DVD Rebuilder (CCE One-Click Beta v0.46) Comments and Suggestions

[wmansir]

Originally posted by lamster
OK, just checked, and all the commentaries were fine.
I just did this disc myself and everything seems to work fine (on my software players). At first I thought there was a problem with the 'Play All' feature of the select commentaries, but it turns out DVD Remake was the cause (the source files also had the same problem). I used titleset blanker and MenuEdit instead (no way I was keeping all those disney trailers) and it works perfectly.

[nimbles]

ummm... just a quick question regarding the updated mpeg2dec3dg.dll and decomb.dll that have been released in the decodefix1.3.0 and decomb5.2.1 respectively over the last week or so

Is it ok to update the aforementioned dll files into the avi synth pugin folder (will rb still work ok?)

[wmansir]

Originally posted by nimbles
ummm... just a quick question regarding the updated mpeg2dec3dg.dll and decomb.dll that have been released in the decodefix1.3.0 and decomb5.2.1 respectively over the last week or so

Is it ok to update the aforementioned dll files into the avi synth pugin folder (will rb still work ok?)

MPEG2Dec3dg.dll, no. Stick with 1.0 or 1.1.0. I added a warning in the install/setup guide last week.

Decomb, should work as long as the FieldDeinterlace command hasn't changed. Which it has not, AFAIK.

[Joergen]

ot: 10000+ DL's on v0.49. Congratulations jdobbs your tool is soon second to none!

[jdobbs]

Originally posted by Joergen
ot: 10000+ DL's on v0.49. Congratulations jdobbs your tool is soon second to none! Wow. I never noticed that. Cool.

[Joergen]

Wow 0.50 already! :cool:

jdobbs: About the 80% bitrate reduction for interlaced, does this apply to PAL aswell? Did the bitrate stutter happen with PAL?

I can see how the NTSC 23.9 out of 29.976 could need a 80% reduction but surely PAL 25/25 doesnt?

Thanks.

[rayvt]

Looks like 0.50 fixes the time glitch at 2:01:46 of this movie. Did you make a change to fix this, or did it just go away on its own? I didn't see anything in the changelog about it.

BTW, 0.49 was ok if I used rejig instead of CCE as the encoder.

[jdobbs]

Originally posted by Joergen
Wow 0.50 already! :cool:

jdobbs: About the 80% bitrate reduction for interlaced, does this apply to PAL aswell? Did the bitrate stutter happen with PAL?

I can see how the NTSC 23.9 out of 29.976 could need a 80% reduction but surely PAL 25/25 doesnt?

Thanks. Yes it does, and you're right, it shouldn't. I'll change that for 0.51.

[Joergen]

Ok, thanks :)

[Rombaldi]

@jdobbs

Kudos on where you've taken this tool in such a short time. Incredible.. I've been away for a while (0.37 was the last ver I
had tried).

Two requests (one old, one new).

1) (and I addressed this before, but you may have gotten busy) On the 4:3 LBX > 16:9 conversion, mentioned the need for an 'advanced' setting on tweaking the er.. 'window'. I've seen some LBX discs that may shift the actual LBX window as much as 15 pixels either direction. At the least it would be nice to be able to put in a +/- or up/down shift for the LBX window (it would REALLY be nice for a visual adjustment, ie. grab a frame from the VTS and display the window and up/down it for fine tuning... but that could get a little hairy... hmmm external tuning tool??). On a per VTS basis.

2) (something new) Would it be possible to an advanced AviSynth 'tweaking block'?? Basically just an interface to insert user defined AviSynth commands into each generated AVS file for whatever 'tweaking' the user wants. Example: I've gotten some discs that were done with a really nasty color balance on the transfers, that look fairly decent when you asjust the colors on playback, but would be nice to do the necessary color/contrast/brightness adjustments ONCE (when re-encoding). This would have to be an 'uber-advanced' setting, a 'Don't cry to me if it messes up the world, it's your problem not mine' setting...

[jdobbs]

@Rombaldi

It may be a while for #1.. but I've been thinking about doing #2..

[DDogg]

Before OPV, I was thinking allowing the insertion of an import line after the mpeg2source line was a simple first stab at it. You would have needed two import statements, one for progressive and one for interlaced.

Import("MyProgressiveFilters.avs") Inserted in all progressive AVSs
Import ("MyInterlacedFilters.avs") Inserted in all interlaced AVSs

Since OPV, it would be mandatory to include any filters in the sample runs as the contents will effect the compression and resulting Q-Value, so that will not work anymore - I fear it will get a bit more complicated for you.

[lexsteele]

I'm a little new to this whole expert encoding thing and was soooo happy when I stumbled accross this great program. Thanks for making all these wonderful, but hard-to-use, encoding programs functional for a user like myself.

Anyway, I was just reading the revisions for the new version 0.50b and I was curious to know a little bit more about this One-pass VBR option. It intrigued me because I only have a 900mhz PC so encoding time for me is slooooowww. Anyway, the revision mentioned that this one-pass VBR will reduce encoding time considerably, something that got me reall excited. I currently use CCE SP and have my CCE Customized Settings set as follows in Rebiuilder:

VBR_Bias: 10 VBR_PAsses: 3 Quality_Prec: 24

I was wondering if I were to enable the one-pass VBR option in the new Rebuilder, does this mean that CCE will perform only one pass (with almost equal quality) as opposed to the 3 (which if I understand correclty is really 2?) that I have in entered in my settings?

ANy help would be greatly appreciated. Thanks so much.

[DDogg]

I was wondering if I were to enable the one-pass VBR option in the new Rebuilder, does this mean that CCE will perform only one pass (with almost equal quality) as opposed to the 3 (which if I understand correclty is really 2?) that I have in entered in my settings? Yes, it will only work with SP versions of CCE and will use that versions special OnePassVbr mode which is a slightly different method than multipass vbr as you have been using. On a slower machine, and when the sizing becomes accurate and I'm sure it will, this OPV mode may be very helpful for you.

[Joergen]

New version 0.51! Great new features in there jdobbs, thanks!

[d62ks821]

With the ripped source and working directory on the same physical harddisk, the DVD Rebuilder progress is at the mercy of harddisk seek/access time. It is easy to solve this problem for steps I and II by placing the ripped source and the working directory on different physical harddisks.

HOWEVER, there is presently no solution for step III. I would like to be able to specify a separate path (on a different physical hd) for output of the VIDEO_TS directory. By my estimatation, this would reduce the time required for step III by an order of magnitude.

[jhmac]

Originally posted by d62ks821
I would like to be able to specify a separate path (on a different physical hd) for output of the VIDEO_TS directory. By my estimatation, this would reduce the time required for step III by an order of magnitude.

I 2nd this request...

[jdobbs]

Originally posted by d62ks821
With the ripped source and working directory on the same physical harddisk, the DVD Rebuilder progress is at the mercy of harddisk seek/access time. It is easy to solve this problem for steps I and II by placing the ripped source and the working directory on different physical harddisks.

HOWEVER, there is presently no solution for step III. I would like to be able to specify a separate path (on a different physical hd) for output of the VIDEO_TS directory. By my estimatation, this would reduce the time required for step III by an order of magnitude. I agree you will save time, but nothing significant.

Step three takes about 15 minutes. Let's take a leap of faith and say you can cut that time in half... Then you'd be taking a 3 hrs and 30 minute encode and make it 3 hrs and 23 minutes.

The only place to make a significant time improvement in the encode is in the encoding phase -- and that part is CPU bound, not I/O bound.

[d62ks821]

I understand your point about overall time. Nonetheless, here is my arguement in favor of a separate path for VIDEO_TS: ...On my computer a recent phase III took 30 minutes, where as I can copy a 4G file HD-->HD in about 7 minutes. So, I suspect phase III could be cut to 7 mins. ...There is also the additional concern of thrashing the HD unneccissarily, which happens when you copy large files from/to the same HD.

[d62ks821]

.vaf file generation not needed for OPV encoding. I manually set vaf_out=0 in REBUILDER.ECL prior to encode and appear to get some performance boost. ...it's probably worth dvd-rb doing this automatically.

[jdobbs]

Originally posted by d62ks821
.vaf file generation not needed for OPV encoding. I manually set vaf_out=0 in REBUILDER.ECL prior to encode and appear to get some performance boost. ...it's probably worth dvd-rb doing this automatically. If you do that, though, and then decide to perform another pass (e.g. because the output was oversized) you cannot.

[d62ks821]

Originally posted by jdobbs
If you do that, though, and then decide to perform another pass (e.g. because the output was oversized) you cannot.

Ok, you are correct here, ...i would save having to generate a .vaf again if I choose to re-encode with multipass VBR. ..but planning for success :), i dont need it.

Although, you seem to be saying that I could change encode mode to something other than OPV, in DVD-RB and re-encode. ...say OPV is too big, so switch to VBR N pass and simply press the encode button again. ...is this what you are indicating? ...and will it work?

...or perhaps another alternative would be to reencode with q-factor incremented on largest asset. (..by manually editing the .ecl file). Might nice to have an "Opps too big" option which would adjust q-factor and auto restart an encode.

[jdobbs]

Originally posted by d62ks821
Ok, you are correct here, ...i would save having to generate a .vaf again if I choose to re-encode with multipass VBR. ..but planning for success :), i dont need it.

Although, you seem to be saying that I could change encode mode to something other than OPV, in DVD-RB and re-encode. ...say OPV is too big, so switch to VBR N pass and simply press the encode button again. ...is this what you are indicating? ...and will it work?

...or perhaps another alternative would be to reencode with q-factor incremented on largest asset. (..by manually editing the .ecl file). Might nice to have an "Opps too big" option which would adjust q-factor and auto restart an encode. Yes. You'd have to change one parameter in the ECL. My plan is for that to happen automatically at some point (in the event output was oversized).

[begu]

About the discussion of optional different outputdirectory for phase III.
Well, it might save couple of minutes, but the more important thing for me is the random access of hard disc. I would be happy to set the output to different HD, because it would make less mechanical stress to hard disc. It wouldn't be only not the time saving for me. :)

But this is only a minor request, I believe that jdobbs does great job, here, and has his own priorities for programming. :)

[jdobbs]

Mechanical stress on the hard drive? ;) I'm an engineer...

[Harm]

Maybe already asked for...........

I'd like the rebuilder display show me what kind of audio track it is besides only the language (e.g. DC)

I'd like rebuilder to tell me approximately the size of the audio tracks

I'd like rebuilder to give me an estimate for the compressionpercentage before I run the prepare phase (if that's possible - I've seen it for example with DVDshrink)

I'd like rebuilder to have a 2 step CCE process: step 1 prepare - step 2 encode + rebuild; this being more logical because the prepare phase gives you the compressionpercentage and if you are not happy with that you can rerun it with some audio tracks deselected leaving the encode and rebuild to be executed overnight.....

I'd like rebuilder to give me the possibility to replace an entire subtitle (I know this is already asked for but I still include it here to be complete........ ;) )

I think rebuilder is a great tool! Hurray for jdobbs! :cool:

[Joergen]

Originally posted by jdobbs
Mechanical stress on the hard drive? ;) I'm an engineer...

Buy samsung, they still have 3 year guarantee when others (should be in quarantine) have only 12 months. I've lost 3 maxtors and 2 IBM's and I will never buy from them again. WD seems ok too.

[jdobbs]

Interesting. I buy whatever is on sale.... I've never had one go bad, even though I run them 24hrs a day, 7 days a week... in my environment I usually get rid of them when they get too small... (or more accurately the demand gets too large).

[Joergen]

You've been lucky.. or like you said they get out of the cycle. IBM has gotten very unpopular due to crappy products, as many people would say from experience.

This is evident in the shortening of guarantees, Maxtor and IBM used to have 24 or 36 months but now have 12, from which one can conlude their lifetime expectancy has equally diminished and they are getting losses from replacements.

I usually keep my HD's for years, moving them to other PC's where they serve files till they drop (and they will). On the other hand, the 1, 2, 4 and 8GB models from any maker havent failed to date. They're cutting corners in HD hadrware these days I suppose.

Then again HD space is so cheap it doesnt matter, if only the data stored on them was equally uninportant, and the manhours wasted reinstalling the system. Backup, backup, backup!

[goshawks]

Last edited by jdobbs on 28th March 2004 at 12:45

[jdobbs]

Originally posted by goshawks
Last edited by jdobbs on 28th March 2004 at 12:45 :confused: What am I missing in this statement?

[goshawks]

In my above post, it would be nice to have a periodic update/repost of your 'Installation and Usage Instructions' to be certain of the most recent versions of the associated programs (AVISYNTH, DVD2AVI/MPEG2DEC3DG.DLL, CCE, etc.) that are proven to work with your most-recent release of DVD Rebuilder. If no changes, you could simply cut/paste the older Instructions into these later pages. This would help newbie fears. Thanks.

[jdobbs]

I think wmansir keep this (http://forum.doom9.org/showthread.php?s=&threadid=74308) thread pretty much up-to-date.

[goshawks]

"I think wmansir keeps 'this' thread pretty much up-to-date."

Thanks, I was not aware of that thread. I will bookmark it. (You might want to insert a reference to that thread in your 28th March 2004 03:36 'Installations' post, to avoid other folks also questioning updates to it.) I am looking forward to giving your program a workout, when time allows.

[rayvt]

"I buy whatever is on sale..."
Me, too. As long as it's a major name--WD, MAxtor, Seagate.

" I've never had one go bad..."
I have. In fact, Thursday my wife's IBM drive started imitating a sewing machine. And we lost everything on it."

"... even though I run them 24hrs a day, 7 days a week"
Next thing, you'll tell us that your car engine works great even though you change the oil every 2000 miles.
Continual power-on is the *best* and least-stressful way to run a disk. Turning off and on is much much harder on it.

[jdobbs]

Yep --- I agree.

[stargazer]

I don't know if this was already discused, and I appologize if it was. It's hard to thoroughly read few thousands of posts...

As I understand, .d2v file describes/serves whole video record of a movie. Then through Avisynth, video is cut to cell ids or some other segments (via trim function). Every cell ID is separately encoded, where "dynamically assign cell bitrates" option should assure correct bitrate distribution through the movie.

It seems to me, as long as number of segments is different from infinite (hope this is a right definition, I don't know the word opposite to infinite :o ), the bitrate distribution will never be as good as it would be if we encoded the whole movie without cutting.

So my question would be: is it possible to reencode whole movie, and cut it to cell IDs (or whatever segments are) afterwards?

P.S. Many thanks to jdobbs for making a dream come true (when high quality DVD backups are concerned, ofcourse ;)) As soon as my country will be "enabled" on PayPal, I will certainly make a donation.

:thanks: and excuse me for my bad english ;)

[Joergen]

Originally posted by rayvt
Thursday my wife's IBM drive started imitating a sewing machine. And we lost everything on it.
Another happy IBM customer I see. :o

[HanSolo00]

Originally posted by Joergen
Buy samsung, they still have 3 year guarantee when others (should be in quarantine) have only 12 months. I've lost 3 maxtors and 2 IBM's and I will never buy from them again. WD seems ok too.

You never know though. There was one year I bought around 30 new Western Digital hard drives at the same date... and every single last one failed in the span of 2 weeks to 3 months (controller failure.) With this staggering failure rate the most I was able to dig out of WD warantee department was 'um well we had a problem...' In other words they had a serious manufacturing defect with that run of drives. These days they don't seem to last as long as the old ones--they are cheap and they perform like it. However, my expensive SCSI drive has outlasted pretty much all my IDE drives.

[PINOBIRD]

My second 120gb maxtor (this year) is clicking when storing data on it.
Is he going to die soon ?:scared:

[onesoul]

Hi, could you consider implementing two field editor? One would be for progressive (like the one existing now) and other for interlaced sources which means a clip that would require converttotyuy2(interlaced=true) unless it was specified interlaced off at a specific vts in the menu.

It would be nice since interlaced frames should be treated differently.

Btw, I have found a good filter chain for clean sources with almost none detail damage, (preserving more detail than deen which to my eyes seems to darken certain scene areas like flaming red), and removing noticeably some noise.
undot()
DctFilter(1,1,1,1,1,1,.5,0)
VagueDenoiser(threshold=0.8,method=1,nsteps=7,chroma=true)
edit: For Interlaced sources should be:
separatefields()
e=undot().selecteven().DctFilter(1,1,1,1,1,1,.5,0)
o=undot().selectodd().DctFilter(1,1,1,1,1,1,.5,0)
interleave(e,o)
weave()
VagueDenoiser(threshold=0.8,method=1,nsteps=7,chroma=true,interlaced=true)

[nwg]

Nice improvements with v51C. :)

My second 120gb maxtor (this year) is clicking when storing data on it.

I would say yes. I have the same problem and it causes the system to freeze and I have restart. It sometimes does it again within minutes.

The HDD is 7 years old though and used everyday.

[Joergen]

Originally posted by HanSolo00
You never know though.

Thats why I said "seem ok" :D My bro had a 80GB WD drive that started giving bad sectors after 3 months of use. So far from personal experience, Samsung is the only manufacturer that hasnt failed me yet.

[The_Flash]

Any chance we could get a "Make New Folder" option after clicking the button to specify the Working Path? Would save me from a few sighs this week. Thanks.

[Patjen]

Originally posted by jdobbs
I agree you will save time, but nothing significant.

Step three takes about 15 minutes. Let's take a leap of faith and say you can cut that time in half... Then you'd be taking a 3 hrs and 30 minute encode and make it 3 hrs and 23 minutes.

The only place to make a significant time improvement in the encode is in the encoding phase -- and that part is CPU bound, not I/O bound.


to be precise : 7 min for the full 4.32Gb Video_ts folder :-))

Don't know if somebody suggested it already but is it possible to add a log file that would show how long the entire encoding took and at what average bitrate with CCE ? Kinda like the dvd2svcd.log ?
I let dvdrebuilder shut down my pc when its done (encoding while I sleep, but I find it always interesting to checks the speeds and bitrates afterwards, also handy for benchmarking)

[jdobbs]

It's been suggested and is on the "wish list" -- I'll add it in soon.

[The eDealer]

Hi jdobbs,
first of all: it's a great program and I hope you will keep on driving it further so successfully, as up to now. :)

I have two small suggestions which doesn't seems to be extensive:
- If the selected main language (e.g. german) is not available in a stream (e.g. extras), always the only available one (e.g. english) should be chosen.
- An additional (root-)subdirectory for the rebuilded video_ts and audio_ts directory, so you can automatic produce an image.

Thanks in advanced.

[Joergen]

I noticed something that could do with improvement:

-I encoded a 3.45GB original disc with convert lb->16:9 and the output was only 3.47GB aswell. Couldnt/shouldnt DVD-RB fill the disc to its usual target sectors in this scenario?

I used CCE 2.66 3pass, disable interlaced, convert to 16:9, and dynamic bitrate.

EDIT: I think the dynamic bitrate option is causing it. The reported average bitrate is 1000 higher than the "max bitrate" :o I'm encoding now without dynamic bitrate and the average is 5.600 in the .vaf pass which is what it should be to fill the disc.

[sweetness]

I did a sreach but didn't find anything on this, so can i suggest a main movie option. I know jdobbs wants DVD-RB to be a "one-click" 1:1 backup tool and not an editing one.
My arguement is that sometimes people just want the main movie and don't care for the extras or menus. Plus i can always put the original in to watch that stuff(every rare).
I also know there are work arounds on how to do this but it takes time(20 to 30 min)and harddrive space to set everything up. i would like to just que up the folder and go to work and come home to find everything done. Just my suggestion no big deal if it's not in the plans. thanks

PS.great work too. if it wasn't for DVD-RB i might have never went down the CCE path.

[nwg]

sweetness wrote,
My arguement is that sometimes people just want the main movie and don't care for the extras or menus

You can use DVD2DVDR with CCE. This only does the movie.

http://www.davis28.pwp.blueyonder.co.uk/