Constantine problem with AutoQMatEnc

[m1ckran]

I've started a new thread for this because I doubt whether it's an inherent problem for DVD-RB & AutoQMatEnc.

I did a movie-only (plus English subtitles) rip of Constantine PALUK, and encoded with DVD-RB using AutoQMatEnc set at three passes with QMatOp. I also tried it with HCEnc. Everything seemed to work well will both encoders so I burned both versions to DVD+RW using ImgBurn.

The first thing I noticed was that the AQE version was undersized compared with the HC version. On running the AQE dvd on my Pioneer player, I found that playback slowed to a crawl (about one frame per 20 secs) at the part where Constantine cut his rists. It later speeded up but slowed again on the night-time roof scene near the end, and stayed that way until the DVD finished. Yes, it did finish - several hours later!

The HC version played normally.

I used the full download release of AQE (not betas) and HC 18.

I've never heard anyone complain about AQE in this way so I'm inclined to suspect that I've done something silly. Unfortunately, I deleted the log file but I could rerun it again. Any ideas?

[rack04]

I have experienced this same problem with "Hoodwinked". I was using AutoQMatEnc v0.34.0.4 + QMATOpt with DVD Rebuilder v1.09.3. At times the movie would slow to still. I've tried 3 different movies with the same result.

[jdobbs]

What kind of bitrates were you getting? I'm wondering because the description has a "rate control" feel to it. Any chance you might be able to run Bitrate Viewer against the M2V for the section that is showing the problem?

[m1ckran]

Oh bugger! I was hoping it was my fault.

I don't normally like reporting my problems because I can usually find the answer somewhere on this site, and I hope the software authors realise I'm not having a go at them because I really appreciate their work.

Perhaps we both did something wrong?

[m1ckran]

@jdobbs
Sorry, I was writing my post when you did yours so I've only just read your question.

I'll run it when I get home tonight.

Thanks.

[rack04]

What kind of bitrates were you getting? I'm wondering because the description has a "rate control" feel to it. Any chance you might be able to run Bitrate Viewer against the M2V for the section that is showing the problem?

How do I associate the "problem" area on the final DVD to the M2V for that section? When I left this morning I started "Munich" using AutoQMatEnc. I can check the bitrate when I go home for lunch.

[jdobbs]

The only way to associate it is to play back the M2Vs with Media Player and find the one that has that scene in it.

[rack04]

What kind of bitrates were you getting? I'm wondering because the description has a "rate control" feel to it. Any chance you might be able to run Bitrate Viewer against the M2V for the section that is showing the problem?

Here is the output from Bitrate Viewer of a sample M2V for my latest encode of "Munic". Using Media Player Classic the M2V play flawlessly but viewing the DVD the video starts and stops randomly. Let me know what other information I can provide.

http://i11.photobucket.com/albums/a199/rack04/bitrate.jpg

[dirio49]

is this the scene/cell that you saw that gave you the problem?

[m1ckran]

@jdobbs

Sorry for the delay but I think I have the info you asked for.

Extract from DVD-RB log:

[13:56:45] Phase I, PREPARATION started.
- AutoQMatEnc selected.
- VTS_01: 2,537,680 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 173,728 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 87.2%
- Overall Bitrate : 4,852Kbs
- Space for Video : 4,116,184KB
- HIGH/LOW/TYPICAL Bitrates: 5,438/3,935/4,852 Kbs
[13:58:53] Phase I, PREPARATION completed in 2 minutes.

Extract from AVS log at the suspect scene:

Adjusted Min bitrate: 2875301
Adjusted Min bitrate: 2819775
Tolerance : 1478466
Final Adjusted Min bitrate: 1273250
Processing VAF file
Count Q1 : 1620 , Count Q2 : 843 , Average Q : 2.15426
Q[1]=1620 Q[2]=843 Q[3]=1646 Q[4]=0 Q[5]=0 Q[6]=0 Q[7]=0 Q[8]=0 Q[9]=0 Q[10]=0 Q[11]=0 Q[12]=0 Q[13]=0 Q[14]=0 Q[15]=0 Q[16]=0 Q[17]=0 Q[18]=0 Q[19]=0 Q[20]=0 Q[21]=0 Q[22]=0 Q[23]=0 Q[24]=0 Q[25]=0 Q[26]=0 Q[27]=0 Q[28]=0 Q[29]=0 Q[30]=0 Q[31]=0 Q[32]=0 Q[33]=0 Q[34]=0 Q[35]=0 Q[36]=0 Q[37]=0 Q[38]=0 Q[39]=0 Q[40]=0 Q[41]=0 Q[42]=0 Q[43]=0 Q[44]=0 Q[45]=0 Q[46]=0 Q[47]=0 Q[48]=0 Q[49]=0 Q[50]=0 Q[51]=0 Q[52]=0 Q[53]=0 Q[54]=0 Q[55]=0 Q[56]=0 Q[57]=0 Q[58]=0 Q[59]=0 Q[60]=0 Q[61]=0 Q[62]=0 Q[63]=0 Q[64]=0
maxQpercent : 40.0584 ; partialQsum : 1646
Initial Average : 2.000000
Calculated Min Qmin : 1 ; Max Qmax : 3
Expected size : 104635685 ; Real Size : 129228238 ; Error : -24 %
Quality Factor correction : 0.809697
Corrected Qmin : 2 ; Corrected Qmax : 5
New Tolerance : 1478466
Done
Final Result :
Expected size : 104635685 ; Real Size : 100294925 ; Error : 4.1 %
Total time elapsed : 469 seconds

Incidentally, the DVD produced from this run was still misbehaving. I loaded DVD-RB's output into DVDShrink and it played the scene but it was very jerky.

Hope the logs help. I hesitate to post the full logs due to their size but I can do so if you wish, or I can PM them.

[rack04]

is this the scene/cell that you saw that gave you the problem?

All scene has this same "jerky" playback. What other information can I provide to better understand this problem?

[jdobbs]

Hmm... if someone can provide me with an NTSC title that does this consistently I'll pick it up and try it. The fact that it works with other encoders makes me wonder what AutoQMatEnc is doing differently. I'll have to defer to SAPSTAR to see what he might think. I tested several DVDs with AutoQMatEnc and didn't see any of these types of problems.

Are these problems only happening when the reduction level is minimal? I noticed the post above had a reduction of 87.2%. It still sounds like a rate control issue to me.

@Rack04
One thing I see in the Bitrate Viewer screen is that the problem area is identified as BFF -- that's possible but fairly rare, which makes it suspect. Are you using the freeware version of DVD-RB when encoding? I think AutoQMatEnc may possibly read the meaning of the CCE ECL file incorrectly. One of the confusing things about CCE is that different versions read the same flags differently. CCE SP versions before v2.70 could only output TFF, so they would shift the source up on line on a BFF source and make it TFF. So DVD-RB had to treat it as TFF when reading it back in and applying the original flags.

Is your playback showing quick forward/backward motion on objects that are moving?

Also, could someone who is experiencing this please try an encode without QMAtOp -- I'd like to make sure the changing matrices aren't associated with the playback issue.

[jdobbs]

One other thing I'd like to remind everyone. These versions of AutoQMatEnc you're using are all beta.... so bugs are expected.

[rack04]

Hmm... if someone can provide me with an NTSC title that does this consistently I'll pick it up and try it. The fact that it works with other encoders makes me wonder what AutoQMatEnc is doing differently. I'll have to defer to SAPSTAR to see what he might think. I tested several DVDs with AutoQMatEnc and didn't see any of these types of problems.

Are these problems only happening when the reduction level is minimal? I noticed the post above had a reduction of 87.2%. It still sounds like a rate control issue to me.

@Rack04
One thing I see in the Bitrate Viewer screen is that the problem area is identified as BFF -- that's possible but fairly rare, which makes it suspect. Are you using the freeware version of DVD-RB when encoding? I think AutoQMatEnc may possibly read the meaning of the CCE ECL file incorrectly. One of the confusing things about CCE is that different versions read the same flags differently. CCE SP versions before v2.70 could only output TFF, so they would shift the source up on line on a BFF source and make it TFF. So DVD-RB had to treat it as TFF when reading it back in and applying the original flags.

Is your playback showing quick forward/backward motion on objects that are moving?

Also, could someone who is experiencing this please try an encode without QMAtOp -- I'd like to make sure the changing matrices aren't associated with the playback issue.

jdobbs,

I am using DVD Rebuilder Pro v1.09.3. I will try to re-encode "Munich" today using AutoQMatEnc v0.32b without QMATOpt. Attached is my Rebuilder.inf

[m1ckran]

I'll rerun it now without QMatOp. Unfortunately, I can't describe any other results since this is my first time using AutoQMatEnc.

One other thing I'd like to remind everyone. These versions of AutoQMatEnc you're using are all beta.... so bugs are expected.
I still feel a little guilty raising issues after someone has worked so hard to produce the goods. On the other hand, I'm happy to make any contribution (however small) to assist in improvements.

Once this is sorted, the next problem is likely to be more dificult: which encoder to choose, out of QuEnc, HCEnc and AutoQMatEnc!!! :)

[jdobbs]

The main reason I added AutoQMatEnc support is because I think it has huge potential. The NASA DCTune algorithm has been proven to dramatically improve picture quality -- and I love the fact that it is being applied to MPEG video. The work that SAPSTAR has done is outstanding. Given some time to settle down, get optimized, and find the right settings I think it will help get the most out of low bitrate encodes.

I think you should still consider it "experimental" -- as is true of all betas. So for your archived DVDs you may want to consider one of the other encoders (CCE, HC, ProCoder, or QuEnc) right now. But keep an eye on it as it evolves.

[rack04]

The playback problem still exists with AutoQMatEnc v0.32b without QMATOpt. Any other ideas?

[jdobbs]

No other ideas right now. I'm encoding "Munich" in Movie-Only mode using AutoQMatEnc right now to see how it works on my system.

[rack04]

I have a feeling that it is something on my end because the two movies I've tried "Hoodwinked" and "Munich" both exhibit the same problem.

[Fishman0919]

@rack04,

update to the lastest beta, lots of bug fixes and a much better rate control

LINK (http://www.autoqmatenc.com/BetaTesters.html)

[jdobbs]

I have a feeling that it is something on my end because the two movies I've tried "Hoodwinked" and "Munich" both exhibit the same problem.I just finished "Munich" and it plays back fine using PowerDVD. I'll burn it and test it on my standalone. I encoded using AutoQMatEnc v0.33.0.4 (which is the latest beta -- even though the download link says v0.34.0.4 erroneously)

[rack04]

@rack04,

update to the lastest beta, lots of bug fixes and a much better rate control

LINK (http://www.autoqmatenc.com/BetaTesters.html)

I've tried the lastest beta but I had the same problems. I even installed the beta as advised in another thread. I can't seem to figure out what the problem is, though I have yet to actually burn the ISO and test in my standalone.

[Fishman0919]

What ver of Avisynth are you using?

What Decoder ver are you using?

[m1ckran]

As promised, I ran Constantine again with AutoQMatEnc but without QMatOp.

It completed and the DVD-RB log was identical to the previous run (with QMatOp), but the AVS log was different. I've included an excerpt of the AVS log from the same scene as before.

D:\DVD Rebuilder Work\D2VAVS\V01002800002011.avs: 720x576, 25000/1000 fps, 4109 frames
converting ? -> RGB24
Assigned new RGB clip
AVS File opened successfully
Ratio : 16/9
Dark Limit : 0.0795781
Darkmasking value : 0
Quantizer value : 4 (delta: -5)
Qmin : 1 ; Qmax : 4
Adjusted Min bitrate: 300000
Adjusted Min bitrate: 2875301
Adjusted Min bitrate: 2992954
Tolerance : 1478466
Final Adjusted Min bitrate: 1273250
Processing VAF file
Count Q1 : 4015 , Count Q2 : 80 , Average Q : 1.3713
Q[1]=4015 Q[2]=80 Q[3]=12 Q[4]=0 Q[5]=2 Q[6]=0 Q[7]=0 Q[8]=0 Q[9]=0 Q[10]=0 Q[11]=0 Q[12]=0 Q[13]=0 Q[14]=0 Q[15]=0 Q[16]=0 Q[17]=0 Q[18]=0 Q[19]=0 Q[20]=0 Q[21]=0 Q[22]=0 Q[23]=0 Q[24]=0 Q[25]=0 Q[26]=0 Q[27]=0 Q[28]=0 Q[29]=0 Q[30]=0 Q[31]=0 Q[32]=0 Q[33]=0 Q[34]=0 Q[35]=0 Q[36]=0 Q[37]=0 Q[38]=0 Q[39]=0 Q[40]=0 Q[41]=0 Q[42]=0 Q[43]=0 Q[44]=0 Q[45]=0 Q[46]=0 Q[47]=0 Q[48]=0 Q[49]=0 Q[50]=0 Q[51]=0 Q[52]=0 Q[53]=0 Q[54]=0 Q[55]=0 Q[56]=0 Q[57]=0 Q[58]=0 Q[59]=0 Q[60]=0 Q[61]=0 Q[62]=0 Q[63]=0 Q[64]=0
maxQpercent : 0.340716 ; partialQsum : 14
Initial Average : 2.000000
Current Average : 1.500000
Calculated Min Qmin : 1 ; Max Qmax : 2
Expected size : 104635685 ; Real Size : 90394959 ; Error : 14 %
Quality Factor correction : 1.15754
Corrected Qmin : 1 ; Corrected Qmax : 2
New Tolerance : 1478466
Done
Final Result :
Expected size : 104635685 ; Real Size : 90394959 ; Error : 14 %
Total time elapsed : 405 seconds

I've not watched the full film from this run, but the dodgy scenes now play without stuttering of any kind.

Strangely, output without QMatOp was more undersized (3.86Gb) than with it (4.22Gb).

Don't forget that I'm using AutoQMatEnc version 0.32b.

I still have the full logs of both runs, so I could zip them up and send them if you wish but, in light of Fishman0919's post, perhaps the later betas have fixed this problem already (although rack04 still seems to have probs).

Please let me know if there is anything I can do to assist.

[danpos]

@m1ckran

Hi mate ! As a beta-tester from almost all AQE releases, I' suggest to you give a try to AQE 0.33 beta 4 coz it presents a huge improvement on rate control module, implying more quality and precise file size. If you don't mind, you can take a look at my results with AQE 0.33 beta 4 / RB-Free (http://www.vmesquita.com/forum/index.php?topic=6538.msg43935#msg43935) there.

Cheers,

[m1ckran]

@danpos

Thanks. I've already seen that post. On my LCD monitor, I can't tell any diffeence between the original and the re-encoded versions.

Don't get bigheaded, but that post is the reason I decided to try AutoQMatEnc in the first place!

I'm not sure whether to install the betas or to wait for jdobbs and sapster to finalise things in a new DVD-RB release. Sod it! I've just decided not to wait.

Thanks again.

[rack04]

What ver of Avisynth are you using?

What Decoder ver are you using?

I'm using AviSynth 2.5.6 and DGDecode Version 1.4.5

[Fishman0919]

I'm using AviSynth 2.5.6 and DGDecode Version 1.4.5

I've had some pb's with AviSynth 2.5.6... try AviSynth 2.5.5... thats the one that jdobbs includes with DVD-RB or dl it here (http://prdownloads.sourceforge.net/avisynth2/AviSynth_255.exe?download)

[SAPSTAR]

@rack04,

update to the lastest beta, lots of bug fixes and a much better rate control

LINK (http://www.autoqmatenc.com/BetaTesters.html)
Sorry for the delay...I was a bit off (real life :) ) But yep I would recommend the usage of the very last 0.33.0.4 with high bitrate encodes....as jdobbs said it sounds like a RC pb. I always had big issues with high bitrates (At the beginning, I optimized AQE for high compressions => low bitrates). The 0.33.0.4 solves most of the high bitrate issues....

[rack04]

Sorry for the delay...I was a bit off (real life :) ) But yep I would recommend the usage of the very last 0.33.0.4 with high bitrate encodes....as jdobbs said it sounds like a RC pb. I always had big issues with high bitrates (At the beginning, I optimized AQE for high compressions => low bitrates). The 0.33.0.4 solves most of the high bitrate issues....

What would cause a rate control problem? I was able to encode "Hoodwinked" with CCE SP, Procoder 2, HCenc, and QuEnc without the slightest glitch. I see how powerful AutoQMatEnc is and I like to diagnose this problem so I can see it at it's fullest. As suggested earlier I reencoded "Munich" with AviSynth 2.5.5 and still the same problem. Thanks to all those who are walking my through this.

[SAPSTAR]

What would cause a rate control problem? I was able to encode "Hoodwinked" with CCE SP, Procoder 2, HCenc, and QuEnc without the slightest glitch. I see how powerful AutoQMatEnc is and I like to diagnose this problem so I can see it at it's fullest. As suggested earlier I reencoded "Munich" with AviSynth 2.5.5 and still the same problem. Thanks to all those who are walking my through this.
The rate control is the mechanism in an encoder which allocates bits to each frame. Sometimes, if the RC doesn't strictly follow the constraints of max or min bitrates some peaky DVD players can have pbs to play the movie (stuttering, ...). AQE was designed for low bitrates (high compressions) so I must admit I worked on the high bitrates (>> 4000kbps) only lately. The last versions improved, but may still be "buggy". The 0.33.0.4 is very close to be "perfect" for and low and high bitrates...the 0.33.0.5 will be improved. There is always room for improvments :) (that's my moto)
EDIT: I will try this weekend "Hoodwinked" and see If I had any issue. My usual "config" for encoding movie in DVDRB is : Half-D1 for extras / 25% steal space + removal of every audio except English 5.1 and French 5.1 (if DTS I choose it instead of 5.1)

[m1ckran]

Thanks Sapster. I was beginning to think I should have been using the beta versions. Your post confirms this. I think I may have raised an issue that no longer exists, because I was using an older version of AutoQMatEnc.

Sorry. :o


PS How do you software authors find the time to develop software and reply to so many queries at Doom9? Are you all insomniacs?

[rack04]

Here is an update. I uninstalled DVD Rebuilder 1.09.3, AviSynth 2.5.6 and AutoQMatEnc 0.33b. Then I installed AutoQMatEnc 0.33.0.4 and DVD Rebuilder 1.09.3 using the AviSynth and DGDecode included in the install. After encoding "Munich" using AutoQMatEnc 0.33.0.4 and QMatOp the studdering still exists. I guess this narrows the problem down to PowerDVD since it seems I the only one experiencing this problem.

[SAPSTAR]

Here is an update. I uninstalled DVD Rebuilder 1.09.3, AviSynth 2.5.6 and AutoQMatEnc 0.33b. Then I installed AutoQMatEnc 0.33.0.4 and DVD Rebuilder 1.09.3 using the AviSynth and DGDecode included in the install. After encoding "Munich" using AutoQMatEnc 0.33.0.4 and QMatOp the studdering still exists. I guess this narrows the problem down to PowerDVD since it seems I the only one experiencing this problem.
I will try Munich with the latest version (not yet available) and see....Do you experience this stuttering with any other encoder on the same movie?!

[m1ckran]

@rack04

I found that stuttering was reproduced by DVD Shrink and my standalone. I'm trying again with the newer betas.

Have you tried burning to DVD and trying it on a standalone?

[SAPSTAR]

Thanks Sapster. I was beginning to think I should have been using the beta versions. Your post confirms this. I think I may have raised an issue that no longer exists, because I was using an older version of AutoQMatEnc.

Sorry. :o


PS How do you software authors find the time to develop software and reply to so many queries at Doom9? Are you all insomniacs?
No reason to be sorry...thanks to the numerous feedbacks I'm able to tweak AQE. Thank you for the feedback.
FYI: I did encode Hoodwinked this weekend...and given the differents conditions, some parts are definitely "high bitrates", which could have been a problem in the older version...but it worked flawlessly with the current version I'm working on...which should be available this week....

[m1ckran]

No reason to be sorry...thanks to the numerous feedbacks I'm able to tweak AQE. Thank you for the feedback
You're very kind.

the current version I'm working on...which should be available this week....
Can't wait.

[rack04]

I will try Munich with the latest version (not yet available) and see....Do you experience this stuttering with any other encoder on the same movie?!

No stuttering when encoded with Procoder 2.

[m1ckran]

Downloaded and installed the latest AutoQMatEnc betas. Ran Constantine again. Played Constantine. No problems.

I even did it again, using LanczosResize to reduce the top&bottom bars whilst maintaining the aspect ratio. I've done this before with a different encoder (don't remember which one) but I found that the picture was fuzzy compared with the original so I only did it once. In contrast, AutoQMatEnc produced a fabulous picture: crisp, sharp, detailed and somehow "tangible".

I also tried the same resize approach with National Treasure with total success. AutoQMatEnc was reporting this film as interlaced (which, I believe, is supposed to respond badly to LanczosResize) but I acheived an excellent backup.

Thank you all.

[danpos]

@m1ckran

I'm glad that you've got success with RB / AQE combo DVD backup.

Enjoy it!

Cheers,

[m1ckran]

I'm beginning to feel guilty and disloyal now because I still think highly of HcEnc and QuEnc.:o :confused:

[SAPSTAR]

Downloaded and installed the latest AutoQMatEnc betas. Ran Constantine again. Played Constantine. No problems.

I even did it again, using LanczosResize to reduce the top&bottom bars whilst maintaining the aspect ratio. I've done this before with a different encoder (don't remember which one) but I found that the picture was fuzzy compared with the original so I only did it once. In contrast, AutoQMatEnc produced a fabulous picture: crisp, sharp, detailed and somehow "tangible".

I also tried the same resize approach with National Treasure with total success. AutoQMatEnc was reporting this film as interlaced (which, I believe, is supposed to respond badly to LanczosResize) but I acheived an excellent backup.

Thank you all.

Thank you for the feedback! It's always interesting to have negative or good feedbacks.....I tend to prefer the last ones, I wonder why?! :D

[SAPSTAR]

I'm beginning to feel guilty and disloyal now because I still think highly of HcEnc and QuEnc.:o :confused:
Why guilty?! I also like HCEnc, I think it has the best rate control. No spikes...perfectly on target. That's an asset for an encoder. Even CCE SP isn't as good as HCEnc from that point of view.....Each encoder has its strength and weaknesses.....

[m1ckran]

@Sapstar
some parts are definitely "high bitrates", which could have been a problem in the older version...but it worked flawlessly with the current version I'm working on...which should be available this week....

This week is beginning to drag :D

(Bet you hate being chased :p )

[SAPSTAR]

@Sapstar


This week is beginning to drag :D

(Bet you hate being chased :p )
Yep...I know but I found another pb with high bitrates, and I prefer dealing with it before the next release...sorry for the delay given the current pb I'm facing, dont expect it before next week... :o
Sorry for that.....

[m1ckran]

No need to apologise. I'm just being a pest! ;)

Incidentally, I did notice some blocking with Contantine, in one of the hell scenes where everything is swirling about and nothing at all seems to be stationary. Tricky scenes for encoders, I imagine, but you might find them interesting.

By the way, I donated through PayPal but don't think I'm mean - I just didn't realise there are so many CAD to the pound. :D

All the best.

Edit: The blocking only happened after resizing so it may well be a lanczosresize issue.

[SAPSTAR]

...
Incidentally, I did notice some blocking with Contantine, in one of the hell scenes where everything is swirling about and nothing at all seems to be stationary. Tricky scenes for encoders, I imagine, but you might find them interesting.

By the way, I donated through PayPal but don't think I'm mean - I just didn't realise there are so many CAD to the pound. :D
...
Thank you very much for your donation!
About the blocking you saw in Constantine, it may be the same "bug", I'm working on. It comes from my damn High bitrates rate control.....I'm tweaking it. But I think I'm on the right track now..:)

[SAPSTAR]

I just released the 0.33.0.5, it should correct some pbs you met at high bitrates. I also improved the RC engine when QMatOp is off.

I recommend everybody using 0.33.0.4 to upgrade to .5.

I'm currently working on the audio encoding, so a .6 may be expected soon including the audio encoding plus any corrections to any pb reported to me with the 0.33.0.5.

[m1ckran]

Thanks.

I'll try it out on Constantine and let you know.

[SAPSTAR]

Thanks.

I'll try it out on Constantine and let you know.
Good...I wonder what are to be your results, because Dirio49 had a similar issue, and it's not solved yet. I'm not able to reproduce it yet.