New DVD2ONE released
 

Morning Kwag,

I just picked this up from your Anouncements Area.. under:
--> A new toy to play, for DVD freaks!

http://www.digitalfaq.com/archives/error.gif


1, I just finished D/L'ng it, but I think its only for DVDR/RW drive, but the
.. file was pretty small (trial version) only 92k in size 8O

2, I take it that this is ONLY for DVDR/RW and NOT CDR/RW ??
.. wouldn't mind seeing a small sample clip though, he he...

4, Have you taken a CLOSER look at the specs in (bitrate viewer) BV.EXE and
.. compared ?? Maybe there's some differences that explains your conclusions ??

5, Since you have a DVDR/RW, why don't you do BOTH conversions ie, one YOUR
.. tipicle way, and the other (NEW) in DVD2ONE's way and THEN, COMPARE specs.
.. Maybe there are some differences ??

If you have the time to do this, that is.. was just curious too !

Have a good day.
-vhelp

Re: DVD2ONE
 

So is the full version :mrgreen: Run a test with the trial. The result is exactly the same as with the full version. Yes, there is :!:. Nominal bit rate is the same on original VOB and new VOB. Average bit rate on original is 5,732 and 3,638 on new one. The Q values are almost identical on both 8O. The peak bit rate on original is 9,474 and 7,871 on new one. These values are from "The Green Mile" movie Yes, 8+ hours encoding with TMPEG against <30 minutes with DVD2ONE 8O :mrgreen:


-kwag

@SansGrip,

Maybe we can think of a new "Transcoding" program, where instead of opening the .avs with TMPEG, we open it with this "New" KVCDTranscoder (:mrgreen:) and do something similar to DVD2ONE, but transcode to MPEG-1 for one or two target CDs :idea: :idea: :idea:

-kwag

hay Kwag,

sorry, but I lost my connect, and couldn't respond back (revise) to my
own post - I needed to revise it.

I realize that its probably ONLY for DVD backup, hence the NEED for a
DVDR/RW drive, but I thought I'd try it out anyways and see for myself.

However, this app didn't like my Modem and caused it to hang, hence my
lost of connection.
So, while I'l was offline, I decided to give it a try. Seems to not want to
work though. I set it to my C:\ drive instead (just wanted to see a 30 min
test sample - that's all.

Anyways, I'm glad this made you enthusiastic (spelling) about your
encoder project again.. haven't heard you bring it up in quite some time
now.. and you idea was a good one - your "KVCD Transcode" that is he, he...

A :idea: Hypothisis.. :idea:
or thought.. maybe the encoding "method" is sampling the scenes
that needs to have bitrage lowered/raised just enough, for a 4.3g project,
and fits within the "file-size prediction" ??

-vhelp

I don't know how it does the calculations. But so far, I've tried it on two different DVDs, and it always fits exactly on 4.3GB 8O. So it must have a very good algo. We could certainly use that algo here :mrgreen:

-kwag

Kwag,

Addendum to my previous :idea: Hypothisis.. :idea: ..

or thought.. maybe the encoding "method" is sampling the scenes
that needs to have bitrage lowered/raised just enough, and encodes those
scenes BUT with a ROUGH algos (but a really FAST one) and coupled w/ a
good "file-size predictor" .. works out to approx 4.3g project sizes.
and fits within the "file-size prediction" ??
I mean, the bitrate is obviously high enough to obtain good quality right ??

It looks like that SOME scens are being re-encoded (transcoded) and
others may be left alone, another UMMFF to its speedy transcoding.
Maybe theres a scale its using w/ it's algos and if any scenes falls
with this scale, it transcodes them.. example: observe the scale below..

1,2,3,4,5,6,7,8,9,10
......O.O.|.O.|.....
1,2,3,4,5,6,7,8,9,10

"." are frames (outside 'scale' params)
"O" are those frames within 'scale'
"|" are frames within 'scale' but do not need any adjusts, and are skipped.

Notice that in between the 1,2,3,4... the "O" 's these fall in between
a threshold, and notice the "|" 's these are w/ in the threashold BUT
maybe there "motion" is minimal, so those "|" get ignored. But, the "O" 's
get transcoded, but with so called ROUGH algorithym (but a really FAST one)
So, couple all this up and you may have some ideas how it's occomplishing
its feet.

What are your thoughts on my HYPOTHs he, he...

-vhelp

Hi vhelp,

Yes I believe that program is doing some "bracket" analysis on frames as it goes. Then on low action scenes, it just passes the data "untouched" to the output, but when action scenes increase bit rate, it encodes those scenes "on-the-fly". That has to be the way it works, because it's the only way it can encode(transcode) so fast :!:
So it's really an active/passive transcoder. Meaning it passes input data to output without affecting it, and only touching the necessary parts of the stream. If this is what it does, it's a very clever (and optimal ) design. :)

-kwag

Kwag,

Yes,
Yes, and
YES !!
That's exactly what I was thinking. My problem is, I can't always get my point
or ideas out in words. Anyways..

Now, as far as your KVCDcoder goes ( frankenocder?? ) apply this to it, and..
and, for ANY conversion, ie, VCD, SVCD, (KVCD) DVD and beyond (don't forget the
new SUPERBIT too, as they come out even better, thanks in part to the HIGHER
bitrate, which by the way, allows one to user even lower bitrates, due to the
HIGH bitrate in scenes - as they have MUCH lesser artifacts to encodes or shall
I say, transcode to lower format ie, VCD etc.. )


1,2,3,4,5,6,7,8,9,10
......O.O.|.O.|.....
1,2,3,4,5,6,7,8,9,10

"." are frames (outside 'scale' params)
"O" are those frames within 'scale'
"|" are frames within 'scale' but do not need any adjusts, and are skipped.

Here is a homemade blue-print I found (i mean, made up) that describes in an
analigy'wise demonstration of (in other words, ROUGH'ly speaking)

--------- --------- --------- --------- --------- --------- ---------
1, create a pre-defined scale, and have a "SOFT" area for frames who needs
.. to be transcoded

2, those that need to be transcoded, get checked against some "motion"
.. estimates for bitrtae level determination etc.

3, if they pass the "motion" test, they get passed on (in other words, copied
.. into the encoding project untouched)

4, if they DON'T pass the "motion" test or threshold, they get transcoded,
.. based on some calculations etc.

5, verfy the level or needs for transcoding each of those frames (in Ln 4)

6, continue on with the transcoding process.. .. .. more spider goola..
--------- --------- --------- --------- --------- --------- ---------

This is something that can DEFINATELY be added into the methods/processes
of an new Encoder. And, the so called, "New Encoder" ie, Frankencoder would
not have to be complicated in nature :wink:

Well, that's it for me for now. Just my two cents worth Kwag, on
this DVD2ONE topic. Thanks for putting up w/ me crazy theories thus far..
-vhelp

This is a post I just replied to at vcdhelp.com.

I have some suggestions for all of you. Do your own encodes and tests, and don't rely on other people results :wink:
Those screenshots were probably made taking the average bit rate produced from small tests. Do a complete encode of a movie and then compare. Here's my contribution on MY own test after encoding the complete "The Green Mile" with DVD2ONE including full AC3 audio, subtitles, all chapters and Closed Captions:

This is the original frame from the original ripped VOB:
http://www.digitalfaq.com/archives/i.../2003/01/3.png

And this is the same frame from the VOB made by DVD2ONE:
http://www.digitalfaq.com/archives/i.../2003/01/4.png

Yes, you'll see some macro blocks in this scene, because it's a fast action scene. But on the original, it's blurred too. Do you guys watch a movie in slow motion or pause ??? I don't! So for practical purposes, the movie looks identical during playback. Even on a HDTV.

-kwag

MPEG-2 -> MPEG-2 transcoding = easy(ish, i.e. possible with a 92kb executable)
MPEG-2 -> MPEG-1 transcoding = 8O

It probably doesn't need to do any prediction since it's only changing the bitrate. It knows exactly what average bitrate it must obtain to fit on the DVD-R, and it knows the existing average bitrate. Not very difficult math, even for me :mrgreen:.

Wrong term 8O. I guess transcoding wouldn't appply there. Too many different factors. Bitstream flags, etc :cry:.
Some things are the same in both bitstream layers, but some are different.
It would really be a TransEnCoder :lol:

-kwag

Kwag, and ..gowsters

I was getting ready to share a tool I came up with, and for this thread,
I thought it appropriate.. but dab-nabbit !! my PC frooze again !! 3rd time
today (early this afternoon) .. .. .. I'm hoping that you all can use it, cause
I think it will definately benefit many of us all testing our encodes etc.

After another days CRASH, I'm back up again. I tell you, that blastid
McAfee Internet suite just ain't worth it !! Pfew!! Anyways, I'll stick it
out till I setup my Internet ONLY pc. Waiting for an item to arrive :)

Kwag, I didn't see a difference in those two pics, but I got'ta tell ya.. just
looking at them at NORMAL distance, and you can't tell the difference, but
looking closely (and, using this tool) you can, but you have to look.
Anyways, if you HAVE to look, then of course, you'll find something.. you
always do !!

Never the less.. I had discovered another theory while taking a shower,
he, he.. rubber-duck-ducky he, he.. Anyways, if you're curious, I'll post
my additional theory.. .. :idea:

Have a good evening/night all.
-vhelp

Kwag,

I hope you don't mind a small contribution to the FORUM peoples here.

I know you can always use another TOOL (toy) to play with, for your video encoding
testing and so on, and so forth..

I figure, you must of gone out of your mind wen trying to obtain a good comparison
of your clips when you are testing them for quality.
I too, have gone through many frustrations in this area of testing. And, the best
that I could come up with, was loading the assumed PICS into a web browser and
alternating between them (Alt_Lft and Alt_Rt arrows) but at times, even
this was a pain. I kept wanting something better. Something that was small and
quick and didn't take up much resources to begin with.

Well, here it is. An app that answers yours (and mine) and others prayers..

http://www.digitalfaq.com/archives/error.gif

--> Download picswitch v1.0.rar
--> Download borland.files.rar - - (if app doesn't work, D/L these support files)
.. .. .. and add them to your c:\windows\system folder

It's called picSwitch..
Basically, you load in a bunch of pics and begin your swaping (though its really
ment for 2 pics, for obvious reasons) or what I like to call, Sequencing..
But, with a few twists and tweaks. Basically, it's a build up of an older version
app I used, for quickly viewing pics. I've incorporated a few enhancements above
the older version.

And Kwag, and others..
This WILL tell you, or HELP to tell you, which pic (from your video encode) is
better !! It really works!!
Oh, why not give one of your kvcd encodes a test run !! he, he... its a great
tool !! Definately a strength giver.

File types suported are:
------------------------
* BMP, JPEG and PNG, though *.BMP are highly recommended for better performance.

There are a few more enhancements that this gem could use, I'm sure, but that
those can be suggested from the users - and they are welcomed. Bugs too.
Report those if or when you find any.

And, last but not least, Enjoy :) :)
-vhelp

Hi vhelp,

Looks great :D, but when I ran the app, I get the "rtl60.bpl" missing :cry: Could you re-package the app with thal .dll :?:

-kwag

Evening Kwag, others..

Lets see if we can figure out how to get picSwitch working. It works fine
under Windows 98, provided..

picSwitch under WINDOWS 98 ( NT ? )
----------------------------------------------
1, D/L << borland.files.rar >>

2, install all three files ( rtl60.bpl / vcl60.bpl / vclx60.bpl ) in
.. your c:\windows\system directory

3, install the LPNG.DLL file in the c:\windows\system directory as well
.. note, you got lpng.dll from my picswitch v1.0.rar app

4, now try and run picSwitch

Running picSwitch under Windows 98 and from a freshly clean PC (my 2nd pc)
since it does not have DELPHI 6 Standard Ed. was a success. I only copied the 3
files from the borland.files.rar archive, and the lpng.dll (above) and picSwitch ran.

I will have to try and see if picSwitch works on my NT at work tomorrow w/ the
above steps. This shouldn't be happending, as Delphi is suppose to create
Stand-alone EXE files, w/ no requirements OR resources what-so-ever !! So, I don't
know what to say, except, if ANYONE knows why this is happening, I'd sure love to
know, and HOW to resolve this. Just to note, FitCD ran smoothly w/out a hitch on
my NT at work. So, NOW I'M REALLY AT A LOSS !! 8O 8O

Ok, NOW back to picSwitch !!

Kwag, gosh.. I'm really sorry the app didn't work for you so far. I
never expected this to happen on anyone's pc !!
I hope it finally works for you. If not, do you have a windows 98 platform
as a backup or something ?? Well, let me know of these apps finally work for
you today. Maybe we can figure out how to get it going for others too..
that's IF it ends up not working for you.

Has anyone tried the app yet ??

Thanks a lot for your patiance.
-vhelp

Very good vhelp :D
Works great now 8)

-kwag

Kwag'ee..

Hip Hip, Horray ! !! !! ! !!
(I'm sorry, I don't have any images right now..)

I just sent you a PM cause I was 8O on your outcome.

I'm glad !!

Ok, at first, you might think it's a tipicle image app - plenty of them all
floating around. BUT.. just think about it as a Tool. SansGrip's
greate filters.. how picSwitch can be used as a gauge in fine-tuning the
filters.. by observing each level of revision in filter adjustments, a better
perspective perhaps, or something. I'm having a hard time expressing
my point in words - sorry !!

But, I need some input and suggestoins and ideas from users. Well, I
hope.

Of course, I'll have to revise my website to reflect this app (and others
I will be creating) I have some app that I put on hold for time being.
Anyways..

Go on and mess around with it. If it's not comfortable or is handycapped
in some way or another, let me know, and I'LL tweak it to get it just right,
and then some. :) :)

Well, it's a releafe anyays.. Thank you again for taking part in my app.
-vhelp

You know, this program (DVD2ONE) is really a little jewel. I've backed up a couple of DVDs with wonderful results. But you know what? It's not fun at all :D. It kills the joy of compressing a complete DVD to ~800MB, and that IS FUN :mrgreen:
So I think that what we have accomplished here with all the techniques ( filters, matrixes, GOPs, etc ) still beats the crap out of DVD2ONE :!:
I can't imagine DVD2ONE puting ~6 hours on a DVD-R with the quality we do with KDVD and company ( company = filters, GOP, matrix, SansGrip :mrgreen: etc. ). It if does, hats off to DVD2ONE :D ( Buy I seriously doubdt it 8) )
The only down side is that it takes us a hell of a long time to encode :?

kwag

If anyone is interested, here's a link (translated from German to English through BabelFish) to a 3-page article describing how DVD2One works. The article also does some bitrate/image quality/macro block analysis. You lose a little in the translation, but it's still fairly readable.

http://babelfish.altavista.com/babel...p=de_en&tt=url

---------
stempy