Objective Quality Measurement of MKV?

[hcour]

Pardon if this is an ignorant question, I'm new to this stuff. I found a program here that purports to give objective measurement of the quality of avi files:

http://compression.ru/video/quality_measure/video_measurement_tool_en.html

Assuming something like this actually works, is there such a program that would do the same for mkv files?

Thanks much,
Harold

[Kurtnoise]

Matroska is just a container, not a codec...So, quality is relative to the streams embedded.

Reading the page, you can open mkv files with this tool by using an avisynth script.

[Dark Shikari]

It's not a very useful application--at least no more useful than anything else that relies on PSNR/SSIM.

The graphs of quality over time are sometimes useful to compare the ratecontrol strategies of various encoders, however.

[laserfan]

I found a program here that purports to give objective measurement of the quality of avi files...Funny you would post this today--I just found it myself and am running it to compare a 26Gb file with a 7Gb x264 re-encoding, hoping it will identify all the pixelated frames in my re-encode (I'm trying to sort why my re-encodings are failing in this way).

Anyway this comparison has taken 8 hours so far on my 3.2GHz P4 though it looks like I should see some results before too long now! :p

This is where I see potential value in the tool, in comparing backups to originals.

[Dark Shikari]

Funny you would post this today--I just found it myself and am running it to compare a 26Gb file with a 7Gb x264 re-encoding, hoping it will identify all the pixelated frames in my re-encode (I'm trying to sort why my re-encodings are failing in this way).

Anyway this comparison has taken 8 hours so far on my 3.2GHz P4 though it looks like I should see some results before too long now! :p

This is where I see potential value in the tool, in comparing backups to originals.You know that x264 prints SSIM and PSNR too, right? ;) And unlike MSU's, its implementation isn't horribly slow...

[laserfan]

You know that x264 prints SSIM and PSNR too, right? ;) And unlike MSU's, its implementation isn't horribly slow...But Dark that doesn't help for my purpose: as I said I'm looking to compare my re-encode to the original on a frame-by-frame basis.

The MSU VQM tool was a complete bust: crashed after 9 hours with an MS runtime error, posted an "imcomplete" csv file that is incomprehensible to me, and wrote-out a couple of BMPs that show nothing.

If anyone here knows of a tool that will compare two files frame-by-frame, by all means let me know as MSU's doesn't work. :o

[mikeytown2]

http://avisynth.org/mediawiki/Compare

RGBDifference(clip1, clip2)
http://avisynth.org/mediawiki/ScriptClip

[laserfan]

http://avisynth.org/mediawiki/Compare

RGBDifference(clip1, clip2)
http://avisynth.org/mediawiki/ScriptClipHey thanks, I will give Compare a try! :)