Creating 2.0 surround

[Xoanon]

OK, so what does this mean:

LS -->(50%L + PhaseInverted 50%R) w/ +90º PhaseShift

Do you mean I invert the right channel of waveform? How can I do that if its mono, which it is?

[Cornucopia]

LS -->(50%L + PhaseInverted 50%R) w/ +90º PhaseShift

Add it to 2 tracks:

1st Track: LS -->100%L, but down by -3dB (so to equal 50%)
2nd Track: LS --> 100%R, Down by -3dB, AND Phase Invert (Yes, you do Invert--Positive waveheight becomes negative and vice-versa)

Do similarly appropriate things for the RS.

Scott

[3dsnar]

OK. This is exactly what you should do:
Lt = L + 0.7071 C + 0.7071 LFE - 0.8944 BL - 0.4472 BR
Rt = R + 0.7071 C + 0.7071 LFE + 0.4472 BL + 0.8944 BR

---------------------------------------------------------------------

More here:
http://forum.doom9.org/showthread.php?s=&threadid=57988

[3dsnar]

(Static) Phase shifting is where the waveform's SINE wavefronts are shifted by a constant Phase Angle. E.G. a textbook "Sine Wave" will become a "Cosine Wave" when shifted 90º. It's fairly easy to do with analog Electronics, but not so easy to do digitally.
Why?
Because a shift by degrees is related in phase, not in time. Shifting a single frequency like above is equivalent to a delay of the same SINE--but only for that frequency. That means that to duplicate using standard delay techniques, you have to have a frequency-dependent time delay.

Thankfully, there is a VST plugin that can do the same thing easily:
http://www.savioursofsoul.de/Christian/VST/PhaseBug.zip

Like I said, if this is too much extra work, you can try skipping the phase shift.

Scott

In fact, such operation is pretty commonly used, and easy to obtain in digital domain, thus popular in various DSP applications.
You obtain 90 deg. phase shift of all signal components by applying Hilbert transform on it. This is simply calculating complex spectrum via FFT algorithm.
This will allow you to represent signal spectrum components as complex numbers, which exactly describe phase and amplitude of each bin (cartesian and polar representations are equivelent).
Than, you modify the complex representation of each spectrum bin to preserve amplitude, but to obtain the 90 deg. shift
And, finally you recalculate the signal to time domain via IFFT.
I hope this helps in any way...
-------------

[Cornucopia]

Sorry, you are correct 3dsnar. This is basic FFT dsp stuff.

However, for an End-User, there hasn't been many plugins (like DirX, VST, etc.) which apply those transformations in a straightforward way. That "PhaseBug" plugin is one of the few I could find that did explicitly that, and only that. There are others that use a phase shift as a component in a more complicated algorithm.

Your formula works out very close to what I mentioned (guestimated) earlier, except the coefficients of the rear channels.

(BTW, keep up the good work!)

Scott

[Xoanon]

LS -->(50%L + PhaseInverted 50%R) w/ +90º PhaseShift

Add it to 2 tracks:

1st Track: LS -->100%L, but down by -3dB (so to equal 50%)
2nd Track: LS --> 100%R, Down by -3dB, AND Phase Invert (Yes, you do Invert--Positive waveheight becomes negative and vice-versa)

Do similarly appropriate things for the RS.

Scott

Cheers man. So I dont need to worry about phase shifting? What difference will using it/not using it make?

[3dsnar]

Sorry, you are correct 3dsnar. This is basic FFT dsp stuff.

However, for an End-User, there hasn't been many plugins (like DirX, VST, etc.) which apply those transformations in a straightforward way. That "PhaseBug" plugin is one of the few I could find that did explicitly that, and only that. There are others that use a phase shift as a component in a more complicated algorithm.

Your formula works out very close to what I mentioned (guestimated) earlier, except the coefficients of the rear channels.

(BTW, keep up the good work!)

Scott

Indeed. There aren't much applications for such phase shifting for end users, because this is usually useful as some processing stage of more complex signal operations (i.e. enhancing multichannel panorama, etc.).

Thanx

[3dsnar]

Sorry, you are correct 3dsnar. This is basic FFT dsp stuff.

However, for an End-User, there hasn't been many plugins (like DirX, VST, etc.) which apply those transformations in a straightforward way. That "PhaseBug" plugin is one of the few I could find that did explicitly that, and only that. There are others that use a phase shift as a component in a more complicated algorithm.

Your formula works out very close to what I mentioned (guestimated) earlier, except the coefficients of the rear channels.

(BTW, keep up the good work!)

Scott

Indeed. There aren't much applications for such phase shifting for end users, because this is usually useful as some processing stage of more complex signal operations (i.e. enhancing multichannel panorama, etc.).

Thanx

First of all, I provided incorrect formulas for downmix.
The correct ones are:
Lt = L + 0.7071 C + 0.7071 LFE - 0.866 BL - 0.5 BR
Rt = R + 0.7071 C + 0.7071 LFE + 0.5 BL + 0.866 BR


You must do phase shifting operations. But only by 180 deg,
which is equal to multiplying the signal by -1.
Therefore complex operations are NOT required. Simply apply
the formulas that I suggested to each channel (each sample of each channel), and you will obtain a DPLII downmix.

This formulas include LFE, which is a sort of extention to
what Dolby Labs recommends... But my experiance tells me that it is worth to downmix the LFE as well.

=================

Please note that after multiplying the signals of various channels by the specified weights and adding them to each other will cause clipping. I.e. normalization (of the signal or weights) is necessery.

wgF=1.0
wgC=0.7071
wgLFE=0.7071
wgA=0.0.8944
wgB=0.5
nrm=(wgF + wgC + wgLFE + wgA + wgB)

So the normalized downmixing weights are:
wgF=wgF/nrm
wgC=wgC/nrm
wgLFE=wgLFE/nrm
wgA=wgA/nrm
wgB=wgB/nrm

[Xoanon]

Thats way above my head Im afraid.

[3dsnar]

OK, to simplify.
This is the final (normalized) DPLII downmix set of equations:

Lt = 0.2645 L + 0.1871 C + 0.1871 LFE - 0.2291 BL - 0.1323 BR
Rt = 0.2645 R + 0.1871 C + 0.1871 LFE + 0.1323 BL + 0.2291 BR

Please let me know if this is not clear.

[Cornucopia]

So I dont need to worry about phase shifting? What difference will using it/not using it make?

Try it both ways and see what you like better. IIRC, adding the phase shift increases the ability of a decoder to separate the discreet channels, particularly the surrounds from the mains.

Scott

[Xoanon]

Ok Ive installed that VST thing, I havent looked at it tho.

[mikiem]

"...mix a 5.1 mix into a 2.0 surround mix...
...converting to AC3 Pro Logic..."

FWIW, perhaps use graphedit on the original audio using Cyberlink filters to convert to wav in your choice of Dolby surround formats [options vary according to the version of PowerDVD etc. purchased as well as optional audio packs]. Does the downmixing and AFAIK creates a stereo file with the surround info included, & then use that for ac3 setting encoder to match wav.

Need DTS/AC3 source -> Cyberlink Decoder -> Cyberlink Effects -> Wav Dest -> File Write I think.

[Xoanon]

What are these Cyberlink filters?

[mikiem]

"What are these Cyberlink filters?"

When you install Power DVD, or many programs that use Direct X, the DX portions (often referred to as filters) are available to anything in Windows. Graphedit simply lets you connect them in a chain or chains. In this case the Decoder & Effects I believe are the two portions of Power DVD that would 1st decode the original audio, then downmix/convert to dolby spec.

Cyberlink is popular, already on a lot of systems, and has some excellent audio handling -- there's a bit of info out there about this older method if you want to research it.

[Xoanon]

I might look into that...