GUIDE: Stereo to 5-Channel Surround

[Eye of Horus]

A GUIDE TO MAKE 5 CHANNEL SURROUND FROM STEREO
+++++++++++++++++++++++++++++++++++++++++++++++

This method provides excellent results. In some aspects it even sounds better than AC3 files ripped from DVD. Give it a try and share your thoughts with us !

Here is the software you need.....

- Cooledit
(http://www.syntrillium.com/download/)
- Aurora plugins
(http://www.ramsete.com/Aurora/download/)
- Tools Ambisonics
Included here - see below message

The Tools include 3 wav files you need further in the process and the free programs pack2x2to4.exe and ambidec.exe.

-> a B format signal is a 4 channel signal. W is the sound pressure. X, Y and Z are the harmonics on the pressure field. X is a vector directly in front of the listener, Y is a vector on the side, and Z is the vector of height. Since we'll decode a 2D signal (remember, surround is not 3D audio, it is 2D!), we can forget about the Z signal.

-> The goal here is to convert your stereo CD signal, which is UHJ - encoded, into a B format signal, which carries all the data so that you can decode all the channels you'd want to.

Okay, let's start working !!

-> Rip a stereo track from a CD

-> Using CoolEdit and Aurora, do this.

-> Start CoolEdit.

-> Load the UHJ-W filter file ,a 44.1 Khz file . Copy it to the windows clipboard. Press CTRL+6 for Windows Clipboard and the CTRL+C to copy to that clipboard.

-> Load the stereo signal (the ripped stereo track you want to convert).

-> Launch the "Convolve with Clipboard" Aurora module, and set the following settings:

Channels to convolve: Audio data: Both. Imp. Res: Both. No Crosstalk cancel mode. No remove DC component. No autorange result. No timeReverse Imp. resp. Gain =0.

-> Convolution begins. When it ends, you'll have a stereo file. We need it in mono.

-> Convert Sample type, from Stereo to Mono, with a 50% mix on each channel and 32 bit resolution.

-> Close and restart Cooledit. This is necessary to be sure that the clipboard is empty again ! Make sure you set the Windows clipboard (CTRL+6)

-> Repeat the procedure, but load the UHJ-X and UHJ-Y files, so you end up with three mono files (which will be your w, x, and y signals).

-> Now remember, the "Z" component means "height". So we really don't need to decode it.

-> Close and restart Cooledit

-> Now we have to pack the W signal along with the X signal.
Re-transform W in a stereo wave file, with 16 bits depth, using the Convert Sample Type module in Cooledit. Be sure 100% of the signal goes to the left mix, 0% to the right, and that resolution is 16 bits. Enable dithering.

-> Now switch to the X signal (the X file we've been working on). Resample it to a 16 bit mono file and Copy it to a Cooledit Clipboard. Go back to the W signal, select the entire right channel and paste the X signal file into the right track of the W stereo signal. You'll end up with a stereo file with the W signal on the left channel and the X signal on the right channel. Save it !

-> Close and restart Cooledit

-> Switch to the Y signal. Convert it to stereo, 44.1/16. The Y signal will be 100% on the left channel. Since we don't have a height (Z) signals, the right channel signal should be left empty. Save it !


-> Now you have two stereo signals. Change the extension of the WX file (should be WAV) to "WX". Change the extension of the YZ file (should be WAV, too) to "YZ".

-> Run the Pack2x2to4.exe file from a DOS prompt. It will ask you the names of the WX and YZ files. It will then save a 4 channel WXYZ file suitable for use with AmbiDec.
! So make sure the extension of the saved file is .WXYZ

-> Use ambidec. It's a program that is used to decode a B-Format file (the one you made) to produce a Wave file with speaker feeds. Those files will be suffixed in .wxyz.
You can decode the signal into any rig type. If you are using a standard 5.1 system, you'll want to decode it in "Pentagon".

-> Use it like this:

ambidec -r Pentagon (input file) (output file)

-> The file you'll end up with is a 5 channel wave file. Now we must encode it to DD 5.1
Load the entire file into SoftEncode. The program automatically recognizes the feed as a 5 channel wave file, and presents it as such. Encode it with a setting of 640 K/s to get the higest possible quality in DD5.1 .


-> Burn the result and enjoy ! The results are really superb !!

OR

Use Bspilt to extract 5 wav files from the wxyz signal and use Surcode DTS to make a DTS.

Commandline for Bsplit :
besplit -core( -input song01.wav -prefix g:\song01 -type wav -demux )

song01.wav = song01.wxyz
g:\song01 = the place where the 5 mono wavs are saved

Use the files in this order :

Left front : song0101
right front : song0103
center : song0102
lfe : stays blank
left surround : song0104
right surround : song0105

NEW METHOD FOR THE LAST STEPS
=============================

After making the wxyz file, there is now an alternative for Ambidec and Besplit. My friend Andreas brought this goodie to my attention !
First the link where you can download the program VVMIC, which stands for Visual Virtual Microphone.
http://mcgriffy.com/audio/ambisonic/vvmic/
Download, install and run the program.
Here follow the steps :
1. Click on input file and chose your wxyz file. You can chose between a wxyz extension or a wav extension.
2. Click on output file and give it a name
3. Select multiple mono files
4. Change number of outputs to 5
5. Select link pairs and select all directives
6. Click on G-format
7. Now click on PLAY and you will end up with 5 mono files !

In surcode, use the same order to make a DTS from those 5 files as described above (1-3-2-blank-4-5)

Results are a fraction less then the Ambidec/Besplit method, but that's mainly, because there is no gain control. The volume is lower and it sounds a little bit les crispier, brighter.

But.... it is very easy and saves a lot of time and it is very fast !



Good luck and let us know something about your results !

Eye of Horus with the much appreciated help of Dr. Freudiana

DSPguru : We hope you will be able to include this