CQM comparison - V5 - Open

[Soulhunter]

Codename Teegedeck... ;D

Again with full length samples !!!


1. How is it supposed to work ???


I did a encode of this source...

The Matrix (German/PAL/R2) - The lobby shoot-out (03:08 min.)


With this AVS...

LoadPlugin("C:\Stuff\MPEG2DEC3.DLL")

Mpeg2Source("C:\Stuff\Lobby.d2v", idct=2)

Crop(0,78,720,420).LanczosResize(720,288)


And this settings...

XviD v.1.0.3 (Koepis) / Defaults / 2Pass @ 2200kbps (VHQ4)


With 5 different quantization matrices !!!


Then I named the files like this:

CLIP_NO_003.avi



2. But how to vote ???


Simple, watch the samples under your "usually" playback conditions...

Note: Please make sure that all post-processing is disabled !!!


If you are not able to see a difference between the samples...

- Reduce the distance to your screen

- Raise the brightness & contrast


If you are still not able to see a difference...

- Reduce the playback speed



When you got a good impression, vote like this...

5 = Best sample
4 = Above average
3 = Average
2 = Below average
1 = Worst sample


Then send me your results (via pm) like this...

CQM_NO_001 = 3.0

CQM_NO_002 = 1.5

CQM_NO_003 = 5.0

etc...


And after some time (~8 weeks), I post the results !!!



3. How to get the samples ???


1st. Click here... (http://service.gmx.net/mc/KOIlroyeVUQDWvBB9Rcy6nTnh3E8D6)

2nd. Press this button -> http://img60.exs.cx/img60/6854/bt_gmxmediacenterstarten.gif

3rd. Download all 5 zip files from the folder and unpack them !!!



Tia n' Bye

[kurt]

ah, nice! at least some higher bitrates :D

[niamh]

Here we go.......... :)
Any guideline on the voting marks?

[obieobieobie]

It's luck that the host you're using is such a good one. ~1 MB/s. Lovely. :)

[Soulhunter]

Originally posted by niamh

Any guideline on the voting marks?

- The best sample gets a 5

- The 2nd best gets a 4

- The 3rd best a 3

- etc...


Originally posted by obieobieobie

It's luck that the host you're using is such a good one. ~1 MB/s. Lovely. :)

And it has a 2GB transfer limit... >.<

So I have to prepare some alternative hosting asap !!!


Bye

[Shinigami-Sama]

ohh I beter get mine fast then *yoink!*
ok this one is gana be harder probly less blocking crop and smudging :D, good thing you fixed the rating hunter :D
*goes to get samples*

[708145]

Originally posted by Soulhunter

And it has a 2GB transfer limit... >.<

So I have to prepare some alternative hosting asap !!!

Bye

Another alternative is here (http://service.gmx.net/mc/5tnKToTIrNdzxBOHz68jcBn17wK9xc) (including another 2GB transfer limit :D )

bis besser,
Tobias

[Soulhunter]

Originally posted by 708145

Another alternative is here (http://service.gmx.net/mc/5tnKToTIrNdzxBOHz68jcBn17wK9xc)


Thanks, makes 6GB now... :)

Just added a additional link to my 1st post !!!


Bye

[Torian]

Hi,
I don't mean to offend anyone and the amount of work put into these tests but I was quite disappointed by the quality of _all_ the clips and would not call them high bitrate. I wonder what the q2-comptest percentage would be? (I guess something around 30%)

I finally wanted to participate but I don't see the point of scrutinizing these clips when the matrices lack bitrate so dearly that I would never encode like that.

I know that it is more difficult to compare matrices at 80% q2-comptest (of say the standard mpeg matrix) because I did that extensively for my own encoding a while ago but since I generally encode in at least that bitrate (for special movies I use Selur's at q2-80% ct) I saw no alternative.

Or did all of You cast the "near-transparency" aside and decided to go 1CD?

[Didée]

Originally posted by Torian
Or did all of You cast the "near-transparency" aside and decided to go 1CD?
Ermh, in case of The Matrix, 2200 kbps is in the range of a 3CD-encoding ... however this chapter draws some more bitrate than movie's average, that's true.

Too strong compression was not the problem here - the matrices surely were not forced to use (too) high quantizers.
Did you encode The Matrix yourself? Problem is, it just doesn't want to look good. The noise is not to underestimate. And worse, the source is, I dare to say, blurry. Both together are tricky.
For a nice looking encoding, I'd throw a filter or two at it - but that wasn't the purpose of this test...

[Soulhunter]

Originally posted by Torian

I don't mean to offend anyone and the amount of work put into these tests but I was quite disappointed by the quality of _all_ the clips and would not call them high bitrate. I wonder what the q2-comptest percentage would be? (I guess something around 30%)

Erm, here is the result of a Q2 (defaults + MPEG) encode... :rolleyes:

http://img73.echo.cx/img73/2157/mpeg1st3pb.png


Bye

[Torian]

Didée Did you encode The Matrix yourself?
I have to admit that I didn't but I couldn't imagine that a DVD would look so pathetic. (for example there is this grid between Neo putting down his bag and the two sitting security guys which gets totally blurred out and is only visible as a slight staircase-pattern on the guards.)

BTW: Thanks for 6o9, it is a good compromise for black and white and other grainy material (I don't like to blur all the grain)

Soulhunter
Erm, here is the result of a Q2 (defaults + MPEG) encode...
I stand corrected

Maybe You could enclose at least a small segment of the original mpeg2 stream (or all to really blow the filesize ;) ) for a better comparison because I usually examine also the truthfullness of a matrix aside from the general visual appearance.

[Soulhunter]

Originally posted by Torian

I stand corrected...

Nope, not really...

Your guess was "around 30%" but its more like 60-90% !!!


Originally posted by Torian

Maybe You could enclose at least a small segment of the original mpeg2 stream (or all to really blow the filesize ;) ) for a better comparison because I usually examine also the truthfullness of a matrix aside from the general visual appearance.

Well, its too late now...

Maybe I gonna upload a source sample for the next comparison !?!


Bye

[Soulhunter]

Three votes so far... :\


Bye

[*.mp4 guy]

I can't really see much difference between the clips on my crt, thats why I haven't voted. I could check on my lcd, but I usually avoid using it because it makes everything look bad.

[Shinigami-Sama]

I lost the clips in an accidental purge so I gatta get then again and find time to test for an hour or so

[Soulhunter]

...but I usually avoid using it because it makes everything look bad.

In this case you could try to sort them like this:

Bad -> Very bad -> Ultra bad etc... :\


Bye

[Backflip]

Have the results been posted for the V5 - CQM comparison?

[Soulhunter]

Have the results been posted for the V5 - CQM comparison?
Congratulation, you are the first one who noticed that this comparison never finished!

Unfortunately the number of participants (4?) was too low for reliable results... :\


Bye

[Backflip]

We need more participants! I might participate depending on the status of my data/transfer cap :(

[JarrettH]

Bump, mwahaha

Hey Soulhunter, I was looking for the "official" topic where you posted your matrices. Is there a topic you maintain like Shark's? ;)

I just replaced the versions I had with the ones in your sig since I haven't been to these forums in a long time. Can you tell me the difference between V2 and V3? Does V3 differ in that it is HVS compliant (like Shark's).

Thankee!

[Sharktooth]

Just a note... the "HVS compliant" thing on my matrices means "IEEE HVS Compliant".

[Soulhunter]

Hi JarrettH,

Soulhunters v3 is a "lets keep all the noise/grain" high-bitrate version of the standard MPEG matrix. It was made to encode hollywood flicks at 1024x576 -> 1 DVD5! But it works same good for D1 -> 1/2 DVD5 scenarios... Soulhunters v6 and v8 were made for mid/low bitrates (the usually 1 or 2 CD encodes). They are not IEEE compliant, the bracketed "HVS" just tells you that they make use of the HVS principle (quantizing high frequencys much stronger etc.). And no, I never posted a "official thread" coz I prefer that ppl play around with my CQMs to find out its advantages/disadvantages on their own (educational) plus, I really suck in advertising... XD


Bye

[JarrettH]

Hmm, I'm encoding Curb Your Enthusiasm and I think I'll stick with the MPEG one. I want to keep the same feel as the DVD.

Results for this test?:)