February 01, 2015, 10:18:49 AM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Topics - Neutral

Pages: [1]
 I am wondering when image processing softwares (LR, DXO, C1 etc.)  will start using GPU for image processing .
 This could drastically increase performance and processing capabilities and could help to implement more complicated and more resource demanding algorithms. Especially using NVIDIA CUDA – when it is possible to utilize more than 1500 processors on latest NVIDA cards for processing instead of just 4 or 8 cores on main CPU.
One who first implement this could have great advantage over other competitors.

This question comes to my mind each time when there are news about new major S/W releases – e.g. now with the information that Lightroom 6 will be coming soon.

Earlier Adobe was telling about difficulties to implement parallel processing but this does not reflect current realities.

Simple search on WEB shows that there are patents existing for GPU image processing as well as number implementations and API Libraries to utilize CUDA technology for image and video processing including RAW files processing  and some implementations which provide amazing processing speed.

Here are some references:

1. GPU  Raw image processing patent  US 8098964 B2

2. http://www.ximea.com/de/technology-news/gpu

3. http://on-demand.gputechconf.com/siggraph/2013/presentation/SG3108-GPU-Programming-Video-Image-Processing.pdf

Lenses / Canon EF24-70mm f/2.8 M2 vs Sony Zeiss 24-70 on Sony a7R body
« on: April 03, 2014, 12:20:12 PM »
  I was always  interested to know what is the resolution power limit of Canon  EF24-70mm f/2.8 M2  combined with the high resolution body and  how this compares to  similar lense type from different vendors.
On Canon 1DX this is just perfect piece of glass.

  Few days back I received Matabone III adaptor which gave me ability to mount Canon lenses on Sony a7R body.
With that I was finally able to test Canon  EF24-70mm f/2.8 M2  on 36mpx   Sony a7R body.
As result there  is also comparison of image quality  between Canon  EF24-70mm f/2.8 M2    and  Sony Zeiss 24-70 f/4 OSS.

There are two 100% crops at 24mm focal lens and two 100% crops at 70mm  focal lens.
All shots  are done  hand held  using camera autofocus.
Focus speed of Canon  EF24-70mm f/2.8 M2 with Matabone III is very slow - few seconds of hunting and searching for focus peak.
Sony Zeiss 24-70 f/4 OSS  focuses almost instantly.

 RAW files  from  Sony a7R were process using Phase One  Capture One with default settings, no adjustements at all

Test results demonstrate that Canon  EF24-70mm f/2.8 M2 is much superior to Sony Zeiss 24-70 f/4 OSS in resolution and color reproduction and overall image quality .  Especially this is noticable at 70mm focal length .
From test pictures it is obvious that even on 36mpx   Sony a7R body  Canon  EF24-70mm f/2.8 M2 still does not reach it's limit in resolution power.  This combination is perfect one when focus speed is not important.
On the other end  Sony a7R with  Sony Zeiss 24-70 f/4 OSS  is a fairly good combination as general purpose  walk-around point and shoot camera.

I also tested Canon TSE 17  on a7R and image quality , color rendering and resulution is also of top possible quality. 
It seems that Canon TSE 17 the same as Canon  EF24-70mm f/2.8 M2 still does not reach it's resolution power limit on a7R
Canon TSE 17  on Sony  a7R  looks as  perfect  combination for Architecture and Landscape photograpy.

Sony is pushing limits with it's  new MF 50mpx CMOS  sensor - vey high IQ at ISO6400:

The IQ250's new Sony-built sensor offers options for medium-format shooters, including high-sensitivity performance up to ISO 6,400; exposures as short at 1/10,000 of a second and as long as one hour; and a dynamic range spanning 14 f-stops for a better ability to capture details in both shadows and highlights. It doesn't take advantage of CMOS sensors' superior ability to shoot video, but that's no surprise or even shortcoming given that Phase One sells machines designed solely to capturing the best still photos possible.

Found  interesting info on  Sony 7R image quality vs. 5Dm3, Nikon D800E and Pentax 645D with 100% comparison crops at different ISOs at imagine -resource:
 Very interesting to see comparison of Canon 5D M3 with Sony a7R  at ISO 3200  both  low contrast high details shorts and for high contrast ones.

EOS Bodies - For Stills / 1Dx ISO100 high DR shadows SNR improvement
« on: November 04, 2012, 12:28:45 PM »
One of the weak sides of the 1Dx is noise   in deep shadows at base ISO.
It seems that nobody is arguing about that any more.
 DXO 1Dx measurement results also show that very clearly on their measurement curves.
And I believe that nobody would be arguing that performance in this area is very critical for not just few people but for many people  - for landscape photography, evening/night  city shots and other sort of photography.
Though it is easy to argue that that Nikon D800 would be the better choice BUT why not to see how we could get better deep shadow SNR from 1Dx.
In one of my posts I mentioned that 1Dx has a very nice feature that could be used for that.
I will not name it now (will do at SOME next post) but here is some magic - example how it works (snapshots from LR screen - noise reduction is totally disabled  )
1.  First image shows 1dX  shot at IS100  with EV=(-3)   the Data Color noise calibration chart  which was  used  for Noise Ninja NR S/W and PS plug-in. Shot  with  manual Out of Focus  to blur target texture so it would not interfere with noise
2. Second one is the same with Exposure slider in LR set to +3  - looks the same as normal shot with EV=0
3. Third  image is 200% crop  of shot with EV=0 showing normal 1Dx ISO shadow noise
4. Fourth one is 200% crop of the EV=-3 shot with LR exposure set to +3 to raise shadows to normal level.
     Level of 1Dx noise at ISO100 at  -3 EV shadows is very clearly seen
 And next post is  little bit of magic  - the similar EV=(-3) shoot from 1Dx  recovered in LR - 200% crop  to see the difference  in SNR for random nose with the previous shot.

 For those who are interested are examples of 1DX low light performance in high DR environment.

  Also this is illustration why high DR performance is required for those who keeps repeating  " why on earth your need high DR and why you need to underexpose images"
Image RAW conversion is in LR4.
Screen snapshots for posting by Win7 Snipping tool - to be quick

1.  Image "AS IS" from 1Dx - evaluative metering - camera is trying to protect highlights so average weighted image exposure shifts to the left (to the shadows) - so most part of the image is underexposed
2. 1Dx Image is normalized as close to human vision - image DR compressed in LR using basic tone sliders
3. 1Dx  ISO1000 crop 100% - no NR
4.  1Dx ISO1000  crop 100%  - NR Lum=20
Please observe noise in shadow recovered areas and sky and what we could do with it in LR
Next post is the same for ISO100

EOS Bodies - For Stills / 1Dx simple DR stress test
« on: October 19, 2012, 06:04:27 AM »
There are a lot of heated discussions about Canon versus SoNikon sensor DR performace, sensor technology, about DXO not measuring sensors correctlly in favour of Nikon  etc.
  So here is a simple DR stress test results for 1Dx shadow recovery limits for those who are interested:
  I did these tests to see  myself what is 1Dx is capable of.
Attached are 4 pictures (snapshots from LR4 screen by Win7 snipping tool):
1. 1Dx shot at ISO100 with EV = (-3 )   - showing 100% crop area shown below
2. Corrected with +3 EV in LR4, no shadow correction, no noise reduction applied
3. The same as #2 above with shadows raised to max 100%
4. The same as #3 above but with NR applied in LR  with L=40, C=40.
Results are self explanatory -  no further comments from me on this subject.
  Despite for what we see here regarding 1Dx DR  performance here my experience with 1Dx is very positive - almost perfect tool for low light sport  and events shooting.
In daylight 1Dx images also look much cleaner and having more 3D look than 1DsM3 and 5DM2 that I used before (just my personal perception)  and allow more image sharpening to be applied safely.
Also seems that 1Dx has a bit weaker anti-aliasing filter  - with extreemly sharp 24-70 2.8L II I could see moire on  some areas of ISO resolution chart  where image elements (projected on sensor)  become comparable in size with pixel size on camera sensor.  I have not observed that with other lenses including very sharp 70-200 2.8L IS II

Pages: [1]