All you have to do to get a 6 channel version is to go into windows control panel sound settings and change your speakers to 5.1 Doesnt even matter if you have got a stereo soundcard (I do) then the app will rip the 6 channel version if thats what you choose in the dvd menu. I have done a 6 channel wav from REM's In Time DVD at 96khz and a 192khz stereo version as well. Both worked but the 192khz version came out at double speed and needed a 200% stretch resample in Audition. Anyone else have this issue with the high res stereo tracks?


EDIT: After trying another 192khz, this time from the decrypted dvd rather than the original source, it worked fine.

Halleck

Good question. Someone should have asked the industry this a few years ago.

Recording or mastering at higher bitrates/sampling rates is technically defensible. Playback at >Redbook is probably pointless. Or do you really think you can hear stuff up above 24 kHz and need a dynamic range >96 dB for playback?






The difference between 'higher res' DVD-A and 48/24 is far less likely to be audible than CD-DA vs 193 kbps mp3. The analogy really doesn't hold.

Bump, I'm hoping someone can answer the questions above. Thanks.

I highly doubt it, because the tools aren't extracting the audio the way EAC does it. The tools fire up WinDVD, which plays the audio back, and that is what is being recorded to the hard drive. I've also come to doubt the reliability of the tools: I launched PPCMRipper with '192000' as argument on 96000Hz content, and the frequency analysis of the resulting .wav files looked like as if it was genuine 192kHz content as well...

Anyway, the only way I can think of to make sure that the resulting .wav files are bit perfect copies, would be to be able to decode .mlp files (extracted with DVD-A Explorer) and compare them to the files extracted by PPCMRipper. But an MLP encoder/decoder costs $2,500, and if you had one, you wouldn't be using PPCMRipper in the first place :-/

I remember Roberto saying he had a copy of a MLP decoder: maybe he could do such a test, for the rest of us who don't own one?

I don't own a DVD-A.

And even if I did, in light of recent events, I wouldn't rip it and announce my feat.

I imagine it would be bit-perfect, because dvdaripper is essentially copying (with decryption) some files from a DVD. Being "files", they must be identical to the source. It's like copying a .zip file from a DVD - if it's not bit-perfect then you'd have big trouble.

A disc can't refuse a decryption key. A disc is a passive device.

Disc in the future may not contain the encrypted Media key for WinDVD's decrypting key,
so WinDVD can't calculate the Media key anymore.

Please try to use an exact language.

We're talking about the resulting .wav files, which are generated by having the decrypted DVD-A played back by WinDVD. Since I don't know exactly what's going in on during that process, I'm not sure of anything. Especially since setting up WinDVD correctly for playback is an important part of the process. Anyway it certainly can't be compared to lossless digital extraction as with .zip or .flac files.

The one guy who gets a copy of a MLP decoder, doesn't own a DVD-A :/
Seriously though, I understand your reserve about that, you've already taken some substantial risks by hosting the tools and advertising them on this forum. But for what it's worth, a decoder (MLPPROOF.exe in your case, if I'm not mistaken) should not leave any kind of watermarking on the resulting files...

I missed you too, Frank

This is the stuff I have to deal with now:
http://www.rarewares.org/rja/mail1.png

I wonder if being sued wouldn't have been nicer :B

I've seen this kind of behaviour for years, yet I still can't picture it IRL. What kind of moron actually talks like that? or even writes that way?

Hi Roberto.
Do you still have 10 fingers on your hands? ;-)

Honestly, I have 10 in both my handS. Only five in each hand... :-B


Edit: And - sorry if I don't get it but - what was the purpose of that question?

Amazing, this is what I get for not visiting for a few days - I miss out the chance to properly extract my few DVDA discs.

Oh well, I guess it wouldn't make a difference anyway, on my crappy SB Live. (I'm just overly anal about it, hehe. And I can't even properly get the ac3/dts streams out of some discs either, because they are parts of the menu itself... blah)

edit: kekeke nevermind, it DID got around on the web quite fast.

If a file can be perfectly copied, it can be perfectly decoded. Unless the decoder is buggy...

It's like copying a password-protected .zip file from DVD to your hard drive, removing the password, burning it on another DVD, then extracting its content. As long as the file can be perfectly copied back and forth between your hard drive and the DVD, the contents are safe.

I guess you can test its bit-perfectness by repeating the ripping process multiple times and comparing the results.

edit: typos

What, am I not important enough for you to read my post?
http://www.hydrogenaudio.org/forums/index....ndpost&p=311874


Edit: I would assume from other recent posts that this is as lossless as using an audio recording program (like TotalRecorder) to save audio from CDs. Perhaps someday the relevant parts of WinDVD could be extracted and it would be easy to make bit-perfect DVD-A rips.

If the only issue at hand was frequency response, then that'd be true. There are also issues of impulse response, which has been shown to be audible up to an 80khz equivalency in regards to time domain placement and soundstage; and also the fact that the "96dB" number only happens 'in a perfect world'.

It doesn't take much to make a believer out of someone who's a member of the "Good Enough Police", some of whom even say "32khz/14bit is good enough!". Again, maybe in a perfect world, with perfect mastering.

Have you ever seen an impulse, quick transient (pin drop) or series of quick transients (guitar strum) go in and out of 44.1/16, or even go into 192/24 and then be -resampled and redithered- to 44.1/16?

Try recording a 10khz square wave at 44.1/16 versus 192/24. Try looking at the pre and post echo of a 10uS 'click' at 44.1/16 vs 192/24. Record a 980hz sine at even 48/24 and resample to 44.1, and drop bit depth to 16 from 24 without dither. Look at the frequency response and aliasing. And then do the same WITH noise shaping and notice how much noise you have to *add* to shift energies in the frequency domain away from that aliasing.

This will fall on "deaf" ears I'm sure; but I"m just so tired of hearing the "44.1/16 is more than we'll EVER need" defense. Either your playback setup is as cheap as a pair of iPod earbuds, or your ACTUAL beef is with the fact that higher resolution formats and systems are either one, too "expensive", or two, too "copy protected".

44.1/16 would be an OK delivery method if the mastering steps going into it were PERFECT, and the resultant output actually *was* a slow-rolloff, 20-20khz 96dB signal. In actuality, this is rarely the case, and as I said, it would be "good" or "ok" not great. I could go on for hours about this; 44.1/16 allows for very little "wiggle room" in the final product, and on sources that are mastered in a mediocre manner, the result is horrendous.

Nothing like compressed 8-bit equivalency with a brickwall cutoff filter to sound oh-so-pleasant. (I don't know who the record companies employ to mix and master these days, but they -must- be going for the cheapest bidder or the guy with the best beer.) HR PCM has its place in the output chain beyond high frequencies and 'bat hearing'. And this sort of proof doesnt lay in the realm of voodoo and incident waves; the benefits described are strictly in the audible region.

In other words no one here have the capability to show that the copy is a bit for bit identical to the DVD-A?

Question: is it possible to rip CPPM protected DVDS discs, WITHOUT reburning them? I lack blank discs (and the required double layer discs due to the size and sector order, or even a DL burner for that matter) to do that.

Or can DAEmon mount the physical files created by Gear Pro? It does not list dv1 files as supported ones...

edit: Daemon CAN progress the dv1 files. And it plays fine in winDVD from that. With the "silence after 30 secs" protection though.

No one here has a proper MLP decoder, so no one can compare the decrypted files to the professional decoded ones.

so yea, I mean, no, no one can. But it would be a safe bet if you ask me.
Perhaps the author of these tools could be asked about this?

Good luck finding him, and then, getting him to answer you

He got very, very freaked out with the publicity.

Well, I kinda thought that you could've asked this - but I guess you can't contact him anymore then.

back to ripping. Is it normal that the channels are all mixed up in the extracted wavs?

DaemonTools can mount the Gear Images

Sorry.. I skipped out for a couple days

The only two samples I have access to are the Doors and Yes clips posted earlier in this thread.. I looked for DVD-As at the local music store but was unsurprisingly not sucessful at finding any.

I have an order in for a couple from an online store (AcousticSounds) that seems to stock 300-500 of them. I'm looking forward to analyzing Porcupine Tree's "Deadwing". The CD is smashed, clipped and limited to all hell (-9.78 for a prog rock album!) I don't expect the stereo DVD-A track will be much better, but I have high hopes for the 5.1 mix, which I'll then convert to stereo with foobar. I'll be sure to report back when they arrive.



Or... maybe I'll play it safe and send you a PM.

I've got some spectral analysis pics up of some DVD-A rips I've done if anyone's interested in seeing more...

EDIT: Previous post sounded a bit "fanboy"

I don't mind... Aren't those records a bit older ones and not so suitable for that test because they're "naturally unclipressed"?
I yearn for your report. Unfortunately, I don't own a DVD drive, but maybe now I'll need it soon for my new DVD-As.
From the few examples that I've read about at HA (e.g. guruboolez' "Dance of death" DVD-A) they also seem to be affected by the loudness war.

I have that disc, I'll check it out.

for the record, my other disc (Linkin Park - Reanimation) sounds clear and crisp, no wonder though as it was (probably) mastered on computers to begin with, being a lame electronic remix album of LP's previous works. No clipping and the like, the difference between that and the AC3s I previously extracted are quite easily audible even on a simple set. The LFE is a bit too quite on the MLP version though, possibly because they tuned up the AC3 version to sound louder on the regular cheap DVD Video sets. Well, its only 48khz 24bit for the 5.1, and 24bit 44,1khz for the Stereo mix (!!!), so no wonder that it's not that breathtaking, it sounds better then the AC3 mainly because the latter is lossy.
I haven't done comparison between the CD version and the stereo MLP, it would be irrevelant - its stored in the same sampling rates. And it's a crap album anyway. (only has one nice track, incidentally that one is probably their only good vocal track since the EP)

Yeah, I think the Doors and Yes examples are pretty well mastered on CD anyway.

As far as what we'll all start to find when we look at DVD-As, I imagine it's going to be a mixed bag. Mastering engineers charge by the hour, and I can't imagine that many record companies would pay to have two separate stereo masterings done. After all, they're already getting "the best" - in their minds - in the form of hypercompressed CD masterings. So they probably just have the compressed 192khz/24-bit stereo digital master (for example) resampled to one frequency for the DVD-A and 44.1 for the CD.

I think that most surround mixes will be much better. It's still more of a niche interest, and since the only place that mixing/mastering will be played will be home theater systems, the engineers can rationalize that they should match the levels of DVD movies. Also, not compressing the surround mix as much emphasizes the surround effect. If a listener hears compressed, constant loudness coming from every speaker, it's more difficult to place instruments. Less compression in surround allows room for instruments to breathe, you might say.

Of course, this all assumes competent recording practices, too! I've heard depressing stories about bands insisting that the engineer apply Waves L1 limiter on every track of the mix. Smash the drums, smash the bass.. mix them, and then apply MORE limiting at mastering.

It's interesting to note, that as long as the disc is not watermarked, You can listen to the MLP tracks at full quality on any card, not just on a Audigy 2 - bypassing the very annoying limitation in WinDVD.
If the audio is watermarked, though, the sound will kill after 30 secs of playback as reported before.

At least, my Dance of Death DVDA disc played back even better once decrypted, then it did on the original, full 5.1 instead of just the two front speakers. (It appears that it had no watermarking)

Ok, messing around with these tools a little bit. I only have one dvd-a and it's essentially a classical recording so there's not much high-frequency content anyway - but while the wav itself comes out 96k 24bit, it looks like the actual audio was just upcoverted from 16 bit.

In Audition you can check this through Analyze > Statistics, at the bottom it'll list 'actual bit depth'. For 32 bit stuff (and since Audition sees 24 bit stuff as 32 bit ) this should read 'float'.

I'm assuming it's just a mastering error on this disc, but it's something to watch out for - it's easy to add the extra bits to make a file 24 bit but it doesn't necessarily mean the extra resolution is being put to use.

I think you're skirting rather clsoe to TOC violation with some of these claims.
Show me the DBT results that confirm the *audibility* of >22kHz stuff, preferably with musical signals, and the deaf ears will start listening. *That* is my beef. And lets' not talk about effects of bad mastering -- that's an implementation issue, not an inherent one, with 16/44.1. Nor about the old 'look how distorted a square wave is"! move. It does not require 'perfection' to implement Redbook properly.

icedragon,

that was a very clear explanation, thanks for giving that amount of detail.

I remember someone claiming that he can hear clipping (at least on the stereo track), I can't find the thread. Guruboolez asked him for some clipping positions. If you're still in this thread here, guruboolez, what did you find out ?

Oh yeah, I agree. It's even noticable on 2 channels. Hypercompression sounds like the space is filled with a mush of instruments.
Here's a direct link to a post by skamp. He ripped the 5.1 audio of a music video DVD and unfortunately, the tracks are badly compressed.

Right, stop there. You're heading straight into the "little knowledge is a dangerous thing" area. Really. Even if you've spent the last 20 years in a recording studio.


The accuracy of impulse responses in regards to time domain placement and sound stage has no equivalence in the frequency domain. What you're talking about are interaural cues from transients. They work down to about 10 microseconds, which you might believe implies a frequency response of 100kHz, of a sample point spacing of 10 microseconds which implies a sampling frequency of 100kHz (and hence a frequency response of 50kHz).

Wrong wrong wrong wrong wrong. You can represent that 10 microsecond delay between two signals using a standard 44.1kHz sampled system. Try it in Cool Edit if you don't believe me.




It might do for me, but there are plenty here who would hear that 16kHZ low pass, and probably the time domain ringing it introduces too!



Yes, have you ever seen the impulse response of a real microphone? A real loudspeaker? They're not that impressive, are they? How about the impulse response of the human auditory system? It doesn't strictly have one in the linear sense, but have you seen signals the brain receives from a single audible impulse (click)?



Better still, trying recording a 10MHz square wave - wow - it completely wrecks that! What does it prove, in audible terms? Well, I don't know - can you hear all the features of a 10MHz square wave? How about all the features of a 10us click? etc etc

You're implying an understanding of human hearing which isn't generally accepted. Of course low pass filtering a signal with high frequency components changes what it looks like - but does it change what it sounds like?

This is the fundamental question which all your facts simply do not address.



OK, now you're just being silly. You finally suggest an example with a sensible audible signal, but then you break one of the basics of sampling theory, show the predictable nasty results, and suggest that proves something?

Anyway, enough. The scepticism here stems from the lack of objective or even rigorous (double blind ABX) subjective proof that more than ~20kHz bandwidth is required.

If you have some new evidence on this subject, please bring it.

Cheers,
David.

How to do this then ?

Hi
Can someone please anonymously upload DVDARipper and PPCMRipper to a binary group somewhere on usenet? I missed it and I have about 15 DVD-A's that I would like to try it on.

TIA

I'm not sure if CoolEdit can do this (I don't own it), but from a theoretical standpoint it is perfectly possible - just delay one signal by 0.441 samples. An implementation that supports sub-sample delay would write the output as if it were simply sampling different points on the continuous waveform.

In other words, it would be something like upsampling to 44.1 Mhz, delaying by 441 samples then downsampling back to 44.1 Khz. Although a real implementation wouldn't be nearly that slow.

I would like to rip all my DVD Audio disc and add them to my lossless CD collection.

I can create 6 mono waves, but I cannot easily combine them and compress to FLAC. I can combine them into one multichannel wav into Audition but that takes ages. I tried besweet but from 6 48/24 wav files it created one multichannel 48/16 full of static noise. On top of that I was not able to compress the Audition multichannel wav file to FLAC (in Foobar).

Could anybody suggest how to do this?

Thanks

Tomas

Next time, please start a new thread.

The kludgy solution I found to this problem was:

1. Multiplex the wavs with Windows Media Encoder into one WMA Lossless file (the AVI solution Microsoft recommends added a gap on the end of my files)
2. Decode the WMA Lossless file with Microsoft's Lossless to PCM Converter
3. Use foobar2000's diskwriter to convert the multichannel WAV into a WAV format FLAC accepts
4. Encode the WAV to FLAC

A sample 2ch DVD-Video is being posted to usenet for the forum to analyze comment on.

24 bit 96KHz stereo in a 3.8gig DVD_Video image in
alt.binaries.dvd.music

You will need a good usenet server to participate

I misunderstood the -s switch in ppcmripper.exe command. Without it you get multichannel wav that is compatible with FLAC frontend. No need for any muxing.

Thanks anyway.

Tomas

You can do exactly that in Cool Edit, but I'd choose easier numbers to prove the point! i.e. just 10x sampling, rather than 1000x.

Cool Edit's native tone generator will do sub-sample interchannel delays easily. Try generating a 20kHz sine wave with an interchannel phase difference of 30 degrees. At 44.1kHZ sampling, this is much less than 1 sample - zoon into the waveform to see how well it works.

Cheers,
David.

could someone please post hashes of the tools so people who don't feel comfortable leeching random binaries out of p2p networks can feel a bit safer?

also, http://tor.eff.org/ would provide anonymous hosting[1], with the drawback being that anyone accessing the stuff would need to use tor as well.

thanks

[1] not the guys provide it but one can setup a tor node that offers a service under a given name without being able to be traced back to an ip address.

Here are the md5's of the executables taken directly from rarewares when they were still there.

Probably the lawyer office that made him remove the files, just to check

There is another sample rip for you all to analyze. It is posted on the usenet in:
alt binaries dvd music

I'm sure that by now, someone has noticed the channels are out of order on playback. This is because DVD-Audio's layout is LF, RF, LS, RS, C, LFE, and any PC media player uses LF, RF, C, LFE, LS, RS. So to correctly play back DVD-Audio in Foobar2000, you will need to use a WAV editor such as Audition to swap LS and C, and RS and LFE.

How do I do that?

For the people who doubt that PPCMripper really delivers bit-perfect lossless audio copies from MLP encoded DVD-audio discs, I have performed some tests using my own made DVD-audio recordings to see if this statement is true.

Conclusion: Yes, it's lossless

The original 24/96 audio files are encoding to 5.1 MLP format with MLP Encoder, subsequently burned to DVD-R with Discwelder Chrome II and finally extracted from the DVD-audio disc back to WAV format by PPCMripper. Bit-by-bit comparison of the extracted WAV files show that audio tracks are indeed exactly identical to the original files. Alignment of the tracks in a multitrack editor is however necessary to do a correct phase inversion summation check. Inaccurate DAE offsets are unfortunately present during audio extraction of DVD-audio discs as is the case for audio CD's.
When properly aligned the sum of the phase inverted copy and the original file however yield perfect cancellation (sum zero in 24 bit).

Three remarks.
1. The order of the six extracted mono WAV files after PPCMripper with the -s option is the following:

0: Lf
1: Rf
2: C
3: LFE
4: Ls
5: Rs

using the following input order group in Discwelder Chrome II: Group 1: Lf, Rf and group 2: C, LFE, Ls, Rs. The latter order is standard for PC media players, but not necessarily correct for standalone DVD players (Lf Rf Ls Rs C LFE).

Remark 2:
For continuous (live) audio tracks on DVD-audio, the method of ripping by PPCMripper is not entirely lossless. The copied tracks when sequentially glued together show gaps of 10 msec of missing audio between the separate audio tracks. This unaccuracy is probably related to limitations in the WinDVD media playback routines. Same effect of small audible gaps sometimes occurs in real-time listening of the DVD-audio discs on my PC via WinDVD.

Remark 3:
The DVDAexplorer program allows you to extract individual audio tracks in MLP format, which files turn out to be perfectly playable in the Surcode MLP decoder. However it does not seem possible to extract one big single MLP file spanning more tracks, so it is not possible to binary compare the input MLP file (normally all tracks are embedded in one single MLP file as e.g. on mine own DVD-audios) with the extracted shorter MLP files taken from the DVD-audio copy.

Software:
Windows 2000 SP4
WinDVD 5.3 including DVD-audio pack
PPCMripper 007
DVDAexplorer A7

Hope that this information is helpful and may spark some more interest in the high-resolution, but slowly dying, DVD-audio medium. But remember: please buy when you think the disc is worth its superior audio quality.

Hippo