The Motherload of Audio questions

[MooreFox]

Okay, I can understand how aggravating this thread may sound, but perhaps it can be a comiliation of nearly all audio related questions.

I, like many AMV makers, am attempting to use the vocals from the anime and incorporate them into the AMV. Let me see if I follow this correctly (I have been reading for two hours, the achrvies of this forum):

Using DVDdecypter, I select an audio track alongside the video, then transport it into DVD2AVI or VirtualDUB (does anyone know which one is first or better? Do you need them both?)

The audio track will be separate from the video files correct? Is there any easy way to tell how to match them up to their original sources, so that they are in sync?

Anyway, save the audio as a wav. fomat - because it has been failry well emphasized that Mp3's suck (what have I been doing all these years!?!??) Do you do this in DVD2AVI/VirtualDUB as well, or in another program?

From my understand, you can then import them into your editing software of choice (in my case, Adobe Pro 1.5).


Okay, now one last part to my UBBER questions - Again, like a lot of editors, I would really really love to just isolate the vocals, and diminish the background noise/music as much as possible. And don't worry, I've seen clearly that this is a very difficult undertaking, but I am committed to attempt it. Now, here are some other links to thread topics I have been referring to throughout this rather intense journey:

http://www.animemusicvideos.org/phpBB/v ... hp?t=45257
http://www.animemusicvideos.org/phpBB/v ... hp?t=40242
http://www.animemusicvideos.org/phpBB/v ... hp?t=48028
http://www.animemusicvideos.org/phpBB/v ... hp?t=65676

In particular
http://www.animemusicvideos.org/phpBB/v ... hp?t=51855
http://www.animemusicvideos.org/phpBB/v ... hp?t=40770


So (especially in the last two), there are several theories and methods for trying this. It seems like a very sound idea to remove the other channels (everthing except center in my case, since I want the vocals), but does this only work with Dolby Digital 5.1? Let's just assume for now that my copies are not in 5.1, is that way still possibly (btw, is there anyway you can check - besides listening of course, my computor only has two speakers anyway)

And the idea put forth by "post-it" in the last link seems very intelligent, does anyone know if that way holds any merit, or has the potential of fixing my problem? Like I said, I know how unlikely it will be to remove all of the background music, but I want to do all that I can to make it as clean as possible.

Oh yeah (I'm so sorry this is taking so long) equalizer effects were also mentioned to a great degree, but never in any specific sense. Can anyone go into more depth on them, especially any that exist in Adobe Pro, and how I can utilize them to the best advantage?

Okay, rant done - except for one last thing - Audacity. It is still a free program right? And how does is compare to Audition (I'm a little too poor right now to afford such a pretty program)


Thank you so so much, and any help you can offer at all would put me closer, and I thank you for that.

[downwithpants]

Using DVDdecypter, I select an audio track alongside the video, then transport it into DVD2AVI or VirtualDUB (does anyone know which one is first or better? Do you need them both?)

you use DVDdecrypter to rip the VOBs from the DVD to your hard drive. the vobs contain both video and audio streams.

you use DVD2AVI (now called dgindex) to index the video stream as well as separate the audio streams (usually an AC3 stream).

you use virtualdub to render an AVI file from the the indexed video stream. it's also preferable to use avisynth to frameserve the video stream to virtualdub. however, this is only important for video stuff, not audio.

when the ac3 stream is extracted via dgindex, you can transcode it to WAV using the besweet program. see: http://www.animemusicvideos.org/guides/ ... index.html

The audio track will be separate from the video files correct? Is there any easy way to tell how to match them up to their original sources, so that they are in sync?

the wav file produced by besweet will contain only audio data. usually the audio is synched to the video stream from the same VOB, but sometimes there's an offset. If there is an offset, the streams text file generated by dvddecrypter will tell you what the offset is (i forgot which way is +/- in regards to audio delay or video delay).

Anyway, save the audio as a wav. fomat - because it has been failry well emphasized that Mp3's suck (what have I been doing all these years!?!??) Do you do this in DVD2AVI/VirtualDUB as well, or in another program?

besweet

From my understand, you can then import them into your editing software of choice (in my case, Adobe Pro 1.5).

pretty much anything will read wav.

Okay, now one last part to my UBBER questions - Again, like a lot of editors, I would really really love to just isolate the vocals, and diminish the background noise/music as much as possible. And don't worry, I've seen clearly that this is a very difficult undertaking, but I am committed to attempt it. Now, here are some other links to thread topics I have been referring to throughout this rather intense journey:

http://www.animemusicvideos.org/phpBB/v ... hp?t=45257
http://www.animemusicvideos.org/phpBB/v ... hp?t=40242
http://www.animemusicvideos.org/phpBB/v ... hp?t=48028
http://www.animemusicvideos.org/phpBB/v ... hp?t=65676

In particular
http://www.animemusicvideos.org/phpBB/v ... hp?t=51855
http://www.animemusicvideos.org/phpBB/v ... hp?t=40770


So (especially in the last two), there are several theories and methods for trying this. It seems like a very sound idea to remove the other channels (everthing except center in my case, since I want the vocals), but does this only work with Dolby Digital 5.1? Let's just assume for now that my copies are not in 5.1, is that way still possibly (btw, is there anyway you can check - besides listening of course, my computor only has two speakers anyway)

And the idea put forth by "post-it" in the last link seems very intelligent, does anyone know if that way holds any merit, or has the potential of fixing my problem? Like I said, I know how unlikely it will be to remove all of the background music, but I want to do all that I can to make it as clean as possible.

removing multiple non-centered sounds will be much more difficult than removing centered sounds using the method post-it describes. with off-center sounds, the problem is that you don't simply have an interaural timing difference, but also and interaural level difference, meaning that you would not only have to shift the timing of one track, but also amplify the sound by some amount in order to cancel out the sound. very difficult to do manually. some audio editing programs may have built-in functions that can do this.

Oh yeah (I'm so sorry this is taking so long) equalizer effects were also mentioned to a great degree, but never in any specific sense. Can anyone go into more depth on them, especially any that exist in Adobe Pro, and how I can utilize them to the best advantage?

this is probably what you'll need to do to remove the instrumentals. i'm not familiar with premiere pro, but if you can figure out how to filter your sound by frequency range, you'll want to filter out (reduce gain on) frequencies outside the speech fundamental frequency range (usually around 100-1000hz). also, you may be able to take advantage of the type of sound that is playing during the speech sample you want to use. is it music? noise? high-pitched? low-pitched?

Okay, rant done - except for one last thing - Audacity. It is still a free program right? And how does is compare to Audition (I'm a little too poor right now to afford such a pretty program)

audacity is free. it has all the basic functions - cut paste etc..., and some special effects - frequency filter, reverb, noise removal. it also allows you to script functions using nyquist.

[MooreFox]

Oh my gosh, I never said thank you! That was a ton of help, more than I ever expected from one reply. But I go in such a frenzy afterward to test it all out that I never properly thanked you. Thank you so much. I will see how it goes and keep you uppdated.

[maciko]

“In the Recording Studios of today, each channel ( track ) is out-of-phase to each
other an that being out-of-phase is how AC_5 and AC_6 encodes are
made possible from a Stereo Recording.”

This is absolute dribble!
Surround sound from stereo recordings is derived in 2 ways…
1) Re mixing of the multi-tracks to suit the Surround format.
2) Using a simple delay line for SL and SR channels, mono component to centre les 6dB, original Rand L to R and L, and a low-pass to sub channel.

“By taking the Left Channel, and slowly making the Right Channel out-of-phase, you can slowly erase -or- isolate any sound, voice or drums from ANY STEREO recording. The Method for doing such a thing is called "Roto-Blending" and the concept was developed in the 1970's [ anyone remember Quadrophonic CD-4 “

Lots more Dribble!

Quadraphonic, as spelt correctly, was a 1” (inch tape) format that carries 4 separate channels on the tape.

Also if your recording are out of phase then the studio you have been working for has most likely fired you already. A great deal of care is taken to make sure that stereo recording when summed loose as little as possible, hence clearly disproving the above theory. With the method above you would either the full scale mono component or it gradually fading out as you change the phase relation between L and R channel. Inversely you would hear an increase of the side components.

“Its not a new solution, roto-blend, its just a forgotten science - being capitalized
by Sound Marketeering today. Most people do not remember the 1978 November
issue of Popular Electronic's that had "the curcuit schematics" for doing it.
And, no! there was never a software package that contained this design”

As far as “electronic circuits” are concerned, a delay circuit can be created with capacitors, but simpler still, you can simulate this in audacity by simply slipping the channels agains eachother.

”Roto-blend is what you need and is the process that you are discribing.”

Roto-Blend is what a food processor does when you are making a protein shake!

There are other methods of extracting speach from background noise but the result is only intelligable, it is not musical in any way. Meaning it's something forensic audio people use. If you need to find out more google Audio Lattice Filter.

[Gepetto]

red... letters... burning... my... retina...