposure.

http://www.gdanmitchell.com/images/5DsNoise/canoneos5ds0.jpg

2. "exposure" slider raised by 1.00.

http://www.gdanmitchell.com/images/5DsNoise/canoneos5ds1.jpg

3. "exposure" slider raised by 2.00.

http://www.gdanmitchell.com/images/5DsNoise/canoneos5ds2.jpg

4. "exposure" slider raised by 3.00.

http://www.gdanmitchell.com/images/5DsNoise/canoneos5ds3.jpg

5. "exposure" slider raised by 4.00.

http://www.gdanmitchell.com/images/5DsNoise/canoneos5ds4.jpg

6. "exposure" slider raised by 5.00.

http://www.gdanmitchell.com/images/5DsNoise/canoneos5ds5.jpg

Each successive image was made by reopening the smart object in ACR and readjusting the "exposure" slider on full value higher, which is essentially the same thing in ACR as adjusting the same slider in LR.

Each was converted to 8-bit jpg and the sRGB color space. No sharpening or additional processing was applied in ACR or in Photoshop.

For fun, here is the whole original image with exposure at +5.00!

http://www.gdanmitchell.com/images/5DsNoise/canoneos5dsFinal.jpg

I have no evidence that the 5Ds equals or exceeds the low noise or dynamic range capabilities of the Sony sensors, and no one has made such a claim — to the contrary, it seems like the new Canon sensor will likely have less dynamic range than the current Sony sensors. That said, everything I've seen so far suggests that the 5Ds sensor will work very well. I hope this set of images may help quantify the meaning of this and the actual ability of the 5Ds sensor.

Enjoy!

Dan

Edited on Jun 08, 2015 at 01:24 PM · View previous versions
Jun 07, 2015 at 04:03 PM
mttran
Offline
Details
p.2 #115Ds production samples posted (RAWs and JPEGs)
Gdan, I am not saying that canon cam can't do it. I am saying canon file can be better and look more organic (under the wide DR condition like most exmor sensor does) after post than it is right now. Take a another peek of dpreview tool and try to understand how the canon/sony engine under the hood work: dpreview: http://tinyurl.com/qh7urzy

BTW, you need to remove the default noise reduction before you lift the shadows to see the real data from your last post. Like I mentioned on page 1 of this thread, this sample was taken in normal DR (even bright sun lit) not a tricky lights so it is not a good sample to support what you try to say in all cases.
http://farm9.staticflickr.com/8841/18578661205_0a7c29e16c_h.jpg

Don't forget canon noise signature always changing based on the angle of the sun and your camera parameters setting. This is not new and already illustrated by two samples of my last post. People raw samples are still there...see it for yourself
Jun 07, 2015 at 04:15 PM
darbo
Offline
Details
p.2 #125Ds production samples posted (RAWs and JPEGs)
5DS R is looking good to me. My buy/don't-buy meter's needle, which has been going back and forth, is now into the "buy" zone.

Of course we knew it wouldn't catch up to the Exmor performance, but it still impresses me as a strong upgrade from my 5D III's sensor, which somehow gives me excellent results with great frequency.
Jun 07, 2015 at 07:15 PM
jctriguy
Offline
Details
p.2 #135Ds production samples posted (RAWs and JPEGs)
mttran wrote:
ronno wrotegdan, I am often shooting business owners, etc. who have things to do other than having their photo taken at *optimum light times*.
gdanmitchell wroteJust stand them on the shady (or sunny) side of the trees. :-)
ronno wroteYou've still got blown light areas and muddy shadows, fine (barely) with the D750, not fine with many other cameras.
Anyway I am pretty sure that you are not arguing that more D.R. is useless, so I am going to move on from this debate.

-ron
I am with you on this one, Ron. I rather have less but healthy pixels than more with colors distortion and noisy pixels. The later often defeats the purpose of bigger print whenever our subject was in strong & uneven lights...I guessed, I am a little picky on my tools, so better DR should be a safety bet for me.

http://farm1.staticflickr.com/397/18537125251_d58875e1f1_h.jpg


What processing was done to this image? Please explain what we are looking at. Otherwise it's just useless garbage.
Jun 07, 2015 at 07:52 PM
mttran
Offline
Details
p.2 #145Ds production samples posted (RAWs and JPEGs)
jctriguy wrote:
mttran wrote
ronno wrotegdan, I am often shooting business owners, etc. who have things to do other than having their photo taken at *optimum light times*.
gdanmitchell wroteJust stand them on the shady (or sunny) side of the trees. :-)
ronno wroteYou've still got blown light areas and muddy shadows, fine (barely) with the D750, not fine with many other cameras.
Anyway I am pretty sure that you are not arguing that more D.R. is useless, so I am going to move on from this debate.

-ron
I am with you on this one, Ron. I rather have less but healthy pixels than more with colors distortion and noisy pixels. The later often defeats the purpose of bigger print whenever our subject was in strong & uneven lights...I guessed, I am a little picky on my tools, so better DR should be a safety bet for me.

http://farm1.staticflickr.com/397/18537125251_d58875e1f1_h.jpg
What processing was done to this image? Please explain what we are looking at. Otherwise it's just useless garbage.


JC, this is common process to handle most canon raw file in order to test the hidden noise. Reset all DPP parameters to zero then slide to group DR range (as shown by double arrow heads) for DPP to average the middle grey to display:
http://farm1.staticflickr.com/425/18405354378_fd70005264_h.jpg
another word, you don't alter the data but able to see them bit by bit in its whole DR range.
Jun 08, 2015 at 02:29 AM
alundeb
Offline
Details
p.2 #155Ds production samples posted (RAWs and JPEGs)
molson wrote:
ronno wroteAnd about "too many pixels": in all my 20 or so years shooting professionally, I'm pretty sure that I have never had a client ask for more detail in a finished image.
I realize you're just starting out, so you haven't figured out that the number of pixels on the sensor has little bearing on the amount of noise it produces - especially when you down-sample the images to small file sizes like the D750 has. Of course, exposing the image properly makes a huge difference, too, but we can save that for another thread.


Not only does the number of pixels have little bearing on the amount of noise when the images are presented at the same resolution, but regarding low ISO DR, the effect is the opposite. It is well established knowledge that high resolution helps the low ISO DR. It can easily be examplified with the Sony A7r vs A7s. The older high resolution A7r outperforms the newer "Noise king" low resolution A7s with a full stop for DR at ISO 100. It is also the same effect that makes the 5Ds appear slightly better than the 6D.

The theory behind is a bit complicated, but in short the dynamic range at low ISO for a given sensor technology is constant per pixel, and does not vary with pixel size. This has to do with where in the signal chain the DR limiting noise origins. When a higher resolution image is downsampled to a lower resolution, the noise level per pixel is reduced and the dynamic range per pixel increased. When compared to an image from another sensor with the same technology but lower pixel count, the higher pixel count sensor get the highest DR.
Jun 08, 2015 at 03:30 AM
ronno
Offline
Details
p.2 #165Ds production samples posted (RAWs and JPEGs)
Sounds like molson and alumdeb are saying that squeezing smaller pixels into a given sensor size does not lead to an increase in per-pixel noise. That is news to me. :-)

On a serious note, the philosophy behind waiting for the so called 5D4 is that no only will the pixels be larger (assuming it's lower rez), but also the tech will be newer, and thus presumably less noise and better D.R. will be present. Also 50 megapixel is excessive for my purposes. And I said before I've never been asked for such resolution. People shoot international advertising campaigns with 6 figure budgets on a daily basis with a 5D Mark II, 1Ds Mark II, etc.

And finally I'm sure that the next 5D will be less expensive - I believe this 5Ds is overpriced, though it seems to be quite capable.
Jun 08, 2015 at 09:44 AM
alundeb
Offline
Details
p.2 #175Ds production samples posted (RAWs and JPEGs)
ronno wrote:
Sounds like molson and alumdeb are saying that squeezing smaller pixels into a given sensor size does not lead to an increase in per-pixel noise. That is news to me. :-)



That is not what we are saying at all, and it is not funny that you try to ridicule detailed knowledge. Can you explain in detail how photon shot noise, read noise and conversion noise affects signal to noise ratio and dynamic range at low and high ISO? I can and have done so several times here.
Jun 08, 2015 at 09:51 AM
gdanmitchell
Offline
Details
p.2 #185Ds production samples posted (RAWs and JPEGs)
ronno wrote:
Sounds like molson and alumdeb are saying that squeezing smaller pixels into a given sensor size does not lead to an increase in per-pixel noise. That is news to me. :-)


For more than a decade, each release of a new, higher MP sensor has been accompanied by concerns that the smaller photo sites would result in increased problems with noise in the photographs made with the new sensors. If this had been true, the sum of all of those increases in noise would have left us with unusable cameras by now.

Of course, it hasn't worked out that way, and each generation of sensor has provided noise performance at least as good and often better than the previous.

A second potential noise-related advantage of higher MP sensors is that the size of the grain becomes smaller and smaller relative to print sizes, making it harder to see and less objectionable. The effect is somewhat similar to what we got back in the film era from using finer grained films.

In general (with some exceptions, for sure) the overall news about noise has been increasingly better over time.
Jun 08, 2015 at 09:55 AM
mttran
Offline
Details
p.2 #195Ds production samples posted (RAWs and JPEGs)
BTW, all the raw test samples I have used so far for this thread coming from the Op post which is from this link: http://www.photographyblog.com/previews/canon_eos_5ds_photos/
Jun 08, 2015 at 10:37 AM
ronno
Offline
Details
p.2 #205Ds production samples posted (RAWs and JPEGs)
alundeb wrote:
ronno wroteSounds like molson and alumdeb are saying that squeezing smaller pixels into a given sensor size does not lead to an increase in per-pixel noise. That is news to me. :-)
Can you explain in detail how photon shot noise, read noise and conversion noise affects signal to noise ratio and dynamic range at low and high ISO? I can and have done so several times here.


Am I interested in going on at length discussing pseudo science instead of looking at the pictures? No.
I have work to do, including planning a shoot at this place tomorrow:

mineral pools

I'll check back in here when people actually *own* the camera. (rather than getting red in the face about something none of us has actually tried before.)
P.S. the above shot is from my iPhone, so no need to go on about D.R. or whatnot
Jun 08, 2015 at 11:33 AM
FM ForumsCanon ForumJoin Upload & Sell
  
 

You are not logged inLog inRegister