Quantcast HCEnc: Some Things to HC v0.15b - digitalFAQ.com Forums [Archives]
  #1  
06-21-2005, 01:40 PM
Encoder Master Encoder Master is offline
Free Member
 
Join Date: Mar 2003
Location: (^_^)
Posts: 504
Thanks: 0
Thanked 0 Times in 0 Posts
Hey guys.

Here I want to say some things to this great new version of this encoder.

First I have compared the quality for Low bitrate encodes like KVCD. And HC does the job better then TMPGEnc's MPEG-1 conversion. In VDubMod it looks with HC minimal sharper but I checked both clips with FritzFramalyzer:

http://www.josef-kirner.homepage.t-o...downloads.html

This Programm compares the Videos from the original. For example PSNR and some other things. This is very good to compare Videos and encoders. And the programm says the same: HC gave me better quality then TMPGEnc.


Now, I test HC with only 1-B-Frame with CQ mode and the same value for 2 and 1 B-Frame. 1-B-Frame compressed better then 2. If the samples have the same average Bitrate then 1-B-Frame is the better quality. This also say Fritzframalyzer.


But as I load my movies into VDub I was confused, because of the GOP structure. The GOPs are always closed and I don't check the ClosedGOP button. What I done wrong, or is this a failure because I miss the B-Frame at the end of the GOP?

Thank you for listening.
Reply With Quote
Someday, 12:01 PM
admin's Avatar
Site Staff / Ad Manager
 
Join Date: Dec 2002
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #2  
06-21-2005, 04:34 PM
Prodater64 Prodater64 is offline
Free Member
 
Join Date: Mar 2003
Location: Palma de Mallorca - España
Posts: 2,925
Thanks: 0
Thanked 0 Times in 0 Posts
Well, I downloaded. Now how do you interpret results?

PSNR, RGB distance, etc.
Reply With Quote
  #3  
06-21-2005, 06:00 PM
Encoder Master Encoder Master is offline
Free Member
 
Join Date: Mar 2003
Location: (^_^)
Posts: 504
Thanks: 0
Thanked 0 Times in 0 Posts
I don't test the new version with the new comparison mode but with the RGB value it's the same as the PSNR. 0 is the best and closer to this the picture is nearer to the original. Or what do you mean?
Reply With Quote
  #4  
06-21-2005, 06:35 PM
Prodater64 Prodater64 is offline
Free Member
 
Join Date: Mar 2003
Location: Palma de Mallorca - España
Posts: 2,925
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by Encoder Master
I don't test the new version with the new comparison mode but with the RGB value it's the same as the PSNR. 0 is the best and closer to this the picture is nearer to the original. Or what do you mean?
Excuse me, but I don't know compare streams with this 2 properties, neither its meaning.
Could you explain me it a little bit more?
Reply With Quote
  #5  
06-21-2005, 06:47 PM
Encoder Master Encoder Master is offline
Free Member
 
Join Date: Mar 2003
Location: (^_^)
Posts: 504
Thanks: 0
Thanked 0 Times in 0 Posts
I can only say: Is the value of "Euklidischer Wert in RGB" and/or PSNR closer to "0" then it's better. If you have a test stream and then you encode with for example TMPGEnc and then HC you load first the Main movie and then the TMPGEnc produces and you get some values. Then the same with HC and you have to compare this values. Which encoder gives you lower values this is better. Hope this help you.
Reply With Quote
  #6  
06-22-2005, 04:12 AM
incredible incredible is offline
Free Member
 
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to incredible
What about your EYES?
I do only make comparisons using my eyes on a PC Monitor or my Panasonic 32" 16:9 TV Set. In a clear served RGB input out of the to RGB output settet SAP There you can even see if a GOP is set to 15 or 30 (the picture is that clear and detailed that you can recognise the "pumping" at the change to the next "I" Frame. Even on orig DVDs!
Thats why GOPs longer than 15 is no longer an issue for me and my taste.

Today you got a normal TV and you see your backups in a quality your TV can offer. But some Day, you do purchase a good LCD or a good CRT like mine mentioned above and THEN you can see using your good eyes what your encodings are made of.

So ..... I do compare outputs NOT on mathematical values, but at least on the best comparison technique given: The PC Monitor.
Do import the encoded and the orig stream in an avisynth script like this:

Code:
encoded1  = mpeg2source("encoded1.d2v").subtitle("encoded1",10,10)
encoded2  = mpeg2source("encoded2.d2v").subtitle("encoded2",10,10)
Interleave(encoded1, encoded2)
Now open that in Vdub und move frame per frame forward.
you then can switch to the orig and the encoded one by back and forth.
THATs a comparison you can trust in as your EYES will be the ones which do enjoy the movie finally.
Reply With Quote
  #7  
06-22-2005, 05:59 AM
Encoder Master Encoder Master is offline
Free Member
 
Join Date: Mar 2003
Location: (^_^)
Posts: 504
Thanks: 0
Thanked 0 Times in 0 Posts
You're absolutly right. First I test with my eyes but if it's not sure then I use this new tool. This only verify ma results.

I will test the script.
Reply With Quote
  #8  
06-22-2005, 06:14 AM
incredible incredible is offline
Free Member
 
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to incredible
Quote:
Originally Posted by Encoder Master
You're absolutly right. First I test with my eyes but if it's not sure then I use this new tool.
But your eyes are the ones which will see the movie finally.

And thats what I do not understand. People are testing encoders differencies and do proclaim a breakthrough if an encoding is 0,08% better than another one out of a diff. encoder. And that very often just testet on ONE testmovie.

Instead of testing if the quality rises one permil I do mainly test such encoders on their handling, options and speed. If I do obtain an encoding which is 0,08% worse but done in 70% of the encodingtime I do chosse THAT one. As I would spend more time in "enhancing" a source like avisynthing approaches. As THAT often will for your eyes be recognisably more than 20-30% more quality even on the encoding compared to the source at a targetsize often out of 80-60% of originalDVDStreamsize.

The mpeg1/2 teasement/developement IMHO has reached its "for human eyes differencies recognisable" end.

The next would be the developement in for instance mp4/x264. As these technologies are "younger" and do NOW got more potential/hype.
And some day Im shure if 90% of all offered SAPs will do support mp4 and the included codecs ... my Targets will be mp4 incl. Chapters and Menuegenerating based on XMT/VRML.

And even if only 60% of the Players will support it, no matter .... thats like when I have been doing KVCDs incl. longer Gops and diff. resolutions at mpeg1 .... so I choosed a SAP which supported the proper playback of our & Karls video developement.
Reply With Quote
  #9  
06-23-2005, 12:54 PM
incredible incredible is offline
Free Member
 
Join Date: May 2003
Location: Germany
Posts: 3,189
Thanks: 0
Thanked 0 Times in 0 Posts
Send a message via ICQ to incredible
Seems HC Encoder .... expects the source to be at FULL luma range, means an internal forced "fixed" Luma scaling to a "proportion!" of 16-235 will be done, no matter if the source is already 16-235 .... can someone duplicate that? Would mean 16-235 Sources will result in too less contrast.
Reply With Quote
  #10  
06-23-2005, 01:47 PM
Dialhot Dialhot is offline
Free Member
 
Join Date: May 2003
Posts: 10,463
Thanks: 0
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by incredible
Seems HC Encoder .... expects the source to be at FULL luma range, means an internal forced "fixed" Luma scaling to a "proportion!" of 16-235 will be done, no matter if the source is already 16-235 .... can someone duplicate that? Would mean 16-235 Sources will result in too less contrast.
According to Hank, nothing is done on the luma (this was in 0.13, things can have changed now). When we had this discussionn the toppic was about effect of colormatrix. See there what Boulder said :
http://www.kvcd.net/forum/viewtopic....rmatrix#117191
Reply With Quote
  #11  
06-25-2005, 06:56 AM
hank315 hank315 is offline
Free Member
 
Join Date: Feb 2005
Posts: 38
Thanks: 0
Thanked 0 Times in 0 Posts
No, nothing changed on luma.
HC just takes the YUV planes as input and doesn't know if these planes are 0-255 or 16-235 scaled.
So I think if you input 0-255 scale, 0-255 scale will come out; 16-235 input will result in 16-235 output.

But I never tested this so correct me if I'm wrong.
Reply With Quote
Reply




Similar Threads
Thread Thread Starter Forum Replies Last Post
The gender of non-living things. kwag Off-topic Lounge 0 08-15-2004 12:16 AM
Mencoder: order of things in ini configuration file? Fluffbutt Video Encoding and Conversion 2 05-22-2004 01:38 PM
Avisynth: How to filter away these things? Razorblade2000 Avisynth Scripting 19 08-20-2003 09:37 PM
KVCD: Is there a general Intro to all things video?? miksmith Video Encoding and Conversion 4 07-21-2003 04:56 AM
KVCD: nero goes and buggers things up? tarsus Authoring VCD, DVD, Blu-ray 0 05-29-2002 03:45 PM

Thread Tools



 
All times are GMT -5. The time now is 03:08 PM  —  vBulletin © Jelsoft Enterprises Ltd