Approve the Cookies
This website uses cookies to improve your user experience. By using this site, you agree to our use of cookies and our Privacy Policy.
OK
Index  •   • New posts  •   • RTAT  •   • 'Best of'  •   • Gallery  •   • Gear  •   • Reviews
Guest
New posts  •   • RTAT  •   • 'Best of'  •   • Gallery  •   • Gear  •   • Reviews
Register to forums    Log in

 
FORUMS News & Rumors Camera Rumors and Predictions 
Thread started 21 Jul 2016 (Thursday) 08:39
Search threadPrev/next
sponsored links
(this ad will go away when you log in as a registered member)

5D Mark IV announced August 25th

 
this thread is locked
sploo
premature adulation
2,402 posts
Gallery: 1 photo
Likes: 460
Joined Nov 2011
Location: West Yorkshire, UK
     
Aug 18, 2016 10:46 |  #121

bacchanal wrote in post #18098923 (external link)
What is interesting to me is that the rumored 4K 8 bit 422 MJPEG comes in at about the same bitrate as 4K ProRes 422 (500 mbps vs ~489mbps). Both are intraframe codecs.

I'm not familiar with MJPEG, so I really don't know if it's similar to ProRes or DNxHD from an editing standpoint, but it sounds like it could be. Or possibly a clean 422 signal could be sent over HDMI to an external recorder where ProRes 422 or DNxHD could be recorded. That would be a welcome change for a lot of people I think (assuming the codec is an improvement over what is in the 5DIII).

It would be nice to see C-Log included as well, but since it's not available in the 1DxII I suspect it won't be in the 5DIV either.

Good question; I don't know (anything, frankly) about the internals of ProRes, so had a quick look at Wikipedia ;-)a

I guess it'll come down to efficiency. Wiki says it's DCT based, so that's the same as JPEG/MJPEG (JPEG/MPEG 2000 use a wavelet algorithm instead).

On paper then there might be little difference; though ProRes 422 is claimed to have a 10-bit sample depth, and I'm assuming MJPEG is only 8; so the former might give more detail and better results with heavy grading.


Camera, some lenses, too little time, too little talent

  
  LOG IN TO REPLY
sponsored links
(this ad will go away when you log in as a registered member)
Talley
Talley Whacker
Avatar
11,064 posts
Gallery: 45 photos
Likes: 2720
Joined Dec 2011
Location: Houston
     
Aug 18, 2016 12:11 |  #122

MJPEG is fine for any non cinema movie enthusiasts. It just sucks for the moms/dads type that want to use it for 4K but don't want massive files. You have to encode the video to H.264 to reduce the file size but then you are reducing the bitrate as well.

Canon use to nothing but MJPEG on their cameras and then went to the IPB/ALL-I type H.264 formats which is fine...

I don't see this but an extra step and a lot of memory usage.... ~12min per 64gb. You'll want a 512GB card 90 minutes of video coverage and you can get a Lexar 800x 512GB card for $290.

ONE THING about doing 4K MJPEG is you can then crop the video TO 1080P and have really good IQ still.

I for one... do not mind the MJPEG.


A7rIII | A7III | 12-24 F4 | 16-35 GM | 28-75 2.8 | 100-400 GM | 12mm 2.8 Fisheye | 35mm 2.8 | 85mm 1.8 | 35A | 85A | 200mm L F2 IS | MC-11
My Gear Archive

  
  LOG IN TO REPLY
sploo
premature adulation
2,402 posts
Gallery: 1 photo
Likes: 460
Joined Nov 2011
Location: West Yorkshire, UK
     
Aug 18, 2016 14:41 |  #123

Talley wrote in post #18099131 (external link)
MJPEG is fine for any non cinema movie enthusiasts. It just sucks for the moms/dads type that want to use it for 4K but don't want massive files. You have to encode the video to H.264 to reduce the file size but then you are reducing the bitrate as well.

Canon use to nothing but MJPEG on their cameras and then went to the IPB/ALL-I type H.264 formats which is fine...

I don't see this but an extra step and a lot of memory usage.... ~12min per 64gb. You'll want a 512GB card 90 minutes of video coverage and you can get a Lexar 800x 512GB card for $290.

ONE THING about doing 4K MJPEG is you can then crop the video TO 1080P and have really good IQ still.

I for one... do not mind the MJPEG.

As far as I understand, formats such as ProRes and MJPEG (and, I guess, ALL-I MPEG) are more intended for editing than content delivery. IPB MPEG will give much smaller file sizes for the same visual quality, but compromises editing as more data has been thrown away for the goal of a lower bitrate.

Given how ubiquitous ProRes seems to be in the editing world, I am surprised that more cameras don't support it, but perhaps there are issues with it being an Apple proprietary format (big companies especially seem to not like to use/endorse technology from other, possibly rival, big companies).


Camera, some lenses, too little time, too little talent

  
  LOG IN TO REPLY
frankchn
Senior Member
455 posts
Likes: 152
Joined Jun 2009
     
Aug 19, 2016 02:23 |  #124

Some more information on Dual Pixel RAWs from Digicame-Info: http://digicame-info.com …08/eos-5d-mark-iv-12.html (external link)

Innovative image processing support the concept "dual pixel RAW". "Dual pixel RAW" is a special RAW file that contains two sets of information about dual-pixel CMOS (normal point of view and the disparity information) in one of the RAW file.

By treatment with dual pixel RAW optimizer of DPP, the following three functions are provided.

- Image micro-adjustable (Image Micro-adjustment): resolving power from the depth information of the dual pixel RAW file to fine-tune the position so as to maximize the (micro-adjustable).

- Bokeshifuto (Bokeh Shift): to change the perspective of the previous blur in order to obtain a better blur.

- Ghost reduction (Ghosting Reduction): to suppress the occurrence of ghost and flare.

The dual pixel RAW file now seems a lot more interesting, especially if it allows me to achieve critical focus by AF microadjusting the image after the fact (i.e. a bit of Lytro).




  
  LOG IN TO REPLY
Mark0159
I say stupid things all the time
Avatar
12,868 posts
Gallery: 39 photos
Likes: 239
Joined Mar 2003
Location: Hamilton, New Zealand
     
Aug 19, 2016 04:07 |  #125

Dual Pixel RAW: One of the RAW file recording both the normal image and the parallax information. Digital Photo Professional of the “Dual Pixel RAW Optimizer” in the can: three of image processing

Image Micro-adjustment: by using the depth information, fine adjustment enables the position of the maximum sharpness and resolution
Bokeh Shift: to change the position of the previous blur (original: repositions the viewpoint of foreground bokeh for a more pleasing result)
Ghosting Reduction: reduces the ghost and flare

http://www.canonwatch.​com …ge-detailed-specs-leaked/ (external link)


Mark
https://www.flickr.com​/photos/52782633@N04 (external link)
Canon EOS 6D | Canon EF 24-105mm f/4L IS USM, EF 17-40mm f/4L USM, EF 50mm f/1.4 USM, EF 100mm f/2.8 Macro USM | Tamron SP 35mm F1.8 Di VC USD | Canon Speedlite 550EX -|- Film | Canon EOS 3 | Olympus OM2 | Zuiko 35mm f2

  
  LOG IN TO REPLY
sploo
premature adulation
2,402 posts
Gallery: 1 photo
Likes: 460
Joined Nov 2011
Location: West Yorkshire, UK
Post edited over 2 years ago by sploo.
     
Aug 19, 2016 04:37 |  #126

frankchn wrote in post #18099715 (external link)
Some more information on Dual Pixel RAWs from Digicame-Info: http://digicame-info.com …08/eos-5d-mark-iv-12.html (external link)

The dual pixel RAW file now seems a lot more interesting, especially if it allows me to achieve critical focus by AF microadjusting the image after the fact (i.e. a bit of Lytro).

That kinda makes some sense - with the data from both pairs of pixels you have two slightly shifted images (that's effectively the point of phase autofocus), so there probably are clever things you can do with the data.

I wonder if it's new in the sense that in previous bodies the dual pixel data wasn't available at image capture time (the camera only "sees" the sum of each pair of pixels), or whether it's actually just a firmware change - to request (from the hardware) and then store values for each pair of pixels. It wouldn't surprise me if it's the latter, and that might mean a future firmware upgrade could add that functionality to the 1Dx2; maybe even the 80D. Obviously we can't know how Canon's internals are working (other than the MagicLantern guys ;-)a) so whether it'd be feasible on older bodies too is unknown.

EDIT: With more effective captured resolution, I wonder if that would also allow more aggressive noise reduction (a bit like the 5Ds). Maybe that's not the point and the data wouldn't be applicable for that, but there are smart people out there who can use data to do innovative things (e.g. using multiple frames of video to work out what on a single frame is noise, and then clean up those individual frames). So, it's not beyond the boundaries of reality that in time people will be able to "get" more out of that dual pixel data.


Camera, some lenses, too little time, too little talent

  
  LOG IN TO REPLY
John ­ Sheehy
Goldmember
2,586 posts
Likes: 326
Joined Jan 2010
     
Aug 19, 2016 08:06 |  #127

sploo wrote in post #18099741 (external link)
EDIT: With more effective captured resolution, I wonder if that would also allow more aggressive noise reduction (a bit like the 5Ds).

Well, there's no question that if this were a simple case of RRGGRRGGRRGG on one line (and GGBBGGBBGGBB on the next) with double the resolution in the horizontal direction, there would be noise information in the difference between a pair of reds or greens which would not be present if they were read out as a unit, because the greater the difference between the members of the pair, the more likely it is that it is not due to optics, but to noise impulse. In a normal Bayer sensor, the neighboring pixels have different color filters (except diagonal green neighbors, whose centers are 1.41x as far from each other as horizontal and vertical neighbors), and can have great differences from their neighbors for optical/color reasons. Having data with the same color filtration so close to each other in an RRGGRRGGRRGG situation greatly increases the confidence of noise vs detail. To have that level of confidence with a normal Bayer CFA, you would need a much higher pixel density.

What I do not know is how much difference there can be between the two members of the pair due to what they see optically in a dual-pixel PDAF application, so I don't know how useful the noise confidence can be.




  
  LOG IN TO REPLY
sploo
premature adulation
2,402 posts
Gallery: 1 photo
Likes: 460
Joined Nov 2011
Location: West Yorkshire, UK
     
Aug 19, 2016 09:34 |  #128

John Sheehy wrote in post #18099828 (external link)
Well, there's no question that if this were a simple case of RRGGRRGGRRGG on one line (and GGBBGGBBGGBB on the next) with double the resolution in the horizontal direction, there would be noise information in the difference between a pair of reds or greens which would not be present if they were read out as a unit, because the greater the difference between the members of the pair, the more likely it is that it is not due to optics, but to noise impulse. In a normal Bayer sensor, the neighboring pixels have different color filters (except diagonal green neighbors, whose centers are 1.41x as far from each other as horizontal and vertical neighbors), and can have great differences from their neighbors for optical/color reasons. Having data with the same color filtration so close to each other in an RRGGRRGGRRGG situation greatly increases the confidence of noise vs detail. To have that level of confidence with a normal Bayer CFA, you would need a much higher pixel density.

What I do not know is how much difference there can be between the two members of the pair due to what they see optically in a dual-pixel PDAF application, so I don't know how useful the noise confidence can be.

Thanks John - interesting info as always.

The thing I don't quite understand (with DPAF) is, is it just like having a sensor with twice the horizontal resolution (where each pair of adjacent pixels are "summed" on readout), or is there some optical/filtering thing going on at the sensor level; such that there's (deliberately) a greater difference between the two pixels in order to calculate phase information.

If it is just like a double resolution sensor then it almost seems odd they didn't just produce higher resolution output files in the first place (for earlier DPAF capable bodies). Even if it didn't make a big difference in reality, it's a good tool in the marketing war (more MP).

That said, I suppose the whole pipeline then has to deal with twice the number of pixels; whereas combining the pairs early on means only dealing with half the actual sensor resolution.


Camera, some lenses, too little time, too little talent

  
  LOG IN TO REPLY
Talley
Talley Whacker
Avatar
11,064 posts
Gallery: 45 photos
Likes: 2720
Joined Dec 2011
Location: Houston
     
Aug 19, 2016 09:49 |  #129

It's not double resolution... It's double color data that has been shifted slightly. Each split pixel uses the same bayer layer. but what this does is create a higher sized file with more color data for the same resolution file. If you look at the specs the size didn't increase but the data dramatically did.

Isn't this sort of like Foveon tech? So what we'll end up with is a sharper image with better color and lower noise since each pixel probably cancels SOME noise out...

I'm uber excited to find out what exactly it does though!


A7rIII | A7III | 12-24 F4 | 16-35 GM | 28-75 2.8 | 100-400 GM | 12mm 2.8 Fisheye | 35mm 2.8 | 85mm 1.8 | 35A | 85A | 200mm L F2 IS | MC-11
My Gear Archive

  
  LOG IN TO REPLY
CyberDyneSystems
Admin (type T-2000)
Avatar
49,439 posts
Gallery: 156 photos
Likes: 5992
Joined Apr 2003
Location: Rhode Island USA
     
Aug 19, 2016 09:51 |  #130

re: MJPEG vs. more compressed files like H264/265

Is there a camp asking that Canon drop MJPEG?
I would assume that the pro h264 camp is asking for a smaller file size/more compressed OPTION.

EG: Please give me the option to shoot in a compressed jpeg file size for thousands of images on my 16GB card vs. shooting RAW for a few hundred shots.

People chose to forgo the RAW files superior image quality all the time, and though I shoot RAW, I would never want a DSLR that did not have the jpeg options we now have.

So I can see the desire/need for a more compressed file type, but don't see the argument against the option.


GEAR LIST
CDS' HOT LINKS
Jake Hegnauer Photography (external link)

  
  LOG IN TO REPLY
TeamSpeed
01010100 01010011
Avatar
35,584 posts
Gallery: 102 photos
Best ofs: 2
Likes: 4994
Joined May 2002
Location: Cannelton
     
Aug 19, 2016 09:52 |  #131

Here is a Canon writeup/video on DPAF, not sure if it helps with the questions though. It sounds like now Canon is allowing all the data to be written out so that during post processing you have extra capabilities on micro adjusting your focus?

http://cpn.canon-europe.com …cal/eos_70d_tec​hnology.do (external link)


Past Equipment | My Personal Gallery (external link) My Business Gallery (external link)

  
  LOG IN TO REPLY
bacchanal
Cream of the Crop
Avatar
5,284 posts
Likes: 21
Joined Jan 2007
Location: Fort Wayne, IN
Post edited over 2 years ago by bacchanal.
     
Aug 19, 2016 10:23 |  #132

CyberDyneSystems wrote in post #18099891 (external link)
re: MJPEG vs. more compressed files like H264/265

Is there a camp asking that Canon drop MJPEG?
I would assume that the pro h264 camp is asking for a smaller file size/more compressed OPTION.

EG: Please give me the option to shoot in a compressed jpeg file size for thousands of images on my 16GB card vs. shooting RAW for a few hundred shots.

People chose to forgo the RAW files superior image quality all the time, and though I shoot RAW, I would never want a DSLR that did not have the jpeg options we now have.

So I can see the desire/need for a more compressed file type, but don't see the argument against the option.


According to EOSHD (on the 1DxII faq):
http://www.eoshd.com …n-1d-x-mark-ii-video-faq/ (external link)

Why MJPEG and at what bitrates?

According to Ebrahim’s Canon source the camera’s weather sealed body and lack of fan excluded the use of the more complex XF / H.264 codec of the C300 Mark II. As the 1D X Mark II is built for absolute reliability, overheating must be avoided. If you need 4K H.264 but are willing to suffer 3 minute time limits and regular problems with overheating during a shoot consider the Nikon D5.

The downside of MJPEG is that due to the ‘simple’ encoding it is inefficient. File sizes are enormous but image quality is maintained.

The bitrate for 4K 24/25/30p is 500Mbit/s like the 1D C. The bitrate for 4K 60p is 800Mbit/s. At that data rate a 64GB card fills up in 10 minutes, so 60fps is better kept as an option for short bursts where a mild slow-mo effect will be applied in post.


Drew A. | gear | photosexternal link

  
  LOG IN TO REPLY
CyberDyneSystems
Admin (type T-2000)
Avatar
49,439 posts
Gallery: 156 photos
Likes: 5992
Joined Apr 2003
Location: Rhode Island USA
     
Aug 19, 2016 10:33 as a reply to  @ bacchanal's post |  #133

Ahh, interesting, the compression = CPU overhead, (of course) and thus not a simple firmware update without sacrificing reliability.


GEAR LIST
CDS' HOT LINKS
Jake Hegnauer Photography (external link)

  
  LOG IN TO REPLY
Talley
Talley Whacker
Avatar
11,064 posts
Gallery: 45 photos
Likes: 2720
Joined Dec 2011
Location: Houston
     
Aug 19, 2016 10:54 |  #134

Yeah and if you wanted video go buy a video camera is what they are saying lol. Like their cinema line


A7rIII | A7III | 12-24 F4 | 16-35 GM | 28-75 2.8 | 100-400 GM | 12mm 2.8 Fisheye | 35mm 2.8 | 85mm 1.8 | 35A | 85A | 200mm L F2 IS | MC-11
My Gear Archive

  
  LOG IN TO REPLY
SuzyView
Cream of the Crop
Avatar
32,073 posts
Gallery: 5 photos
Likes: 117
Joined Oct 2005
Location: Northern VA
     
Aug 19, 2016 11:02 |  #135

HAHA! Like I have money for that line of products. :) Still, I am another who cannot shoot just RAW. The JPEG option is always good for me. I can post images straight from stored files. But I love editing in RAW. Just wish Canon would hurry up. I know it's just a few more days, but I'm getting really excited!


Suzie - Still Speaking Canonese!
5DIV, 7D2, SL1, G12, iPad Pro, iPhone 6P, 5 L's & 2 Primes, 25 bags.
My children are the reason, but it's the passion that drives me to get the perfect image of everything.

  
  LOG IN TO REPLY
sponsored links
(this ad will go away when you log in as a registered member)

43,907 views & 83 likes for this thread
5D Mark IV announced August 25th
FORUMS News & Rumors Camera Rumors and Predictions 
AAA
x 1600
y 1600

Jump to forum...   •  Rules   •  Index   •  New posts   •  RTAT   •  'Best of'   •  Gallery   •  Gear   •  Reviews   •  Member list   •  Polls   •  Image rules   •  Search   •  Password reset

Not a member yet?
Register to forums
Registered members may log in to forums and access all the features: full search, image upload, follow forums, own gear list and ratings, likes, more forums, private messaging, thread follow, notifications, own gallery, all settings, view hosted photos, own reviews, see more and do more... and all is free. Don't be a stranger - register now and start posting!


COOKIES DISCLAIMER: This website uses cookies to improve your user experience. By using this site, you agree to our use of cookies and to our privacy policy.
Privacy policy and cookie usage info.


POWERED BY AMASS forum software 2.1forum software
version 2.1 /
code and design
by Pekka Saarinen ©
for photography-on-the.net

Latest registered member was a spammer, and banned as such!
1070 guests, 368 members online
Simultaneous users record so far is 15144, that happened on Nov 22, 2018

Photography-on-the.net Digital Photography Forums is the website for photographers and all who love great photos, camera and post processing techniques, gear talk, discussion and sharing. Professionals, hobbyists, newbies and those who don't even own a camera -- all are welcome regardless of skill, favourite brand, gear, gender or age. Registering and usage is free.