looking to further my understanding of DVD conversion [Archive]

tawraste

24th June 2009, 17:25

hi everyone,
i'm new to the forum, i've joined because all of my searching has led me to here. can any of you help ? i am looking to further understand the process of dvd conversion.

i have been using autogk for a few months now, and having tried loads of different programs, autogk does seem to be the best of them so far.

i've ripped loads of my dvd collection, and i have to say i'm really happy with the results. i put them onto my media centre and watch them on the big flatscreen and 95% of the time things are favourable indeed.

what i want to do is learn the finer points of dvd conversion, and get those file sizes down but i do not know how to begin.
my current methods only allow me to get down to about 1400mb. less than this and the picture quality is simply not as good. the same goes for dark action movies such as the underworld series. i do realise that conversion inherently involves loss, i am simply looking to improve - which considering other people's results, can be done.

i have played around with the hidden settings, but things have not improved much. though some movies do seem more conversion friendly.

lots of searching has led me to some incredibly technical discussions which i lack the groundwork to understand. bitrate calculation for example - what is this for and how does it benefit?
i learn quickly, but if i move ahead too quickly with a subject i get lost.

what would any of you recommend i study next in order to progress along this subject objectively? and have any of you any useful links that do not assume you are a technical genius? lol !

looking forward to hearing your ideas,

tawraste :thanks:

Ghitulescu

24th June 2009, 18:41

tawraste

24th June 2009, 22:21

i'm happy with doing the converting - its become a bit of a challenge !

BigDid

24th June 2009, 22:23

Hi and welcome to the forum,

i've ripped loads of my dvd collection, and i have to say i'm really happy with the results. i put them onto my media centre and watch them on the big flatscreen and 95% of the time things are favourable indeed.

what i want to do is learn the finer points of dvd conversion, and get those file sizes down but i do not know how to begin.
my current methods only allow me to get down to about 1400mb. less than this and the picture quality is simply not as good.
Trying to understand your conversion method: you ask for 2cd (1400Mb), 2 passes xvid encode and 95% of the time it comes out good?
Can you confirm my supposition and give more details?

Does you media-center plays AVC/H-264? if yes you could try an easy x264 converting app that will give you good to better quality with possibly less space BUT long to very long encoding time depending on your encoding machine, Ripbot:
http://forum.doom9.org/showthread.php?p=1298488#post1298488

If you want to stick with AutoGK, there are some tips that could marginally or better improve your quality, use search in the main AGK thread or subsections, or let me know if you want to exchange on the subject but I will need more details on your encoding process as asked above and eventually some AGK logs (No need for full log, the beginning part for the settings and the 1st pass %quality result only).

PS: no discussion are allowed concerning illegally obtained material (rule6)

Did

tawraste

25th June 2009, 09:45

hi !
thanks for your reply ! ok, here is the process that i typically use so far:
i import the vobs using dvddecrypter.
then using autogk, i select in the advanced options for divx support.
i leave the width on auto, and the audio on VBR MP3 128

then, on the advanced settings tab
i select correct colour if needed accurate
i select enable standalone support (sigma is the chipset of my mvix780hd player)
i select override input ar to 16:9
i select force de-interlacing as hybrid

i have just done my copy of underworld, and trial and error meant i had to use 1400mb as a custom avi file size.

however, simple dramas with no action i can happily get down to 850mb.

i am not using xvid conversion in the advanced tab because a bug in AGK prevents me from using the standalone support and this results in reduced file sizes. also some of the films done with the xvid setting have audio that glides in and out of sync.

also, my wife is going bonkers because her favourite film (hidalgo) does not have the subtitles. using a seperate subtitle file is no good because she has to literally enable them and they show all the english dialouge too. is the "display only forced subtitles" option in advanced settings for movie specific subtitles?

many thanks,

tawraste :thanks:

CWR03

25th June 2009, 12:46

If you're not too concerned with the final size of the file, select a width and try the single pass, quality-based encoding. You should not need to override input AR or force deinterlacing as AutoGK will nearly always correctly determine the need. The advantages are encoding time is greatly reduced and the quality throughout the video and across many projects will be the same.

tawraste

26th June 2009, 13:00

thanks for that - i am running through that idea with several rips at various quality settings to see how it all pans out. i am making a note to keep all the log files and such like for future reference too. i like the fact that i do not need those hidden options, you're right about it being quicker. now i am simply choosing to correct colour is needed accurate and sigma support.
in what application would the 2 pass begin to be a necessary option then?
thanks !
tawraste :thanks:

CWR03

26th June 2009, 19:49

in what application would the 2 pass begin to be a necessary option then?
Only when you need an exact size to your output file.

BigDid

29th June 2009, 19:01

Hi,

Solution from Cwr03 is a good one and will permit to get constant quality.
I use 1pass 75% as a pre-encode to get a good estimation for a 2 pass size. So if I get a 1pass size around 1.2-1.3Gb I go for 1.1Gb and 720width 2 pass and get good quality and so forth.
For more difficult (less compressible) sources I use Autogktweaker with some degrain filters and/or different matrixes.
For max quality ask for a 100% 1pass and do a 2 passes encode with the resulting file size. Good chances you will get an undersized file so AGK will redo the first pass without B-frames which will rise the quality a little bit more.

Did

yetanotherid

26th July 2009, 19:37

Solution from Cwr03 is a good one and will permit to get constant quality.
I use 1pass 75% as a pre-encode to get a good estimation for a 2 pass size. So if I get a 1pass size around 1.2-1.3Gb I go for 1.1Gb and 720width 2 pass and get good quality and so forth.

Can I pop in with a question?
Assuming your single pass encode is done at 75% and you then do a 2 pass encode using the file size from the first, and assuming the 2 pass encode is still at 75%, is there any actual difference in the quality of the final video.... single verses 2 pass?
It's just that as a general rule I couldn't care less about file size but I'd prefer to keep the quality consistent, so I'm wondering if I'm wasting my time doing 2 pass encoding.

As a side note, wouldn't it be nice if AutoGK could run the compressibility test before you specify the file size, then let you specify or change the file size to achieve a 2 pass encode of the desired quality?

For max quality ask for a 100% 1pass and do a 2 passes encode with the resulting file size. Good chances you will get an undersized file so AGK will redo the first pass without B-frames which will rise the quality a little bit more.

Could I enquire as to why you do the 1 pass encode for maximum quality?
Wouldn't it be easier just to start with a 2 pass encode while specifying a ridiculously large file size which AutoGK could never achieve?

CWR03

27th July 2009, 01:53

Assuming your single pass encode is done at 75% and you then do a 2 pass encode using the file size from the first, and assuming the 2 pass encode is still at 75%, is there any actual difference in the quality of the final video.... single verses 2 pass?
The compressibility test isn't an accurate gauge for a final file size because it isn't checking every frame. Motion in the video is the biggest factor in determining quality per a file size, and a short sampling from a test won't find that. The whole purpose for a 2-pass encode is to allow a predermined file size, while quality-based 1-pass is for getting the image you want. So no, you won't significantly change quality by using a 1-pass encode to determine a file size for a 2-pass encode.

Wouldn't it be easier just to start with a 2 pass encode while specifying a ridiculously large file size which AutoGK could never achieve?
If you do that, you'll end up with a file that's bigger than the original DVD rip (assuming that's what you're working with), so you would be better off just staying with the .VOB file.

yetanotherid

27th July 2009, 04:17

The compressibility test isn't an accurate gauge for a final file size because it isn't checking every frame. Motion in the video is the biggest factor in determining quality per a file size, and a short sampling from a test won't find that. The whole purpose for a 2-pass encode is to allow a predermined file size, while quality-based 1-pass is for getting the image you want. So no, you won't significantly change quality by using a 1-pass encode to determine a file size for a 2-pass encode.

Thanks, but it wasn't really the question I was asking.
What I was wondering is.... would a 75% single pass encode be of exactly the same quality as a 75% 2 pass encode?

It's just that while I have no doubt a full single pass encode might at times give a more accurate indication of file size than the compressibility test, quite often if you run a single pass 75% encode and then use the exact file size produced for a 2 pass encode you end up with a quality of something other than 75%. More often than not it's pretty close, but I've done that exact thing myself in the past only to find the quality of the 2 pass encode is only going to be 68%, for example.

If you do that, you'll end up with a file that's bigger than the original DVD rip (assuming that's what you're working with), so you would be better off just staying with the .VOB file.

Doesn't that depend on the quality of the source file? I'd be thinking if you're using a good quality source file that's not going to be likely.

I've converted old video tapes at 75% quality and the resulting file has been larger than the original mpeg video which it was originally recorded to. I've also converted good quality DVD video and guessed at a completely wrong file size, which has caused AutoGK to switch off B-Frames and use a sharper matrix, but it still couldn't achieve the file size I'd specified so it definitely wouldn't have come close the the file size of the original mpeg video.

BigDid

27th July 2009, 06:57

Can I pop in with a question?
Assuming your single pass encode is done at 75% and you then do a 2 pass encode using the file size from the first, and assuming the 2 pass encode is still at 75%, is there any actual difference in the quality of the final video.... single verses 2 pass?
Hi,

I have the feeling some people here do not appreciate my contributions, so until that feeling will evolve I will restrain these contributions to answer direct questions like yours.
To the subject, I kind of agree with Cwr03, the only thing I can add is encoding bits are distributed better with 2 passes than with 1.
So if your encoding machine is quite quick you can go 2 passes; if not stay with 1 pass if it suits you. This 1 pass/ vs 2 is even more critic with h264 and longer encoding time.
Last comment, I use the 1 pass to get an idea of the quality, I do not always stick to it; example: Harry POTTER-1 1pass is 2340 Mo (with 320 Kbps audio), I will not go above 1/2 dvd for 2 passes and I will use some avisynth filters to keep the quality. My choice.

...
As a side note, wouldn't it be nice if AutoGK could run the compressibility test before you specify the file size, then let you specify or change the file size to achieve a 2 pass encode of the desired quality?
Yes it could be possible without audio; with audio AGK needs to calculate the final size after audio encoding, not before. This is why I usually re-use the audio (in Vdubmod) from the first pass encoding, no new audio or comptest calculation needed ;)

...
Could I enquire as to why you do the 1 pass encode for maximum quality?
Wouldn't it be easier just to start with a 2 pass encode while specifying a ridiculously large file size which AutoGK could never achieve?
Cwr03 answer is very pertinent, another thing is AGK will redo the 1st pass( B-frames off, max width eventually) if encoding is above 100% so again time consuming.
My suggestion is the sharing of my experience in that matter, but anyone is free and welcome to experiment by him/herself.

Did

yetanotherid

27th July 2009, 19:28

Yes it could be possible without audio; with audio AGK needs to calculate the final size after audio encoding, not before. This is why I usually re-use the audio (in Vdubmod) from the first pass encoding, no new audio or comptest calculation needed ;)

I never thought about having to work out the audio size, although it'd be a good reason to use CBR MP3 wouldn't it? That way AutoGK could work out the size of the audio file from the length of the video and the chosen bit rate without having to encode it first I assume? Same for AC3 as well I guess, if it could look at the DVD audio and extract the bit rate without having to demux it first. Or if it could use an external VBR MP3 and simply add it's file size into the calculation automatically....

Cwr03 answer is very pertinent, another thing is AGK will redo the 1st pass( B-frames off, max width eventually) if encoding is above 100% so again time consuming.
My suggestion is the sharing of my experience in that matter, but anyone is free and welcome to experiment by him/herself.

Yes I never thought about the double first pass. :-)
Although if you're doing a single pass encode in order to work out the file size for a 2 pass encode at maximum quality, then a two pass encode with a double first pass would probably take about the same amount of time as a single pass encode plus a two pass encode with a single first pass. ;-)

Thanks for your comments.

mr soft

27th July 2009, 20:51

VBR MP3 shouldn't be used in AVI.

http://www.animemusicvideos.org/guides/avtech/aviaudio.html

Ghitulescu

27th July 2009, 21:16

VBR MP3 shouldn't be used in AVI.

http://www.animemusicvideos.org/guides/avtech/aviaudio.html

Tell this to the one-click software developers, too ;)

Most people tend to use the easiest way.

manono

28th July 2009, 13:16

VBR MP3 shouldn't be used in AVI.
Nonsense:

http://www.alexander-noe.com/video/amg/en_myths.html

If those animemusic guys are going to provide only part of the quotation, I can only conclude they have some sort of an agenda of their own. When using MP3 audio, I've used nothing but VBR MP3 for over 10 years and literally thousands of encodes and have never had even one problem with it, not with computer playback, not with standalone playback.

BigDid

28th July 2009, 18:07

... When using MP3 audio, I've used nothing but VBR MP3 for over 10 years and literally thousands of encodes and have never had even one problem with it, not with computer playback, not with standalone playback.

+1

Although, when dealing with audio problems in AVI and AGK, switching from VBR to CBR "MAY" work. But switching from a problematic bitrate (144) to another may also work :)

Did

Edit: only 5 years for me, not 10 ;)

Ghitulescu

28th July 2009, 19:18

When using MP3 audio, I've used nothing but VBR MP3 for over 10 years and literally thousands of encodes and have never had even one problem with it, not with computer playback, not with standalone playback.

I had, on all divx movies I've seen (a colleague of mine was so kind to let me use several models of hardware players with HDD, also a divx-capable dvd player). Since there weren't my movies (I do not encode in divx anyway) I cannot tell which software and what settings were used, but the audio was definitively VBR. I belong to the category of people that easily spot the lipsynch, and it's extremely annoying, disturbing even. Of course, the problems appeared when I scrolled back and forth through the movie - playing it from top to bottom kept the synch within tolerable limits.

CWR03

28th July 2009, 23:18

When using MP3 audio, I've used nothing but VBR MP3 for over 10 years and literally thousands of encodes and have never had even one problem with it, not with computer playback, not with standalone playback.
+2. I don't use a standalone, only PC, but I've never had synch issues with files I've encoded from DVD. I've even used DVD Flick on some old TV shows I'd saved in .AVI with VBR audio to share with a friend, and the resulting disk played perfectly in synch.

What I was wondering is.... would a 75% single pass encode be of exactly the same quality as a 75% 2 pass encode?
There is no 75% 2-pass encode. As I already said, the whole reason for 2-pass is to hit a specified file size. If you reverse the process (Do a 1-pass encode at 75% to determine a file size for a 2-pass encode) you're just wasting a lot of encoding time. There's no way to accurately determine an output quality based on a predetermined file size.

manono

29th July 2009, 02:55

I had, on all divx movies I've seen (a colleague of mine was so kind to let me use several models of hardware players with HDD, also a divx-capable dvd player). Since there weren't my movies (I do not encode in divx anyway) I cannot tell which software and what settings were used, but the audio was definitively VBR. I belong to the category of people that easily spot the lipsynch, and it's extremely annoying, disturbing even. Of course, the problems appeared when I scrolled back and forth through the movie - playing it from top to bottom kept the synch within tolerable limits.
Are you saying that they're OK when played straight through but go out of synch when seeking? I've never had the problem with my own, but I have seen the problem with AVIs created by others. But, like you said, there's no way to know how they were created and how they went wrong. I'm not responsible for what other people do, but am only speaking of AVIs with MP3 audio I've created myself using reputable software. I feel fairly confident in saying that if you use either AutoGK to create the entire thing, or GKnot, together with Lame to encode the audio and VDubMod or something equally good to mux it at default preload and interleave settings that, assuming they start out in synch, they should both play and seek fine.

mr soft

1st August 2009, 15:32

I never had anything but grief from VBR in AVI. Seeking or pausing is one of the most common problems for me. As mentioned by Ghitulescu , playing from top to bottom is less of a problem. I'm glad you guy's have success , and your experience in this far outweighs mine or any noobs looking for advice. It should be more of a cautionary thing that you should take care when encoding with VBR.

Another page on it.

http://www.virtualdub.org/blog/pivot/entry.php?id=27

yetanotherid

13th September 2009, 14:36

There is no 75% 2-pass encode. As I already said, the whole reason for 2-pass is to hit a specified file size.

While I admire your determination to tailor the question in order to appear to have the answer, as I already said, it's not what I asked.

Assuming the point of a 2-pass encode wasn't totally obvious, but also assuming the file size I pick just happens to result in a second pass quality of 75%, and assuming I then run a single pass encode at 75%.... even if I'd been blissfully ignorant as to the reason for 2-pass encoding I'll still invariably end up with AVIs of different sizes. Sometimes not much different, sometimes a little more.
To my way of thinking, logically there's only a couple of likely reasons for that. Either a single pass encode compresses the video slightly differently (maybe as a result compressing it more or compressing it less), resulting in a different file size while somehow still producing the same quality, or if it compresses the video in the same way as a two pass encode would then one of the AVIs has a quality of 75% and the other doesn't. Or maybe it's some combination of the two?

If you reverse the process (Do a 1-pass encode at 75% to determine a file size for a 2-pass encode) you're just wasting a lot of encoding time. There's no way to accurately determine an output quality based on a predetermined file size.

I'm not sure whether you mentioned encoding time because you're saying even a 1-pass encode at 75% can't accurately determine a file size for a 2-pass encode, or whether you've got some theory regarding long encoding times adversely effecting the quality....
Assuming it's the former, if I do elect to waste a lot of encoding time and do a 75% 1-pass encode to give myself a file size for a 75% 2-pass encode...... maybe if I ask the question this way.... when I run the two pass encode and AutoGK reports a second pass quality of something other than exactly 75%, even if it's only one or two percent, in what way is the resulting 2-pass AVI different to the first?

manono

13th September 2009, 16:50

...assuming the file size I pick just happens to result in a second pass quality of 75%, and assuming I then run a single pass encode at 75%.... even if I'd been blissfully ignorant as to the reason for 2-pass encoding I'll still invariably end up with AVIs of different sizes.
This assumes that the other variables (resolution, matrix, resizer, etc.) are identical for both encodes). And they probably won't be. You'd better post logs for both encodes where you've noticed these similar or identical percentages, but way different file sizes.

yetanotherid

13th September 2009, 18:32

This assumes that the other variables (resolution, matrix, resizer, etc.) are identical for both encodes). And they probably won't be. You'd better post logs for both encodes where you've noticed these similar or identical percentages, but way different file sizes.

They're not really way different. In fact the only times I've even compared it was just for the sake of experimenting. I found that if I converted a movie at 75% using a single pass encode, then without changing settings, used it's file size for a 2 pass encode, the second pass quality is not always exactly 75%. Sometimes there's a difference of a percent or three. Nothing to worry about, no quality difference I can see anyway, but....

Out of simple curiosity I wondered about single pass 75% encoding vs a 2-pass encode which achieves a 75% quality, irrespective of slight differences in file sizes.
Maybe it's a dumb question, I dunno, but I was wondering if the 2-pass encode would compress the video differently, maybe make different decisions as to how to distribute the bits and whether that'd have an impact on quality, even if it's only theoretical quality and nothing a mere mortal such as myself is likely to notice. Maybe I could rephrase it this way....
Does 2-pass encoding do absolutely nothing more than achieve file size when compared with single pass encoding.... meaning if I performed a 2-pass encode and used it's second pass quality for a single pass encode, would both encodes be identical in quality in every way?

Maybe I'm just not good at phrasing questions, unless of course the answer is an explanation regarding the purpose of 2-pass encoding and the inaccuracy of compression tests, and I've got the question wrong. ;)