Digital Video Forums  

Go Back   Digital Video Forums > Software Forums > DVD Rebuilder

Reply
 
LinkBack Thread Tools Search this Thread Rate Thread Display Modes
Old 29 Dec 2006, 11:11 AM   #1
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default Quality Comparison using "Mulholland Drive"

I've done this comparative test, following in the steps of UncasMS. This is my first post and I would like to share my thoughts with you. I've used DVD-Rebuilder with cce 2.70, Hcenc 0.19 and Canopus Procoder2, all with their highest quality settings, namely 1+3 passes for cce, best (slowest) quality for Hcenc, and mastering quality for Procoder.

I 've used one of my favorite movies,'Mulholland Drive' for DVD-5 conversion. That because it has a lot of demanding scenes with fast motion, and because of its relatively big length (2h and 20 mins).

So, there was a reduction level of 52,4%, with "steal space from extras 50%" enabled. Overall bitrate was 3.140kbps.

My conclusions are that all three encoders do a substantially good job in retaining a good part of the initial information, although with a different manner.
To my eyes, cce tends to retain as much initial information as possible, but at the expense of sometimes introducing significant noise, even in noise-free parts, which is something unacceptable in my humble opinion.
Canopus procoder tries a completely different approach. It often smears (or smudges, i don't know which is the appropriate adjective here) some of the detail and produces a softer looking image (appealing to the eye, I must admit). It even eliminates noise where present in the original picture, as if it were using some kind of filter. This is a good thing, but the macroblocking artefacts are a little disturbing, particularly in high-speed motion and in still background.
Lastly, Hcenc is beyond belief in most cases. I mean this is a professional looking encoder judging from the output. It's a free encoder and yet it surpasses its pricey competitors in some cases.

All I'm saying is that I tend to agree with UncasMS's findings and thus I'm rather in favor of Procoder, hcenc comes very close and cce comes last. I usually don't have much time to do such tests as my job is very demanding (I'm a doctor) but being a great fun of movies urged me to do it. So please don't ask me for further testing using other encoders like quenc and qmat. I think that the ones tested are the cream of the encoders.


I also have to mention that cce doesn't do a good job in low bitrates. The extras (reduced by 50%) were hideous with cce.
Just hideous. I got too many blocks in some occasions, as if i were transcoding instead of re-encoding.
I mean it really screwed up the trailer. I couldn't believe my eyes. Needless to say that procoder shines here.


I've made several caps of the same frames for comparison reasons, but I'm kindly asking for the moderator's permission to post them, because they will overload the page. Would it be OK if I posted them on imageshack or imagevenue and instead of posting the full caps here I should only post the thumbnails? I followed UncasMS's method of naming the screencaps, framenumber_frame_encoder.jpg
The only problem when it comes to free hosting sites is that they rename the uploaded pics.

Thanks in advance.

Last edited by Hlektron; 29 Dec 2006 at 11:34 AM Reason: omission
Hlektron is offline   Reply With Quote
Old 29 Dec 2006, 03:35 PM   #2
Super Member
Super Member
 
Join Date: Nov 2005
Location: Region 4
Posts: 217
Default

You can make the pictures into thumbnails. Any free picture site will do. You can just put what each frame is underneath the picture.
linx05 is offline   Reply With Quote
Old 29 Dec 2006, 03:41 PM   #3
Super Moderator
 
anonymez's Avatar
 
Join Date: Mar 2004
Posts: 5,525
Default

Quote:
I've made several caps of the same frames for comparison reasons, but I'm kindly asking for the moderator's permission to post them, because they will overload the page.
no problem

Quote:
posting the full caps here I should only post the thumbnails?
thumbnails, linking to the full image, will be preferable. that way folks on dialup can at least follow the discussion
__________________
"What were the things in Gremlins called?" - Karl Pilkington

anonymez is offline   Reply With Quote
Old 29 Dec 2006, 11:36 PM   #4
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default

OK, enough with the talking already. Here are the caps.

Please, remember to exclude the first five digits of the filename (those generated by imagevenue) while saving, so they can be sorted with their frame number. You may also delete the numbers following the encoder version, i.e pro, cce, hc.
For example, "92884_7329_b_cce_4passes_122_344lo.jpg" should look like "7329_b_cce_4passes.jpg"


























Hlektron is offline   Reply With Quote
Old 29 Dec 2006, 11:37 PM   #5
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default

Some more:











Hlektron is offline   Reply With Quote
Old 29 Dec 2006, 11:38 PM   #6
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default

And now, onto the "Extras" (the trailer, compressed by 50%).
















Could anyone explain why the color-shift in the re-encodes? The reds are orange-ish and generally there is a slight yellow cast in all the re-encoded videos. Does it have to do with the colorspace? How can I get a more accurate output color-wise?

Last edited by Hlektron; 31 Dec 2006 at 01:56 AM Reason: Correction
Hlektron is offline   Reply With Quote
Old 31 Dec 2006, 06:51 AM   #7
Digital Video Enthusiast
Digital Video Enthusiast
 
Join Date: Sep 2004
Posts: 324
Default

I can't help feeling that something is out of kilter... I've never seen CCE do as poorly in comparison to other encoders as I see in that last set of pictures. Not a criticism -- it just doesn't match with the testing I've done. Maybe something unique? There's also the issue of using JPG for the comparison... JPG will compress on top of the existing compression and can add its own artifacts. I did look, though, and the JPGs are the same size.

All in all, though, an interesting comparison.

Last edited by jdobbs; 31 Dec 2006 at 06:58 AM
jdobbs is offline   Reply With Quote
Old 31 Dec 2006, 07:31 AM   #8
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default

I also have the feeling that something went bad during the cce encoding, because the result was totally unacceptable. I tried selecting 25% for extras, and the results were far better, really good as a matter of fact.
Did it have to do with the number of passes? Maybe a bit of "over-doing"? Dunno. maybe I'll give it a shot with 1 pass and see if the result is better.
As for the vidcaps, they were initially saved in tiff format and then converted to jpgs with photoshop in their highest jpg quality (12).

As for the color shift, any ideas?

Is one pass encoding the same as OPV?
Hlektron is offline   Reply With Quote
Old 31 Dec 2006, 11:18 AM   #9
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default

Just out of curiosity, I'm re-doing the movie with an older version of cce, 2.67.00.23. I'll post my findings as soon as it's finished. I've chosen exactly the same options, 4 passes and 50% steal space from extras. Let's see... I can't wait to see how the extras are going to look this time.
Hlektron is offline   Reply With Quote
Old 31 Dec 2006, 11:32 AM   #10
Digital Video Enthusiast
Digital Video Enthusiast
 
Join Date: Sep 2004
Posts: 324
Default

You also did 50% "Steal Space" with the other encoders? If so, it's pretty impressive how well ProCoder and HC did. 50% kinda pushes the envelope -- because a lot of extras are encoded at relatively low bitrates to begin with.
jdobbs is offline   Reply With Quote
Old 31 Dec 2006, 12:10 PM   #11
Not a god of digital video
 
blutach's Avatar
 
Join Date: Oct 2004
Location: is everything!
Posts: 24,627
Default

I've always thought that Procoder shines on low BR material while CCE is good for the better source and struggles on low BR stuff.

HC is amazing - now that's what you call value, eh? Huge thanks to Hank.

Happy New Year to all DVD Rebuilder'ers!

Regards
__________________
Les

Essential progs - [PgcEdit] [VobBlanker] [MenuShrink] [IfoEdit] [Muxman] [DVD Remake Pro] [DVD Rebuilder] [BeSweet] [Media Player Classic] [DVDSubEdit] [ImgBurn]

Media and Burning - [Golden Rules of Burning] [Media quality] [Fix your DMA] [Update your Firmware] [What's my Media ID Code?] [How to test your disc]
[What's bitsetting?] [Burn dual layer disks safely] [Why not to burn with Ner0] [Interpret Ner0's burn errors] [Got bad playback?] [Burner/Media compatibility]

Cool Techniques - [2COOL's guides] [Clean your DVD] [Join a flipper] [Split into 2 DVDs] [Save heaps of Mb] [How to mock strip] [Cool Insert Clips]

Real useful info - [FAQ INDEX] [Compression explained] [Logical Remapping of Enabled Streams] [DVD-Replica] [Fantastic info on DVDs]


You should only use genuine Verbatim or Taiyo Yuden media. Many thanks to www.pcx.com.au for their supply and great service.

Explore the sites and the programs - there's a gold mine of information in them

Don't forget to play the Digital Digest Quiz!!! (Click here)

blutach is offline   Reply With Quote
Old 31 Dec 2006, 12:39 PM   #12
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default

Quote:
You also did 50% "Steal Space" with the other encoders?
Yes, of course. The same settings were applied to all the encoding procedures, that was the main concept, to have a reference point. The extras were interlaced, so it didn't quite amaze me that procoder did so well. What did amaze me, instead, was the bad output I got from cce which is generally considered an exceptional encoder (and in fact, it is).
Was it due to the interlaced source in association with the "push to their limits" approach (50%) I followed? Because when I selected 25% instead, the extras were considerably better. Or should I say, they was simply no comparison between what I got from 50% and 25%.

Here are the caps from the extras "stolen" by 25%:


Hlektron is offline   Reply With Quote
Old 31 Dec 2006, 12:46 PM   #13
Junior Member
Junior Member
 
Join Date: Dec 2006
Location: Ganymede
Posts: 26
Default

Quote:
HC is amazing - now that's what you call value, eh? Huge thanks to Hank.
I couldn't agree more with that statement. I'm very, very impressed by Hcenc. Unbelievably good value for "money". It is money. Well done, Hank.
To be honest, I still haven't settled down to a particular encoder. All three of them are excellent, but I think Procoder has a slight edge, just due to its performance in low bitrates. But I have to admit that cce is the sharpest of all.
Hlektron is offline   Reply With Quote
Old 31 Dec 2006, 04:02 PM   #14
just farted
 
Join Date: Aug 2004
Location: dumpster behind mcdonalds
Posts: 314
Default

Quote:
Originally Posted by jdobbs View Post
I can't help feeling that something is out of kilter... I've never seen CCE do as poorly in comparison to other encoders as I see in that last set of pictures. Not a criticism -- it just doesn't match with the testing I've done. Maybe something unique? There's also the issue of using JPG for the comparison... JPG will compress on top of the existing compression and can add its own artifacts. I did look, though, and the JPGs are the same size.

All in all, though, an interesting comparison.
woah... i agree... something must have went wrong with that CCE encode. i've never seen CCE do such a bad job before.

Last edited by rahzel; 31 Dec 2006 at 04:04 PM
rahzel is offline   Reply With Quote
Old 31 Dec 2006, 07:49 PM   #15
Junior Member
Junior Member
 
Join Date: Jul 2006
Posts: 12
Default

Quote:
Originally Posted by jdobbs View Post
I can't help feeling that something is out of kilter... I've never seen CCE do as poorly in comparison to other encoders as I see in that last set of pictures. Not a criticism -- it just doesn't match with the testing I've done.
I have seen very similar bad macroblocking as in picture <95493_Extras_11434_p_cce_4passes_122_321lo.jpg> in high motion scenes when CCE was forced to encode at an average bit rate of less than 1400 kbps. I haven't done similar tests with other encoders, though.

@Hlektron:
Do you know the avearge bit rate of your "50% steal space" Extras? Did it go below 1500 kbps?
Sharc is offline   Reply With Quote
Reply

Bookmarks


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Error: "No Disc is in drive" "The Device is not ready"?? Orangeaid186 General Computing 11 6 Dec 2005 08:31 AM



All times are GMT +10. The time now is 10:14 AM.

Kirsch designed by Andrew & Austin


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.
Copyright © 1999 - 2018 Digital Digest

Visit DivXLand   Visit dvdloc8.com