this is why i prefer rebuilder to shrink

[blutach]

And I would support the thumbnail approach.

Regards

[UncasMS]

... max smooth is quite slow, but is the choice for shrinkers at low BRs (say <2Mbps). I'd stick with "sharp" for anything over 3Mbps.

and here lies a severe problem

the average bitrate for this title was: Overall Bitrate : 3.085Kbs
except for one segments (which was 2.200kbps but that segment was only 0.8% of the entire movie) the bitrates were between 2.600kbps and 3.500kbps

this leads me to the conclusion that AEC sharp should need way more than an average of 3.000kbps because we clearly see the result is horrible

the average of 3.000kbps in this title required max smoothness and still wasnt completrely satisfying to me


so what will the shrink user go for?

- will he care to analyse his material before transcoding?
- will he know HOW to analyse the source and draw apropriate conclusions from this analysis - shrink does not provide a solution for this task?
- will he really care for these extra steps and no less importantly will he really use the SLOWEST shrink mode: max smoothness?

i pretty much doubt it

but then on the other hand i might ask how much work will be necessary when using rebuilder with a decent encoding engine?

the answer is: hardly any

loading the material into rb, keeping the matrix one likes best and has used before, inserting a couple of lines into the filter editor in case one might think this is necessary and then press START

of course the encoding process takes more time with an encoder, but i have my machine running overnight or start the conversion before leaving home

i can easily have more than one dvd processed while i'm away from the pc through the *batch processing* in rebuilder

so i may wake up/come home and find 2-3 titles converted in flawless quality - no need to check the average bitrate before, no need to look through my entire movie before burning the backup in order to make sure, there are no crippled scenes

[laserfan]

...i may wake up/come home and find 2-3 titles converted in flawless quality - no need to check the average bitrate before, no need to look through my entire movie before burning the backup in order to make sure, there are no crippled scenes

Your confidence is remarkable, and good to see! I have only done a couple of DVD-RBPro conversions recently using HC0.16, but I haven't been able to watch them yet. One is "Bourne Identity" which I've never seen but expect it to be a slam-bang challenge for the process.

Frankly, making these backups was so easy it's a little scary--could a program really work THIS WELL??? For $10 bucks!!?? Or whatever I paid. If I can see no flaws in Bourne I will likely abandon Shrink at least for anything actually involving "shrinking"!

[2COOL]

Frankly, making these backups was so easy it's a little scary--could a program really work THIS WELL???

I'm glad you saw the light. I have been using HC Encoder more now instead of CCE seeing the quality on some of my backups.

[UncasMS]

Your confidence is remarkable, and good to see!

it is confidence for a small part and years of experience for the rest of it ^_~

no, honestly

i guess i'm one of those guys who have spend hours without end on testing different encoders, different settings and options and transcoders and i am sensible to macroblocks in particular

there are dvds that i have converted some 20+ times (this is nothing compared to jdobbs converting THE MATRIX i guess *g* ) but it is an aweful lot of comparison and apart from being biased and convinced i prefer my conversions flawlessly and perfect and i dont want to sacrifice major parts of the content

i do like dual-language dvds, i do want to keep menus, bonus and extra material and thus i cannot rely on shrink - it cant cope

[tigerman8u]

I must admit their is a big difference with encoded pic compared to using dvdshrink especially without using max smoothness

[Taelon]

Ok guys, you convinced me to give it a whirl. I know jdobbs doesn't support pre-processing but I don't careat this point. I've got 'crash' more or less re-authored the way I want it (way too much crap on that dvd). My reauthored folder is about 7.3 GB. Checking 'movie only' reauthoring in Shrink is around 67% compression (retaining all 3 audio streams 5.1ch, 2ch, +2ch dir comm.) and I'm guessing I won't be happy with the Shrinked results.

From all that you have said, should I really expect that RB (not pro, yet) with HC give better results on the reauthored compilation than shrink will with 'movie only'? Or should I just stick to using RB on the main title only? Are there any settings that I should look at changing from the defaults, without having to run a bunch of test-encodes? Any other pitfalls, problems, quirks and whatnot in the free v094 ?

[UncasMS]

1. make sure to use profile BEST with hc

2. you cannot use matrices in the freeware edition but maybe your title can use some filtering (cf. my sticky on *filters in rebuilder*) in case it may be slightly noisy?

3. in case you have extra/bonus material:
in the avs options set the vts to HALF D1 which do not belong to the main title (beware not to reduce the main title vts to half d1!) and then set mode => steal space to 25%


no need to sacrifice any content btw

[Taelon]

UncasMS,

Well, I read your excellent "quickstart guide" and decided to give it a whirl. The installer was impressive, I remember reading alot of posts in the past about people having tons of installation/setup troubles. This was a breeze.

1. make sure to use profile BEST with hc

Oops, I used the default HC setting, how much quality am I losing?

2. you cannot use matrices in the freeware edition but maybe your title can use some filtering (cf. my sticky on *filters in rebuilder*) in case it may be slightly noisy?

This is a new release of a new movie, hopefully it won't be 'noisy', so I didn't even look at the filtering options.

3. in case you have extra/bonus material:
in the avs options set the vts to HALF D1 which do not belong to the main title (beware not to reduce the main title vts to half d1!) and then set mode => steal space to 25%

I took out about half of the bonus content and set steal space to 25%, but didn't touch the HALF D1 option. Is that half frame size? Steal space alone only netted an extra - Movie improvement from extra reduction = 3.4% - according to the log, what does this mean exactly 3.4% less macroblocking, mosquito noise, smoothing?

It just finished encoding, 200 mins encode time. A few other questions if I may. What is the significance of the "Current" 1101000111###### is it just a random number generated by RB to identify the current D2VAVS script?

What constitues a VTS segment? Is it the Chapter, Cell, something else?

The Encode button changes to Stop/Pause when running, if I stop the encoding can I later resume it where it left off? Does it immediately stop or when it finishes the segment it's on?

- HIGH/LOW/AVERAGE Cell Bitrates: 4,390/600/3,454 Kbs
Is this good, bad, average?

thanks

[blutach]

@Taelon - pay $10 and use matrices. Then download the RME addon by Rockas and really get into it.

After your pre-processing with VobBlanker/PgcEdit/MenuShrink (which leave compliant DVDs - jdobbs' warning is there to caution against some stuff that doesn't leave things all copacetic), you'll look at various segments and find the right matrix for your work. This might take a few more minutes but is worth it. Then Process, Encoe and Rebuild. Burn with PgcEdit and away you go.

As for your DVD, those bitrates are pretty high - you could use a high-high or high-medium matrix.

Always use best. What does time matter - yer sleeping anyway when this stuff is going on. It might be fun to watch for about 3 minutes but after that it's like watching grass grow.

Half D1 is 352x480/576 as opposed to full frame 720 x 480/576 (edit). Steal space can usually go up to 50% but again, you might like to use the right matrix for the job.

A VTS segment in DVDRB is a cell.

If you press pause, you can get going again at where you paused - like in DVD Shrink/Recode.

Regards

[blutach]

I'm glad you saw the light. I have been using HC Encoder more now instead of CCE seeing the quality on some of my backups.

I have come to believe that HC is just so much better than CCE, especially on low BR stuff. I don't use CCE.

But Procoder 2 is best

Regards

[UncasMS]

Oops, I used the default HC setting, how much quality am I losing?

hopefully not that much - i usually stick to the best profiles just to be on the safe side ^_~

Is that half frame size?

it is half the resolution but the picture will be sized to normal when watching on your tv

the advantage is, that muss less pixels are present on which the entire bitrate will have to be distributed

this means when using steal space 25% 1/4 of the average bitrate will be subtracted and only 3/4 will be allocated to this part of the movie

now when this part is in reduced physical resolution and thus having much less pixel the 1/4 reduction may not visibly reduce the quality of this part of the movie as much as it would when the full resoution and thus much more pixel would have to b saturated with the average bitrate (of ~3500kbps in your case)

so reducing the size can be helpful when using steal space n%

what does this mean exactly

it means that the average bitrate was increased by this percentage - no additonal smoothing or else will occur

the longer your bonus material may be in terms of playing time the more impact the option *steal space* may have

What is the significance of the "Current" 1101000111###### is it just a random number generated by RB to identify the current D2VAVS script?

these are the numbers rebuilder gives to all the small segments into which rb divides the dvd - they are no random numbers

What constitues a VTS segment? Is it the Chapter, Cell, something else?

chapters and cells are part of the video title sets

The Encode button changes to Stop/Pause when running, if I stop the encoding can I later resume it where it left off? Does it immediately stop or when it finishes the segment it's on?

it does not stop immediately but halts when the segment currently converted is finished

you can exit rb at thea moment and continue later on!
you must, however, switch to THREE click method when you continue

simply fire up rb, use 3 click method and hit encode again - rb will continue with the next segment in its queue

dont forget to press REBUILD as the final step having finished the encoding part as this will not be done automatically when using 3 click method

Is this good, bad, average?

this is average imho
high bitrate would be some 4500+ kbps and low something less than 2500kbps

[Taelon]

blutach and UncasMS, Thanks alot for the in-depth answers.

I probably will send jdobbs the $10, the quality of the main movie looks great. I'm still wondering about the difference between HC normal and best, is it an extra pass or something else?

Unfortunately some of the extras look like crap. I guess I'm not understanding the 25% Steal Space setting. The one extra I looked at closely seems to be compressed over 50%?

original average bitrate = 6002 Kbps
25% steal space bitrate = 2362 Kbps

The other thing I don't understand is why does it change the # of I frames?
It threw me for a sec. when I went to do a comparison (I normally scroll to the general area in vdub-xxx then use the keyframe button to get the exact frame match). I can post the comparison images if need be.

I can't imagine it would look any worse if I used Shrink to compress it to the same size, maybe I'll test it tonight. If I'm understanding UncasMS correctly if I had used HALF D1 + Steal Space it would actually look better because the bitrate would be higher, even though it was at half resolution?

A VTS segment in DVDRB is a cell.

I see that now, except he seems to start counting at 0 instead of 1 like DVD's, seems kinda wacky.

Apparently the default was 3 click, which was fine since I was checking on it occasionally while watching another dvd.

How do you test your matrix and filter settings since it seems to only work on a full dvd compilation, or is that a feature in the Pro version to be able to test a single cell?

[cynthia]

I'm still wondering about the difference between HC normal and best, is it an extra pass or something else?

I did a test - using the three modes - this is the result.

An example of the time for the various options. This test performed on an AMD XP2000+ 512 Mb:

• 09 h 15 min - Fast
• 09 h 43 min - Normal
• 11 h 11 min - Best

I thought there should be a higher time difference between fast and best. So I can't find any reason for selecting anything else than 'Best' based on the time factor.

[jdobbs]

@Taelon

If you select Half D1 and Half Bitrate -- and then also select 25% reduction -- you are stealing 25% from a bitrate that has already been halved.

Also -- the 25% reduction isn't taken from the original bitrate. First the bitrate is proportionately reduced (feature and extras) in order to make it fit on a DVD-5 -- then if you have 25% extra reduction selected, the resultant value is reduced by 25% and the space recovered is applied to the feature.

Example:

-- Original Extra was done at 6000Kbs
-- The DVD had to be reduced at 67% in order to fit on DVD-5
-- Result is 4020Kbs bitrate.
-- The 25% "Steal space from Extras" flag is set
-- The extra would end up with a bitrate of 3015Kbs

Example at half/half and 25%:

-- Original Extra was done at 6000Kbs
-- The DVD had to be reduced at 67% in order to fit on DVD-5
-- Result is 4020Kbs bitrate.
-- The resolution is changed from 720 on the x axis to 352*
-- The rate is halved for the half/half setting
-- The result is 2010Kbs
-- The 25% "Steal space from Extras" flag is set
-- The extra would end up with a bitrate of 1507.5Kbs

Also, the number of I-Frames is set by the encoder. It depends on the setting for GOP size, what the encoder decides are scene changes, etc...

A segment is usually a cell, but it can be smaller depending upon decisions made by DVD-RB. For example -- a still frame is a segment. Starting your count at 0 isn't "whacky" -- it is the norm in computer processing.

*The DVD Standard only accepts 720, 704, and 352