CQM comparison - V4 - Finished

[Soulhunter]

Once again...

Now 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 @ 1100kbps (VHQ4)


With 5 different quantization matrices !!!


Then I renamed 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, write down a number between 1-5 !!!

1 = Crap
2 = Bad
3 = Tolerable
4 = Good
5 = Perfect


Then send me your votes like this...

CQM_NO_001 = 2.0

CQM_NO_002 = 1.5

CQM_NO_003 = 3.0

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/JG8CzluxLaM8SDMfHCBqKYDZFqoFLI)

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 !!!



4. Finished...

The results !!! (http://forum.doom9.org/showthread.php?s=&postid=648109#post648109)



Thx n' Bye

[Teegedeck]

I'm glad you struggle on with another blindtest! :)

And now that you're closing in to medium bitrates it is becoming much more interesting for me! Thanks.

[CruNcher]

my decission has allready fallen i'll use Sharktooths V3ULR matrix in the next EDP build :P but i atend as allways :D

[Teegedeck]

Ow, and I had so hoped not to see something low quality, this time. :(
But at 1000 kbps, what ever did I expect? How long will it take till we get to a 2000 kbps roundup? ;)

[Soulhunter]

Originally posted by Teegedeck

How long will it take till we get to a 2000 kbps roundup? ;)

You mean the "lost" 2200kbps round I have skipped ???

Well, I gonna post it next after this round... ;)


Bye

[Teegedeck]

Voting will take some time, certainly not easy to judge...

[*.mp4 guy]

Yes they are all quite similar. Except for one that seems quite a bit more smoothed out then the others too me...

[Soulhunter]

Btw, what do you think about my new "less matrices but full length samples" concept ???


Bye

[Didée]

Generally I like the concept very much.

It's only a pity that, due to the fact i live in hi-tec country Germany which CANNOT SUPPLY ME WITH ANYTHING FASTER THAN 56k DIAL-UP, I'll not be able to participate in this round, again.

[*.mp4 guy]

Originally posted by Didée
Generally I like the concept very much.

It's only a pity that, due to the fact i live in hi-tec country Germany which CANNOT SUPPLY ME WITH ANYTHING FASTER THAN 56k DIAL-UP, I'll not be able to participate in this round, again.

:(, I've used dialup before. Definately not fun.

It was nice too have the full clip, however it didn't really seem necessary to rate the one that looked best. I think maybe 30 seconds with them all in a 7-zip file might be more conveniant for most peaople. Testing a smaller number of cqms at once tho seems like a good idea to me.

[Soulhunter]

@ Didèe

Hehe, hi-tec country... :D

Whats with a DSL connection via ASTRA (http://www.ses-astra.com/market/deutschland/internetviasat/nutzen/startseitenutzen.htm) ???


@ All

The next round is 4x50MB huge... :\

Would a additional "4x10MB" dial-up zip do it ???


Bye

[*.mp4 guy]

Imo 40 mb will take hours with dialup. Also 7-zip usually gets a lot more compresion then zip does, fyi. If for some reason you dont like 7-zip... well never mind.:D

[Soulhunter]

Originally posted by *.mp4 guy

Also 7-zip usually gets a lot more compression then zip does, fyi...

Well, not for AVI files... :\

http://img90.exs.cx/img90/9900/zip2ki.png (http://www.imageshack.us)

http://img90.exs.cx/img90/8013/7zip4pf.png (http://www.imageshack.us)

Btw, I zip'ed the files not to make them smaller...

I did this to prevent that you review b0rked files !!!


Bye

[Richard Berg]

600KB/s, nice server! I'll let you know in a few.

[Soulhunter]

Originally posted by Richard Berg

600KB/s, nice server! I'll let you know in a few.

And now imagine the upload speed... :D :D :D

If you know a better free hosting solution, plz tell me !!!


Bye

[*.mp4 guy]

:confused: ...I thought 7-zip was better for everything, I'll have to look into this. Also why would non-zipped files be borked? does zip have error protection?

[Sharktooth]

It's really hard this time. Detailwise they're all there, except one clip.
Can you elaborate some more on the voting?

[Joe Fenton]

Originally posted by *.mp4 guy
:confused: ...I thought 7-zip was better for everything, I'll have to look into this. Also why would non-zipped files be borked? does zip have error protection?

All archive formats have error detection, and many have the ability to include a certain percentage of redundancy for error correction as well.

No archiving format can be as good at all file types as some format is at one file type. If an archiver is designed for certain types of files, it normally outperforms any other archiver on those files.

[*.mp4 guy]

hmmmm. I just made a 7-zip of an avi that was smaller then the equivelent zip file...However I used the PPMd algorithm instead of LZMA
1 xvid in avifile test:

Original: 8,065,314 bytes
Zip: 7,869,867 bytes
LZMA: 7,871,227 bytes
PPMd: 7,822,962 bytes
Deflate64:7,872,532 bytes

2 xvids and 4 other avifiles, 40mb:

Original: 42,422,830 bytes
Zip: 41,585,185 bytes
LZMA: 41,611,095 bytes
PPMd: 41,698,635 bytes
deflate64:41,576,376 bytes

[Soulhunter]

Originally posted by Sharktooth

Can you elaborate some more on the voting?

Yeah, the difference is small...

Should I allow 1/4 playback speed ???


@ *.mp4 guy

Joe Fenton is right !!!

A bad AVI you could watch without noticing that its b0rked...

But a bad ZIP would show you a error message n' tell ya that its b0rked !!!


Bye

[Sharktooth]

Originally posted by Soulhunter
Yeah, the difference is small...

Should I allow 1/4 playback speed ???
Well, dont know if it is "correct", coz you dont watch a movie at 1/4.

[Soulhunter]

Originally posted by Sharktooth

Well, dont know if it is "correct", coz you dont watch a movie at 1/4.

Maybe I should change my 1st post like this... :\


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, write down a number between 1-5 !!!

1 = Crap
2 = Bad
3 = Tolerable
4 = Good
5 = Perfect

[...]

[Sharktooth]

Rise the brightness AND the contrast.

[Soulhunter]

Originally posted by Sharktooth

Rise the brightness AND the contrast.

Yeah, sure... :o

I have edited both posts !!!


Bye

[Sharktooth]

Now, can you send me the source at the usual gmail address?
I wanna play with it and x264...

[Soulhunter]

Originally posted by Sharktooth

Now, can you send me the source at the usual gmail address?
I wanna play with it and x264...

I gonna send you a pm after uploading the vfs files... ;)


Bye

[Sharktooth]

vfs?

[Soulhunter]

Originally posted by Sharktooth

vfs?

Mail providers dont like huge files, you remember ???

Hope you still have this VFS tool... :rolleyes:


Bye

[Sharktooth]

... no ...
multipart .zip or .rar (with no compression)?

[Soulhunter]

Ok, I gonna upload this VFS tool as well !!!

Hope you remember how to use it... ^^


Bye

[Sharktooth]

:scared:

[Richard Berg]

I know there's more detail in this scene (at least on the NTSC R1 disc) but I can't tell the difference between them except for the smoothed one.

[Sharktooth]

It's not easy to spot differences when there's high motion.
That's why they seem all to be the same...

[*.mp4 guy]

look at textures during pans etc to look for sliding. also look for ringing around edges. Not that you really have to look, those artifacts are present in all the encodes, just look for the one in whitch the presence of those artifacts is least offensive to your eyes. Also smoke can be a good indication of detail preservation.

[*.mp4 guy]

Soulhunter, how many people have given there results?

[Soulhunter]

Originally posted by *.mp4 guy

Soulhunter, how many people have given there results?

Except you ???

Zero... :(


Bye

[kurt]

don't be sad, soulhunter....
my votings will follow in the next week :)

[Sharktooth]

mine too. i just need some time to analyze the clips again and again... :)

[Spyn]

compare the vote result with SSIM result could be interesting...

[Shinigami-Sama]

Originally posted by Soulhunter
3. How to get the samples ???
1st. Click here... (http://service.gmx.net/mc/GEmrN9RdeNc4xlIJ5VVZe07lX4JziJ)
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
if only it would let me dl
or I could raed german to tell me what I'm doing wrong:rolleyes:

[Soulhunter]

Originally posted by Shinigami-Sama

if only it would let me dl
or I could raed german to tell me what I'm doing wrong:rolleyes:

Erm, whats your actual problem ???

Can't you reach this (http://i158.exs.cx/img158/4927/113zp.png) page after following step 1&2 ???

Can't you download the zip's via right-click -> "save target as" ???


Bye

[Shinigami-Sama]

ehh?
I think I just hit a 5min hiccup on the server -_-
now it;s letting me dl...:)

[Soulhunter]

Originally posted by Spyn

compare the vote result with SSIM result could be interesting...

Remembers me my first (http://forum.doom9.org/showthread.php?s=&threadid=76021) CQM odyssey... :D


Bye

[Shinigami-Sama]

ok
now that I;ve watched after that weird error >_>

opps
sorry
I was tired when I posted this
^^;

[kurt]

i think it would be better not to post here your results ---> pm soulhunter instead....
people should rate the clips by their own...

[*.mp4 guy]

Yes please remove your results, it could interfere with the comparisons "double blind" nature.

[Shinigami-Sama]

sorry about that
I guess I shouldn't start testing stuff at night anymore

[kurt]

check your inbox :D

[Soulhunter]

Ok, this makes 3 votes now !!!

Keep it coming guys... :)


Bye

[Soulhunter]

Just noticed the 4 weeks are over... ^^;

So, should I prolong the deadline another 4 weeks ???


Bye