Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

5D II Zone Plate...

Jannard

Jannard

Red Leader
Joined
Dec 28, 2006
Messages
8,248
Reaction score
7
Points
0
Just as a reference, this is a Zone Plate shot with the Canon 5D II. Pretty easy to see what's going on with row-skipping. Getting motion from a slow clock-speed sensor is... well.

Jim

1228226829.jpg
 
Here's the above image with a gaussian blur and ImageMagick Lanczos resample to DVD size:
5d2-zone-480.jpg


Inexpensive, large, quality: pick any two.

Canon picked cheap and large, so quality went out the window.

RED is planning large and quality (FF35) or cheap and quality (2/3").
 
Ooo! That's ugly. You could have warned me Jim. I think I nearly threw up...

Graeme
 
Daniel Browning said:
Here's the above image with a gaussian blur and ImageMagick Lanczos resample to DVD size:
5d2-zone-480.jpg


Inexpensive, large, quality: pick any two.

Canon picked cheap and large, so quality went out the window.

RED is planning large and quality (FF35) or cheap and quality (2/3").
Click to expand...

Still butt ugly.

Looks like my mom's third generation vhs copy of a TV show recording.
 
So are they skipping, like, every other pixel?

Sensor resolution is stated at 5616 x 3744 (21.0 MP).
 
Interesting, but I'm still not sure about the line skipping. There can be other causes for this, although it may partially attributable to line skipping. Either way an artifact is an artifact.
 
Yes, the Canon is not the best.....but it is better than HDV. Yes????
 
not symmetric

not symmetric

Let us be explicit on what is obvious to Jim.

Surely you noticed that the Moiré Pattern is not symmetric (horizontal and vertical interference is not the same). This would suggest that

(a) every third line in the 16/9 part of the sensor is read
(b) the lines are read completely and resampled to 1920
(c) luminance and color are interpolated on line level from the rgbgrgbg... pattern

Yes?

(d) These lines compose the image to be compressed, maybe after applying some filters.

5D II users should definitely know this, since they must be scared of horizontal and nearly horizontal lines, less of vertical lines. Such an information should figure in the user manual, really.
 
Can we see the same shot with R1, Jim?
 
Tom said:
So are they skipping, like, every other pixel?
Click to expand...

No, they're using a new DDG algorithm and RR sampling. The sensor read out electronics measure each pixel and assign it a "D" for skipped lines and a "G" for read lines. Then another sample is taken when one of the lines is chosen to be sent over to the other "side", which is known as RR sampling.

If you listen to the camera closely, you can hear this DDG algorithm working. The two "D" sound like a faint "Duck...duck..." and "G" goes "GOOSE!" RR sampling is a sing-songy "Red Rover, Red Rover, send a pixel right over!".
 
Jim, please put out the same shot of R1 & Scarlet tomorrow (3rd Dec) so then a lot of us would cheer up more :-D.
 
RED-Tank said:
Jim, please put out the same shot of R1 & Scarlet tomorrow (3rd Dec) so then a lot of us would cheer up more :-D.
Click to expand...

We have posted a Zone Plate somewhere (looks perfect)... we can post one tomorrow if you don't mind delaying the announcement a day or two...

Jim
 
Please no delay.
 
Again thank you very, very much Jim and all of the Red team!

Your work is truly appreciated by me. It is a dream come true.
 
Wow! It is terrible! Thanks for doing the test, Jim. And, thanks for the R1:love:
 
You must log in or register to reply here.
Back
Top