How to rip these days?

[Wombat]

Now that I finally have the HD space, I want to start ripping my DVDs again. It seems the tools have changed. I used to use FlaskMPEG and Virtual Dub, and make DivX 4 rips.

What are the tools to use these days?

[Jammrock]

The best resource is still Doom9!

http://www.doom9.org/

Jammrock

[Indiana]

Use DVDx and DivX5.02.
The initial DivX5 release was totally bugged, but they really cleared those issues with the 5.02 release.
And those additional features like QPel do make a difference in quality at the same bitrate.

[dZeus]

I've heard that SBC DivX 3.11 (MS MPEG4v3) is still unbeaten in picture quality... then again, it's not iso mpeg4 so no chance ever to play it back on accelerators for iso mpeg4.

xvid (opensource iso mpeg4 implementation) is said to give good quality. Another opensource project, ffmpeg, also is said to give even superior quality and speed at the same bitrate.
If you want to encode in linux, you might want to check out mencoder, part of the mplayer package (www.mplayerhq.hu).

[Indiana]

No, DivX4 and 5 run circles around DivX3.11 regarding quality at least when bitrates around 1000-1500 KBps are used.
xvid is better than DivX3, sometimes also than DivX4 - but still worse than the fixed DivX5.

[dZeus]

I don't know how well the comparison at the following link has been conducted, but it doesn't seem to stroke with your opinion:

http://www.doom9.org/codec3-1.htm

[thop]

gotta agree with dzeus. xvid has still much trouble with very complex scenes (f.e. thousands of people in a concert). now any codec has problem with complex scenes but xvid manages them the worst imo.
other than that xvid is great, so is divx 5.02 but many people still prefer divx 3.11 SBC for dvd->1CD. divx4, i dont like.

now this is my subjective experience i may be totally wrong

[thop]

ps. ffdshow is definetly worth to check out! i've been using it for quite a while now. it has nice postprocessing features, and uses less cpu than the regular filters.

[Indiana]

I'd say the test is flawed in several ways:
1. they don't state the exact setting for the codecs. Here especially the max./min. quantizer values for DivX5 are the most important values to play around with

2. the encoding speeds for DivX5 are much too fast on the type of machine they've used if they'd configured it correctly to give HQ results

3. While the psychovisual filters are an easy/comfortable way (not the best, but IMO very good from a effort put into it/result ratio standpoint) to clean up mediocre quality video like the one captured from antenna, they should not be used for HQ content like most DVDs - and do majorly cause the "blurring" "missing detail" he attributes to DivX5.

4. He did apparently NOT use the Bidirectional encoding and QPel facilities of DivX5 (he only talks about GMC and psychovisual enhancements), where especially the QPel seems to give a lot more detailed picture.

5. I don't like those DivX5 playback filters at all as they introduce much too much blurring. Better encode the vido at high quality and not use any playback filtering at all.

That said Div3 content IMO looks better when played back with the DivX3 decoding and not with the DivX5 one he uses.

All in all Div3.x causes much too much MPEG2-lookalike blocking artifacts even at higher bitrates to be considered HQ IMO, especially in high-motion scenes (and I do like action movies) or when you want to watch the result on a nice big television set.

BTW, I spoke about bitrates over 1000KBps (that's a 2CD Rip, for very long movies even 3CDs), and while DivX3 doesn't really gain much quality from about 800KBps DivX5 does.
Look at his 2CD rip images (Saving private Ryan), those show this quite well with DivX5 giving IMO the best overall compromise between annoying blocking and detail, even though he did use psychovisual filtering and playback filtering and apparently did not set especially the max. quantizer value to the right settings (should really not be over 7 to avoid blocking). I usually use 2/6 or 3/7 depending on movie length for min/max quantizer.

Do a 2CD rip of a DVD with DivX5 with this setting and play them back without those nasty playback filtering: you should get highly detialed movies with not much blocking. do the same with DivX3.x and while you could get high detail you WILL get a hell of a lot blocking artifatcs in high motion scenes that are annoyingly visible even when played back on TV. For DivX3.x movies you have to use the playback-filtering to make the blocking artifacts only partly acceptable and this introduces major blurring.

[Indiana]

P.S.: As you can see here I was definitely NOT happy with the first DivX5 version. But although it was only a minor revision-jump from 5.0 to 5.02 they really got rid of all my problems. DivX5 simply was one those softwares that were released just a few weeks too early...

Look at this page to see another comparision between codecs that comes to an entirely different result - but this one is congruent to the results I get here. DivX3 has just too much blocks to even look fairly nice.

[The_King]

I use gknot and that seems to work fine....

[Evildead666]

I concurr with DivX5.02, its a very good compressor.

btw, did anyone else have trouble getting rid of the Div4 codec?
I had to manually get rid of it from my HDD's, cos it kept on reinstalling itself....and taking over the DivX settings in Flask...

[Gurm]

Hrm.

I've been using CladDVD to rip the DVD video to the hard drive, and then of course VirtualDub/DIVX5.02 to encode, or TMPGenc to make VideoCD's.

What are the advantages of DVDx over Clad? Any?

- Gurm

[Indiana]

DVDx does this in one pass: it reads the DVD and simultaneously converts it to DivX, and Lame for Audio (or any other .avi video compressor) on the fly. So there's no need to create a temporary large DVD MPEG2 file on your HD.

[dZeus]

I hope it doesn't buffer enough data from the DVD to let it spin down and spin up every so many seconds/minutes. That's one great way to kill your DVDROM drive.