Csound Csound-dev Csound-tekno Search About

Re: Quad & Discrete 5.1

Date1998-11-25 19:34
FromRichard Dobson
SubjectRe: Quad & Discrete 5.1
I have some general comments around this;

there is a surround sound discussion list (described on Dave Malham's
Ambisonics website: http://www.york.ac.uk/inst/mustech/3d_audio/ambison.htm

they spend most of their time discussing phantom sources, pinnae reflections,
speaker placments and the DVD/DTS fracas, but some useful information emerges
from time to time. Dave Malham also has Csound examples of ambisonic encoding
of sounds with rotational control. His website covers the basic maths of it,
and has pointers to digest lists for surround sound and Ambisonics.

Ambrose Field, a composer at York, has recently completed an electro-acoustic
piece using Ambisonics, and, I belive, now has a commision for  a new piece,
from ICMA. Ambisonics is a much more sophisticated sorround encoding and
diffusion system than raw 5.1, and may be more what you are looking for,
especially if you want to rotate sounds, or soundfields, periphonically.

With regard to multi-channel files, both WAVE and AIFF can in principle be
extenede to any number of channels (Csound will write a quad WAVE file, for
example, if asked). Soundcards  are now appearing which can play such files
(Sonorus Studi/o, Creamware Tdat-16, and their imminent PULSAR card), so you
will not be obliged to manage separate mono files (unless that suits your way
of working, of course!).

A 8-channel opcode was posted to this group some while back, and I recall it
is one John Fitch's list of Things To Do, to addd this to the canonical
sources in some generalized form. I'm not so sure about quad input though.

A precursor to what you describe is the SPACE unit generator which F.R. Moore
provide in his program CMusic (PC version called pcmusic).

5.1 originated as a Cinema surround format, and the centre channel is
typically used for dialogue (localized), rather than for music. I don't think
in itself it would do much to remove 'hole-in-the-middle' effects - plain
stereo constant-power panning can do that already.

I am also working on multi-channel support in CDP -  do feel freee to email me
privately if you want more info on that.


Richard Dobson

tolve wrote:
> 
> Warning: dull laborious, though real world problem ahead!
> 
> let's begin with the goal:
> start with a discrete quad mix consisting of 4 individual (mono) soundfiles (AIFF's) and enhance it by automatically creating a fifth mono soundfile.
> 
> definition:
> 5.1 usually refers to a speaker set-up frequently used for stereo surround sound (not the topic of this post). 3 speakers are positioned in front, 2 in the rear, and the ".1" refers to a subwoofer which handles the low low frequencies. for purposes here, let's forget the subwoofer as an outboard crossover will handle that nicely. what i refer to as the enhanced part is the extra speaker located at middle front -quite useful in large venues to avoid the "hole in the middle" sound effect and for more precise perception of position of sounds. ok if you prefer to call this pentaud or something. just want to make it clear that we're not adding another dimension.
> 
> challenge:
> surprisingly enough, in this case to spotlight csound's superior efficiency against commercial mixing aps including USD20000+ digidesign protools system.
> 
> desired method:
> pull 4 discrete tracks (mono soundfiles) into csound individually using diskin. these would then be used to generate automatically the fifth (center) track (mono soundfile).
> 
> Important: the positioning of the sounds and the manner in which they pan should not be altered by the addition of this extra track (speakers). the result would be, in the case of 5.1, five separate aiff files.
> 
> to further clarify: the levels of the four discrete tracks would be adjusted accordingly and the new fifth file would be created in such a way as to maintain all positioning of sounds in the quad image as they were in the original quad mix.
> 
> Incidentally, if anyone is feeling additionally ambitious, 7.1 adds to the above additional enhancements: 2 speakers, one on each side midway between the front and rear speakers (middle left and middle right). for 7.1, the result would be seven separate aiff files, again positioning of the sounds and the way in which they travel should be the same as in the original quad mix. and in a large hall, keep adding speakers till your budget bursts!
> 
> ok. 5th file generating orc & sco anyone?
> 
> would also welcome, in a separate post, a mathematical response which i will subsequently hope to one day understand.
> 
> and in case someone thinks this is best addressed with a new opcode, don't blame me (smile).
> 

Date1998-11-25 20:07
Fromtolve
SubjectRe: Quad & Discrete 5.1
thanks for your well detailed response richard. perhaps i'll sign on to
that list for a bit.

in the theater space i just left, i was playing back a simple stereo file.
the speakers unfortunately were not well placed for much of the audience
-too high, too far apart, and too far forward (a few rows of the audience
were even behind the speakers!). among the results: that whole in the
middle i mentioned. not just for sounds panning, but all the time. now of
course the speakers could have both been tilted towards the center a bit
but that would have ruined the imaging for a portion of the audience
further back. next time i expect to have more control over placement and
that should help.

for those jumping on this thread let me repeat that i am not referring to
5.1 as it is usually configured for surround sound. this is about taking a
true quad mix and adding an additional speaker between the two in the front
without altering the quad image.

tolve

>I have some general comments around this;
>
>there is a surround sound discussion list (described on Dave Malham's
>Ambisonics website: http://www.york.ac.uk/inst/mustech/3d_audio/ambison.htm
>
>they spend most of their time discussing phantom sources, pinnae reflections,
>speaker placments and the DVD/DTS fracas, but some useful information emerges
>from time to time. Dave Malham also has Csound examples of ambisonic encoding
>of sounds with rotational control. His website covers the basic maths of it,
>and has pointers to digest lists for surround sound and Ambisonics.
>
>Ambrose Field, a composer at York, has recently completed an electro-acoustic
>piece using Ambisonics, and, I belive, now has a commision for  a new piece,
>from ICMA. Ambisonics is a much more sophisticated sorround encoding and
>diffusion system than raw 5.1, and may be more what you are looking for,
>especially if you want to rotate sounds, or soundfields, periphonically.
>
>With regard to multi-channel files, both WAVE and AIFF can in principle be
>extenede to any number of channels (Csound will write a quad WAVE file, for
>example, if asked). Soundcards  are now appearing which can play such files
>(Sonorus Studi/o, Creamware Tdat-16, and their imminent PULSAR card), so you
>will not be obliged to manage separate mono files (unless that suits your way
>of working, of course!).
>
>A 8-channel opcode was posted to this group some while back, and I recall it
>is one John Fitch's list of Things To Do, to addd this to the canonical
>sources in some generalized form. I'm not so sure about quad input though.
>
>A precursor to what you describe is the SPACE unit generator which F.R. Moore
>provide in his program CMusic (PC version called pcmusic).
>
>5.1 originated as a Cinema surround format, and the centre channel is
>typically used for dialogue (localized), rather than for music. I don't think
>in itself it would do much to remove 'hole-in-the-middle' effects - plain
>stereo constant-power panning can do that already.
>
>I am also working on multi-channel support in CDP -  do feel freee to email me
>privately if you want more info on that.
>
>
>Richard Dobson
>
>tolve wrote:
>>
>> Warning: dull laborious, though real world problem ahead!
>>
>> let's begin with the goal:
>> start with a discrete quad mix consisting of 4 individual (mono)
>>soundfiles (AIFF's) and enhance it by automatically creating a fifth mono
>>soundfile.
>>
>> definition:
>> 5.1 usually refers to a speaker set-up frequently used for stereo
>>surround sound (not the topic of this post). 3 speakers are positioned in
>>front, 2 in the rear, and the ".1" refers to a subwoofer which handles
>>the low low frequencies. for purposes here, let's forget the subwoofer as
>>an outboard crossover will handle that nicely. what i refer to as the
>>enhanced part is the extra speaker located at middle front -quite useful
>>in large venues to avoid the "hole in the middle" sound effect and for
>>more precise perception of position of sounds. ok if you prefer to call
>>this pentaud or something. just want to make it clear that we're not
>>adding another dimension.
>>
>> challenge:
>> surprisingly enough, in this case to spotlight csound's superior
>>efficiency against commercial mixing aps including USD20000+ digidesign
>>protools system.
>>
>> desired method:
>> pull 4 discrete tracks (mono soundfiles) into csound individually using
>>diskin. these would then be used to generate automatically the fifth
>>(center) track (mono soundfile).
>>
>> Important: the positioning of the sounds and the manner in which they
>>pan should not be altered by the addition of this extra track (speakers).
>>the result would be, in the case of 5.1, five separate aiff files.
>>
>> to further clarify: the levels of the four discrete tracks would be
>>adjusted accordingly and the new fifth file would be created in such a
>>way as to maintain all positioning of sounds in the quad image as they
>>were in the original quad mix.
>>
>> Incidentally, if anyone is feeling additionally ambitious, 7.1 adds to
>>the above additional enhancements: 2 speakers, one on each side midway
>>between the front and rear speakers (middle left and middle right). for
>>7.1, the result would be seven separate aiff files, again positioning of
>>the sounds and the way in which they travel should be the same as in the
>>original quad mix. and in a large hall, keep adding speakers till your
>>budget bursts!
>>
>> ok. 5th file generating orc & sco anyone?
>>
>> would also welcome, in a separate post, a mathematical response which i
>>will subsequently hope to one day understand.
>>
>> and in case someone thinks this is best addressed with a new opcode,
>>don't blame me (smile).
>>
>> tolve