or Connect
AVS › AVS Forum › Video Components › Home Theater Computers › 1920x1080 24/25/30fps Test Patterns - detect pulldown issues
New Posts  All Forums:Forum Nav:

1920x1080 24/25/30fps Test Patterns - detect pulldown issues

post #1 of 21
Thread Starter 
The following are updated Test Patterns to help discern any video judder issues in mismatches between frame rates and refresh rate.

Test Files: http://www.megaupload.com/?d=JEH1719D These Frame Rate Test files are a high bit rate 1920 x 1080 MPEG2 (with a blank AC3 sound track so it works with Reclock) in a M2T container taking up about 1GB each (but all three are in the 45MB Zip File) and include files encoded to refresh rates used in:
* Blu-ray (24 unique frames encoded at 23.976fps)
* PAL (25fps)
* NTSC (30 unique frames encoded at 29.97fps)

How To Use: Each file has a set of unique images that repeat every second for five minutes so you have plenty of time to observer how smooth the video is rendered. Change your monitor refresh rate to match that of the file you are going to play, then play it and keep your Eye on how smoothly the Red Dot moves across the screen. It should be nice and smooth but you may see issues such as pauses (repeating frames) or jumping forward a bit (dropping frames).

For "fun" or for those that don't think telecline judder as used by NTSC transfers looks bad, just play the 23.976 file back on your monitor/TV set at 60hz!

EDIT: I've noticed that these test files also highlight more than just the fluidity of video playback:
Pulldown / Frame Interpolation: One thing I've noticed that is interesting with the test files, it shows you what pull down and frame interpolation is being applied by the PC and/or TV. Each of the test files is only encoded with only 1 red square per frame. So when playing say the NTSC test file (29.97) you see two red squares when played at 60hz...in this case (I guess) the video renderer or TV displays each frame twice but overlaps them with the following frame. Now the interesting thing is when I play the Blu-ray test file (23.976) at 24hz I see 3 red boxs on my setup. As this is not being done in the PC, it looks like it must be the TV doing a 3:3 frame interpolation (though the documentation on the PIO LX608 is pretty vague on it's settings and options here).
Pixel Refresh: I also notice a very faint red trail (an erratic scattering of pixels) trailing the red box for a couple of mm at most


For these that are interested and if there are any odd visual artifacts, these test files were made form individual PNG slides, authored in Vegas and rendered using the MainConcept encoder.

Enjoy,
Nathan
LL
post #2 of 21
What do you mean by "Pull Down" issues?

Pulldown (telecine) means interlacing/repeating frames of a 23.976fps or 25.000fps progressive video stream in 3:2 or 2:2 pattern to produce interlaced video streams at 59.94 or 50.000 fields per second for SD/HD broadcast (or DVD). Pulldown detection (inverse telecine) is detecting this pattern in this 29.97fps or 25fps interlaced video stream, removing extra fields and deinterlacing, thus recovering the original 23.976fps or 25.000fps progressive video stream.
post #3 of 21
Thread Starter 
Quote:
Originally Posted by renethx View Post

What do you mean by "Pull Down" issues?

Pulldown (telecine) means interlacing/repeating frames of a 23.976fps or 25.000fps progressive video stream in 3:2 or 2:2 pattern to produce interlaced video streams at 59.94 or 50.000 fields per second for broadcast (or DVD). Pulldown detection (inverse telecine) is detecting this pattern in this 29.97fps or 25fps interlaced video stream, removing extra fields and deinterlacing, thus recovering the original 23.976fps or 25.000fps progressive video stream.

You will also get this any time there is a difference between the fps of the source material and the frequency of the monitor. For instance if you play 25fps material on a 60hz monitor, most frames will be repeated twice but some will need to be repeated 3 times...or, on a 50hz monitor when playing 30fps material 10 frames per second will need to be dropped. Depending how your HW handles this and your sensitivity to it you may be annoyed at the effect.

Some of the Pulldown patterns are listed at http://en.wikipedia.org/wiki/Telecine
post #4 of 21
OK, I see your point. Yours has nothing to do with interlacing/deinterlacing in broadcast. It's not like HQV Benchmark.

Hmm, interesting patterns. Will you be able to create a 23.976fps (the frame rate of the majority of movies) pattern of 60 sec duration?
post #5 of 21
Thread Starter 
Done (updated download link in Post 1) now also includes a:
- 23.976 version of the 24 segment and a
- 29.97 version of the 30 segment test pattern

To make a 60 sec version you could just binary copy 3 of the 20sec clips together (else the download size will blow out)
post #6 of 21
How about a version that is 24.000Hz? Being able to compare to 23.976Hz. And yes, its good if they are longer then 20s. Seing these small hiccups sometimes needs atleast 60s.


Thanks alot for youre work.


Edit: spoke to soon, downloaded the files and saw both 24p and 23,976. However, how about marking them as 24.000 and 23.976. Its a littlebit confusing, "24p" is infact 23,976Hz in the absolute number of cases.
post #7 of 21
Thread Starter 
Ok - I'll make some longer patterns and change the marking to show both the frames they contain and the frequency encoded at, eg:
- 24fps @ 23.976
etc....
post #8 of 21
Thread Starter 
Updated the first post with longer (now 60sec) and more test files. Also please read the note regarding the stuff encoded at 23.976 or 29.97.
post #9 of 21
Downloaded the files and played them on my Samsung C430. (In Europe this model plays video files with the built-in media player)

The 30p files had the smoothest motion, but 24p and 25p were fine, too. They were just not as smooth as the 30p versions.

What am I supposed to see?
post #10 of 21
Thread Starter 
It "should" be smoothest when you TV is at the same refresh rate (or a multiple off) the fps of the test file you are playing. Eg if your TV is doing 50hz (which I presume if you are in Europe) then the 25fps should be the best as each segment of the test patter is just displayed twice. The 30fps when played on a 50hz TV would have some segments played twice and some only once giving a not so smooth result (unless it is changing to 60hz). etc etc

If the 30fps stuff is smoother than the others it may be that your TV is playing them back at 60hz OR that you simply just prefer a higher fps over pull down judder from mismtatched frame rates and frequences!
post #11 of 21
Thread Starter 
Updated the First Post with new (and "better") test pattern files based on Blu-ray, NTSC, and PAL refresh rate encodes.
post #12 of 21
Quote:
Originally Posted by jmone View Post

* Blu-ray (24.976fps)

There's a typo in there. It should be 23.976
post #13 of 21
Thread Starter 
Thanks - fixed and updated some content on the first post again
post #14 of 21
jmone, Excellent resource! I have been using another set of SD files with a vertical bar moving across the screen and found that to be pretty useful in identifying mismatched refresh and source frame rates.

[First link in the external links section here: http://en.wikipedia.org/wiki/Frame_rate , but the link is often down]

Would it be too much to request for 50 fps and 59.97 fps samples too? Thanks!
post #15 of 21
could anyone reupload this please?
post #16 of 21
Quote:
Originally Posted by visca blaugrana View Post

could anyone reupload this please?


Would like them also.

post #17 of 21
Quote:
Originally Posted by Mrkazador View Post


Would like them also.

+1
post #18 of 21
Thread Starter 
Here they are again (45MB in one Zip) - this time at DropBox if that helps!

https://dl.dropbox.com/u/59011278/FrameRateTestFiles.zip
post #19 of 21

Thanks a lot!

post #20 of 21
thanks
post #21 of 21
Thread Starter 
FYI - I tried taking a photo of how my Kuro displays the 24fps test pattern but that just does not work so I mocked up what I see in paint (sort off). It looks like this in with Drive Mode 2 and with PureCinema set to "Advance" and the group of three red squares then travels diagonally. With it OFF it is five squares, if it is the other two settings it seems to alternate between two and three. As mentioned, in the test file each individual frames moves the red square completely so they touch but don't overlap - it seems the Kuro interpolates.


I should say I only see this on the Kuro - if I run the same test patterns on my PC's LCD screen there are no separate images as the red box moves (but other issues like blur and very evident strobing on the black white boarder area).
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Home Theater Computers
AVS › AVS Forum › Video Components › Home Theater Computers › 1920x1080 24/25/30fps Test Patterns - detect pulldown issues