CQ Value selected not right

[Fishman0919]

I just tried to backup Halloween R1 Movie Only
and I noticed the CQ value was not right.
I'm using HC for the Redis

[12:11:41] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.8.0
- CmdLine (.BAT) encoder selected.
- "Steal Space from Extras" mode (33%) is enabled.
- "Movie Only" mode is enabled.
- Source: HALLOWEEN
- VTS_01: 3,022,964 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 174,768 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 74.6%
- Overall Bitrate : 5,690/4,552Kbs
- Space for Video : 4,050,166KB
- Analyzing VTS_01 for optimal CQ factor.
-- TargetSize (sectors):2,055,458
-- Sampling 1752 of 174768 frames.
-- Predicted size (sectors) at CQ=2.2: 2,966,882
-- Predicted size (sectors) at CQ=3.1: 1,512,754
- CQ Value selected: 3.1
- Redistributing using Final_Q: 31

[jdobbs]

A final CQ of 3.1 is equal to a Final_Q of 31.

[Fishman0919]

Yes but what I am referring too is...

-- TargetSize (sectors):2,055,458
-- Sampling 1752 of 174768 frames.
-- Predicted size (sectors) at CQ=2.2: 2,966,882
-- Predicted size (sectors) at CQ=3.1: 1,512,754

Sorry I wasn't most specific before.

[jdobbs]

I guess I should invest in a new set of eyes. Ok, I see it now. Hmmm... not sure how that could happen. I'll take a look at it. Was this the first time through the file?

[Fishman0919]

Yes, first run thru.
I del the D2VAVS folder and rebuilder.log then rerun the movie... Same

I just re-ran a 3rd time...

-----------------
[10:18:57] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.8.0
- CmdLine (.BAT) encoder selected.
- "Steal Space from Extras" mode (33%) is enabled.
- "Movie Only" mode is enabled.
- Source: HALLOWEEN
- VTS_01: 3,022,964 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 174,768 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 74.6%
- Overall Bitrate : 5,690/4,552Kbs
- Space for Video : 4,050,166KB
- Analyzing VTS_01 for optimal CQ factor.
-- TargetSize (sectors):2,055,458
-- Sampling 1752 of 174768 frames.
-- Predicted size (sectors) at CQ=2.2: 2,966,997
-- Predicted size (sectors) at CQ=3.1: 1,512,275
- CQ Value selected: 3.1
- Redistributing using Final_Q: 31
- HIGH/LOW/TYPICAL Bitrates: 6,087/1,052/4,552 Kbs

I also ran it with CCE Basic...

-----------------
[10:27:41] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.8.0
- CCE 2.70.1.16 encoder selected.
- "Adaptive Quantizer Matrices" is enabled.
- "Steal Space from Extras" mode (33%) is enabled.
- "Movie Only" mode is enabled.
- Source: HALLOWEEN
- VTS_01: 3,022,964 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 174,768 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 74.3%
- Overall Bitrate : 5,666/4,532Kbs
- Space for Video : 4,032,966KB
- Analyzing VTS_01 for optimal CQ factor.
-- TargetSize (sectors):2,046,729
-- Sampling 1752 of 174768 frames.
-- Predicted size (sectors) at CQ=2.2: 2,966,997
-- Predicted size (sectors) at CQ=3.1: 1,512,754
- CQ Value selected: 3.1
- Redistributing using Final_Q: 31
- HIGH/LOW/TYPICAL Bitrates: 6,552/1,047/4,532 Kbs

[Fishman0919]

Tried again with avisynth 2.57...

-----------------
[10:42:51] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.7.0
- CmdLine (.BAT) encoder selected.
- "Steal Space from Extras" mode (33%) is enabled.
- "Movie Only" mode is enabled.
- Source: HALLOWEEN
- VTS_01: 3,022,964 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 174,768 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 74.6%
- Overall Bitrate : 5,690/4,552Kbs
- Space for Video : 4,050,166KB
- Analyzing VTS_01 for optimal CQ factor.
-- TargetSize (sectors):2,055,458
-- Sampling 1752 of 174768 frames.
-- Predicted size (sectors) at CQ=2.2: 2,967,497
-- Predicted size (sectors) at CQ=3.1: 1,513,118
- CQ Value selected: 3.1
- Redistributing using Final_Q: 31
- HIGH/LOW/TYPICAL Bitrates: 5,090/1,052/4,552 Kbs

[Sharc]

You may want to try again using RB-Opt for redistribution and see if it makes a difference.

[Fishman0919]

Sure, thanks. That may be a temp fix for me... but DVD-RB still needs to be looked at.... the problem I get with RB-Opt is that it doesn't extract the files that don't need the be re-encoded like DVD-RB pro.


The encoding with DVD-RB goes thru fine... just the Final_Q: is not right (or what it should be).

[Fishman0919]

I ran into another problem with Q acting funny

It's the SIMPSONS_WS R1... HC for Redis... CCE Smooth Matrix

Should DVD-RB see that the prediction is going the another way and go with the highest value

- Reduction Level for DVD-5: 60.6%
- Overall Bitrate : 4,566/3,653Kbs
- Space for Video : 3,444,652KB
- Analyzing VTS_02 for optimal CQ factor.
-- TargetSize (sectors):1,391,238
-- Sampling 1572 of 156091 frames.
-- Predicted size (sectors) at CQ=2.9: 1,308,451
-- Predicted size (sectors) at CQ=2.7: 1,362,159
-- Predicted size (sectors) at CQ=2.6: 1,342,375
-- Predicted size (sectors) at CQ=2.5: 1,302,291
-- Predicted size (sectors) at CQ=2.3: 1,196,052
-- Predicted size (sectors) at CQ=1.9: 1,163,943
-- Predicted size (sectors) at CQ=1.5: 1,114,694
-- Predicted size (sectors) at CQ=1.2: 1,157,175
-- Predicted size (sectors) at CQ=0.9: 1,212,367
-- Predicted size (sectors) at CQ=0.7: 1,212,367
-- Predicted size (sectors) at CQ=0.6: 1,217,688
-- Predicted size (sectors) at CQ=0.5: 1,212,367
-- Predicted size (sectors) at CQ=0.4: 1,212,367
-- Predicted size (sectors) at CQ=0.3: 1,212,367
-- Predicted size (sectors) at CQ=0.2: 1,217,688
-- Predicted size (sectors) at CQ=0.1: 1,212,367
- CQ Value selected: 0.1
- Redistributing using Final_Q: 1

[jdobbs]

Well something else is definitely wrong... because a lower CQ value should always result in a higher file size... no wonder the software is confused. How on earch could an encode with a CQ of 2.6 possibly be smaller than the one for 2.7? If it were the same I'd say it might be related to saturation... but smaller???

But... that wasn't the case on the first post.

I know you are using HC for the redist... but what encoder are you using -- FFMPEG?

[Fishman0919]

[08:12:50] One Click encoding activated...
-----------------
[08:12:51] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.7.0
- QuEnc 0.7.2.0 encoder selected.
- "Steal Space from Extras" mode (25%) is enabled.
- Source: SIMPSONS_WS
- VTSM_02: 117,979 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 8,541 frames.
-- Building .AVS and .ECL files
- VTS_02: 2,915,584 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 147,550 frames.
-- Building .AVS and .ECL files
- VTS_03: 61,636 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 3,317 frames.
-- Building .AVS and .ECL files
- VTS_04: 61,166 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 2,967 frames.
-- Building .AVS and .ECL files
- VTS_06: 202,843 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 10,170 frames.
-- Building .AVS and .ECL files
- VTS_08: 279,529 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 12,660 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 60.6%
- Overall Bitrate : 4,566/3,653Kbs
- Space for Video : 3,444,652KB
- Analyzing VTS_02 for optimal CQ factor.
-- TargetSize (sectors):1,391,238
-- Sampling 1572 of 156091 frames.
-- Predicted size (sectors) at CQ=2.9: 1,308,451
-- Predicted size (sectors) at CQ=2.7: 1,362,159
-- Predicted size (sectors) at CQ=2.6: 1,342,375
-- Predicted size (sectors) at CQ=2.5: 1,302,291
-- Predicted size (sectors) at CQ=2.3: 1,196,052
-- Predicted size (sectors) at CQ=1.9: 1,163,943
-- Predicted size (sectors) at CQ=1.5: 1,114,694
-- Predicted size (sectors) at CQ=1.2: 1,157,175
-- Predicted size (sectors) at CQ=0.9: 1,212,367
-- Predicted size (sectors) at CQ=0.7: 1,212,367
-- Predicted size (sectors) at CQ=0.6: 1,217,688
-- Predicted size (sectors) at CQ=0.5: 1,212,367
-- Predicted size (sectors) at CQ=0.4: 1,212,367
-- Predicted size (sectors) at CQ=0.3: 1,212,367
-- Predicted size (sectors) at CQ=0.2: 1,217,688
-- Predicted size (sectors) at CQ=0.1: 1,212,367
- CQ Value selected: 0.1
- Redistributing using Final_Q: 1
- Movie improvement from extra reduction = 6.5%
- HIGH/LOW/TYPICAL Bitrates: 8,808/400/3,653 Kbs
[09:07:02] Phase I, PREPARATION completed in 55 minutes.

[Options]
Mode=2
OneClick=0
ReduceOpt=2
NoWarn=1
AdditionalOutput=1
LogFile=1
CCEAQM=1
QuEncTargetSectors=2245000
ProCoderTargetSectors=2250000
REDIST_USE_FINAL_Q=1
Redist_Percent=25
QuEncHQ=1
EncoderMinimized=0
RemoveDTS=1
HC_Quality=3
HC_Matrix=1
SkinVersion=12
AudioDub=1
QuEncodeType=0
iDCT=6
GOP=0
DCPrec=2
MainMatrix=Smooth (CG)
LowMatrix=Same as Main Feature
VLowMatrix=Same as Main Feature
ExtraMatrix=QuEnc Lower Bitrate (QLB)
ProCoder_Quality=4
DVD_Label=SIMPSONS_WS
DVD_Name=SIMPSONS_WS.ISO
MovieOnly=0
HalfD1=000000
Convert_16_9=000000
DisableInterlace=000000
ConvertToYUY2=1
Skin=Rockas Original
Redistribute=1
HC_FOR_REDIST=1
HC_LUMGAIN=3
HC_VBR_Bias=25
Encode_Menus=1
RemoveDC=1
Completed=32
QuEncScene=1
Encode_Processes=1
[Paths]
QuEnc=C:\Program Files\DVD-RB PRO\Encoders\QuEnc\QuEnc.exe
ReJig=C:\Program Files\DVD-RB PRO\Encoders\ReJig\ReJig.exe
HC=C:\HC022\HCenc_022.exe
FFMPEG=C:\Program Files\DVD-RB PRO\Encoders\FFMpeg\ffmpeg.exe
CMDLINE=C:\Program Files\DVD-RB PRO\Encoders\CMDLINE\mencoder.bat
ProCoder=C:\Program Files\DVD-RB PRO\Encoders\EclPro\EclPro.exe
DECOMB=C:\Program Files\AviSynth 2.5\plugins\Decomb521.dll
NICAUDIO=C:\Program Files\AviSynth 2.5\plugins\nicaudio.dll
MPEG2DEC=C:\Program Files\DVD-RB PRO\DGDecode.dll
Working=C:\DVD-RB Temp\
Output=D:\Done DVD's\
Source=C:\FULLDISC\SIMPSONS_WS\VIDEO_TS\
[Audio]
Selected=1100001111
Remapping=
[Subpictures]
Selected=1000
[CCEOptions]
VBR_bias=10
Quality_prec=16
eclPasses=2

HC is doing both the Prediction and Redis phase

[Sharc]

Does RB-Opt converge properly in these cases?
........the problem I get with RB-Opt is that it doesn't extract the files that don't need the be re-encoded like DVD-RB pro.

You can - after the redistribution with RB-Opt - just go into the editor of DVD-RB and tick "No reencode" for all segments which indicate 100% - and "Save and Exit".

[Fishman0919]

Yes but me=lazy...sometimes. Would like it to be a one click.

But the fact of me doing that and saying nothing to jdobbs doesn't help jdobbs or DVD-RB

[Sharc]

Well, just wondering. I never run into a similar problem with DVD-RB, but in few cases the number of iteration steps appeared to be rather large (10 cycles or so).

Added: Which matrix did you use in the HALLOWEEN case? Perhaps the problem is matrix/encoder sensitive.

[jdobbs]

I'm running "The Simpsons" R1 right now with HC set to do the prediction. Any special settings? It might help if you mail me your REBUILDER.INI file (dvd-rb@dvd-rb.com).

[jdobbs]

It's acting really odd. The Q values aren't matching what is being passed to the encoder. Are you using HC v0.22? Try going back to v0.21 and see if the issue goes away... I'm about to do the same.

[Fishman0919]

It's acting really odd. The Q values aren't matching what is being passed to the encoder. Are you using HC v0.22?

Yes... I posted my rebuilder.log and REBUILDER.INI in the post above.

I also emailed my .ini file

It's acting really odd. The Q values aren't matching what is being passed to the encoder. Are you using HC v0.22?

Are you are talking about HC.22 Q values bouncing all round?

Most likely is encoder saturation being that it's a Anime/CGI but the CCE Smooth Matrix should make that hard.

[jdobbs]

I think I see what it is... the MAXBITRATE value is very small and it is hitting it too easily. Interestingly it also seems to be getting smaller with each pass, too... there has to be a bug on my end.

HC is only doing what it is told to do.

[Fishman0919]

I think I see what it is... the MAXBITRATE value is very small and it is hitting it too easily. Interestingly it also seems to be getting smaller with each pass, too... there has to be a bug on my end.

HC is only doing what it is told to do.


ahh... that'll do it

[jdobbs]

I fixed it for the next release. It was an odd little glitch -- interesting that it had never shown itself before, as that code has been there for a long time.

[Fishman0919]

Glad I could help :)

[jdobbs]

The fix for the second issue is in v1.26.6... but I still need to look at the first one. I'll need to pick up a copy of HALLOWEEN.

Which version are you trying to back up? Is it:

1. The Unrated Director's Cut (Widescreen 2 disc)
2. Two Disc Special Edition
3. 25th Anniversary Edition
4. One of the other 6000 versions. :)

[Fishman0919]

Unrated Director's Cut (Widescreen Two-Disc Special Edition) (2007)

Link (http://www.amazon.com/Halloween-Unrated-Directors-Widescreen-Two-Disc/dp/B000VKL6Z2)

[jdobbs]

Ok. I'll see if I can pick it up tonight.

[jdobbs]

@Fishman0919

I'm having a hard time repeating your problem. When I run "Halloween", DVD-RB uses 2.3 as the first value -- and selects it.[21:27:49] Phase I, PREPARATION started.
- DVD-RB v1.26.5
- AVISYNTH 2.5.7.0
- CCE 2.70.2.9 encoder selected.
- AVS Filters are enabled.
- "Movie Only" mode is enabled.
- Source: HALLOWEEN
- VTS_01: 3,022,964 sectors.
-- Scanning and writing .D2V & .AVS files
-- Processed 174,768 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 71.0%
- Overall Bitrate : 5,414/4,332Kbs
- Space for Video : 3,854,274KB
- Analyzing VTS_01 for optimal CQ factor.
-- TargetSize (sectors):1,956,043
-- Sampling 1752 of 174768 frames.
-- Predicted size (sectors) at CQ=2.3: 1,952,406
- CQ Value selected: 2.3
- Redistributing using Final_Q: 23
- HIGH/LOW/TYPICAL Bitrates: 6,128/1,001/4,332 Kbs
[21:32:32] Phase I, PREPARATION completed in 5 minutes.

Could you see if you can repeat it, and then send me your INI file? As you can see, the source frames and size is the same... but I'm getting a different reduction level for DVD-5. You have "Extra reduction 33%" on -- but that has no effect on "Movie Only" mode (I tested it just to be sure).

Added: The only way I can think of this happening (my 71% reduction and your 74.6% reduction) is if you had provided a non-standard TargetSectors value or were removing less audio. Is that the case?

Additional added: How long did it take for your Phase I to complete? I forced a CQ of 2.2 to test it -- and it seems to "hang" during encode.

[Fishman0919]

I'm away from home now and flying back Monday night.
Once I get home I'll re-run it and post my results

Edit: I kept only the DD 5.1 and maybe subpicture 1

- Reduction Level for DVD-5: 71.0%
- Overall Bitrate : 5,414/4,332Kbs
- Space for Video : 3,854,274KB
and
- Reduction Level for DVD-5: 74.6%
- Overall Bitrate : 5,690/4,552Kbs
- Space for Video : 4,050,166KB

are a little diff.

Could you see if you can repeat it, and then send me your INI file?

the INI would be the same as the last one I email to you

I used Mencoder and CCE Basic and I don't have any sizing changes for them

[Sharc]

Just as a note:
For repeating the problem and for comparison one should also agree on the matrix that is used for redistribution because it will influence the search for the Final_Q.

[jdobbs]

@Fishman0919

It was an elusive little bug -- but I got it. It's fixed for the next version.

[Fishman0919]

Great :)