AuroraOpenALSoft/Alc/panning.c

441 lines
20 KiB
C
Raw Normal View History

/**
* OpenAL cross platform audio library
* Copyright (C) 1999-2010 by authors.
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
* Or go to http://www.gnu.org/copyleft/lgpl.html
*/
#include "config.h"
#include <math.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <assert.h>
#include "alMain.h"
#include "AL/al.h"
#include "AL/alc.h"
#include "alu.h"
2013-11-04 21:44:46 +00:00
extern inline void SetGains(const ALCdevice *device, ALfloat ingain, ALfloat gains[MaxChannels]);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
static inline void Set0thOrder(ALfloat coeffs[16], ALfloat w)
{
coeffs[0] = w;
}
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
static inline void Set1stOrder(ALfloat coeffs[16], ALfloat w, ALfloat x, ALfloat y, ALfloat z)
{
coeffs[0] = w;
coeffs[1] = x;
coeffs[2] = y;
coeffs[3] = z;
}
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
static inline void Set2ndOrder(ALfloat coeffs[16], ALfloat w, ALfloat x, ALfloat y, ALfloat z, ALfloat r, ALfloat s, ALfloat t, ALfloat u, ALfloat v)
{
coeffs[0] = w;
coeffs[1] = x;
coeffs[2] = y;
coeffs[3] = z;
coeffs[4] = r;
coeffs[5] = s;
coeffs[6] = t;
coeffs[7] = u;
coeffs[8] = v;
}
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
static inline void Set3rdOrder(ALfloat coeffs[16], ALfloat w, ALfloat x, ALfloat y, ALfloat z, ALfloat r, ALfloat s, ALfloat t, ALfloat u, ALfloat v, ALfloat k, ALfloat l, ALfloat m, ALfloat n, ALfloat o, ALfloat p, ALfloat q)
{
coeffs[0] = w;
coeffs[1] = x;
coeffs[2] = y;
coeffs[3] = z;
coeffs[4] = r;
coeffs[5] = s;
coeffs[6] = t;
coeffs[7] = u;
coeffs[8] = v;
coeffs[9] = k;
coeffs[10] = l;
coeffs[11] = m;
coeffs[12] = n;
coeffs[13] = o;
coeffs[14] = p;
coeffs[15] = q;
}
void ComputeAngleGains(const ALCdevice *device, ALfloat angle, ALfloat hwidth, ALfloat ingain, ALfloat gains[MaxChannels])
{
ALfloat tmpgains[MaxChannels] = { 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f };
enum Channel Speaker2Chan[MaxChannels];
ALfloat SpeakerAngle[MaxChannels];
ALfloat langle, rangle;
ALfloat a;
ALuint i;
2014-09-11 00:53:01 +00:00
for(i = 0;i < device->NumSpeakers;i++)
Speaker2Chan[i] = device->Speaker[i].ChanName;
for(i = 0;i < device->NumSpeakers;i++)
SpeakerAngle[i] = device->Speaker[i].Angle;
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
for(i = 0;i < MaxChannels;i++)
gains[i] = 0.0f;
/* Some easy special-cases first... */
2014-09-11 00:53:01 +00:00
if(device->NumSpeakers <= 1 || hwidth >= F_PI)
{
/* Full coverage for all speakers. */
2014-09-11 00:53:01 +00:00
for(i = 0;i < device->NumSpeakers;i++)
{
enum Channel chan = Speaker2Chan[i];
gains[chan] = ingain;
}
return;
}
if(hwidth <= 0.0f)
{
/* Infinitely small sound point. */
2014-09-11 00:53:01 +00:00
for(i = 0;i < device->NumSpeakers-1;i++)
{
if(angle >= SpeakerAngle[i] && angle < SpeakerAngle[i+1])
{
/* Sound is between speakers i and i+1 */
a = (angle-SpeakerAngle[i]) /
(SpeakerAngle[i+1]-SpeakerAngle[i]);
gains[Speaker2Chan[i]] = sqrtf(1.0f-a) * ingain;
gains[Speaker2Chan[i+1]] = sqrtf( a) * ingain;
return;
}
}
/* Sound is between last and first speakers */
if(angle < SpeakerAngle[0])
2013-10-08 23:31:23 +00:00
angle += F_2PI;
a = (angle-SpeakerAngle[i]) /
(F_2PI + SpeakerAngle[0]-SpeakerAngle[i]);
gains[Speaker2Chan[i]] = sqrtf(1.0f-a) * ingain;
gains[Speaker2Chan[0]] = sqrtf( a) * ingain;
return;
}
if(fabsf(angle)+hwidth > F_PI)
{
/* The coverage area would go outside of -pi...+pi. Instead, rotate the
* speaker angles so it would be as if angle=0, and keep them wrapped
* within -pi...+pi. */
if(angle > 0.0f)
{
2013-06-05 08:52:49 +00:00
ALuint done;
ALuint i = 0;
2014-09-11 00:53:01 +00:00
while(i < device->NumSpeakers && device->Speaker[i].Angle-angle < -F_PI)
i++;
2014-09-11 00:53:01 +00:00
for(done = 0;i < device->NumSpeakers;done++)
{
2014-09-11 00:53:01 +00:00
SpeakerAngle[done] = device->Speaker[i].Angle-angle;
Speaker2Chan[done] = device->Speaker[i].ChanName;
i++;
}
2014-09-11 00:53:01 +00:00
for(i = 0;done < device->NumSpeakers;i++)
{
2014-09-11 00:53:01 +00:00
SpeakerAngle[done] = device->Speaker[i].Angle-angle + F_2PI;
Speaker2Chan[done] = device->Speaker[i].ChanName;
done++;
}
}
else
{
/* NOTE: '< device->NumChan' on the iterators is correct here since
* we need to handle index 0. Because the iterators are unsigned,
* they'll underflow and wrap to become 0xFFFFFFFF, which will
* break as expected. */
2013-06-05 08:52:49 +00:00
ALuint done;
2014-09-11 00:53:01 +00:00
ALuint i = device->NumSpeakers-1;
while(i < device->NumSpeakers && device->Speaker[i].Angle-angle > F_PI)
i--;
2014-09-11 00:53:01 +00:00
for(done = device->NumSpeakers-1;i < device->NumSpeakers;done--)
{
2014-09-11 00:53:01 +00:00
SpeakerAngle[done] = device->Speaker[i].Angle-angle;
Speaker2Chan[done] = device->Speaker[i].ChanName;
i--;
}
2014-09-11 00:53:01 +00:00
for(i = device->NumSpeakers-1;done < device->NumSpeakers;i--)
{
2014-09-11 00:53:01 +00:00
SpeakerAngle[done] = device->Speaker[i].Angle-angle - F_2PI;
Speaker2Chan[done] = device->Speaker[i].ChanName;
done--;
}
}
angle = 0.0f;
}
langle = angle - hwidth;
rangle = angle + hwidth;
/* First speaker */
i = 0;
do {
2014-09-11 00:53:01 +00:00
ALuint last = device->NumSpeakers-1;
enum Channel chan = Speaker2Chan[i];
if(SpeakerAngle[i] >= langle && SpeakerAngle[i] <= rangle)
{
tmpgains[chan] = 1.0f;
continue;
}
if(SpeakerAngle[i] < langle && SpeakerAngle[i+1] > langle)
{
a = (langle-SpeakerAngle[i]) /
(SpeakerAngle[i+1]-SpeakerAngle[i]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, 1.0f-a);
}
if(SpeakerAngle[i] > rangle)
{
2013-10-08 23:31:23 +00:00
a = (F_2PI + rangle-SpeakerAngle[last]) /
(F_2PI + SpeakerAngle[i]-SpeakerAngle[last]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, a);
}
else if(SpeakerAngle[last] < rangle)
{
2013-10-08 23:31:23 +00:00
a = (rangle-SpeakerAngle[last]) /
(F_2PI + SpeakerAngle[i]-SpeakerAngle[last]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, a);
}
} while(0);
2014-09-11 00:53:01 +00:00
for(i = 1;i < device->NumSpeakers-1;i++)
{
enum Channel chan = Speaker2Chan[i];
if(SpeakerAngle[i] >= langle && SpeakerAngle[i] <= rangle)
{
tmpgains[chan] = 1.0f;
continue;
}
if(SpeakerAngle[i] < langle && SpeakerAngle[i+1] > langle)
{
a = (langle-SpeakerAngle[i]) /
(SpeakerAngle[i+1]-SpeakerAngle[i]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, 1.0f-a);
}
if(SpeakerAngle[i] > rangle && SpeakerAngle[i-1] < rangle)
{
a = (rangle-SpeakerAngle[i-1]) /
(SpeakerAngle[i]-SpeakerAngle[i-1]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, a);
}
}
/* Last speaker */
2014-09-11 00:53:01 +00:00
i = device->NumSpeakers-1;
do {
enum Channel chan = Speaker2Chan[i];
if(SpeakerAngle[i] >= langle && SpeakerAngle[i] <= rangle)
{
tmpgains[Speaker2Chan[i]] = 1.0f;
continue;
}
if(SpeakerAngle[i] > rangle && SpeakerAngle[i-1] < rangle)
{
a = (rangle-SpeakerAngle[i-1]) /
(SpeakerAngle[i]-SpeakerAngle[i-1]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, a);
}
if(SpeakerAngle[i] < langle)
{
2013-10-08 23:31:23 +00:00
a = (langle-SpeakerAngle[i]) /
(F_2PI + SpeakerAngle[0]-SpeakerAngle[i]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, 1.0f-a);
}
else if(SpeakerAngle[0] > langle)
{
2013-10-08 23:31:23 +00:00
a = (F_2PI + langle-SpeakerAngle[i]) /
(F_2PI + SpeakerAngle[0]-SpeakerAngle[i]);
tmpgains[chan] = lerp(tmpgains[chan], 1.0f, 1.0f-a);
}
} while(0);
2014-09-11 00:53:01 +00:00
for(i = 0;i < device->NumSpeakers;i++)
{
2014-09-11 00:53:01 +00:00
enum Channel chan = device->Speaker[i].ChanName;
gains[chan] = sqrtf(tmpgains[chan]) * ingain;
}
}
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
void ComputeDirectionalGains(const ALCdevice *device, const ALfloat dir[3], ALfloat ingain, ALfloat gains[MaxChannels])
{
ALfloat coeffs[MAX_AMBI_COEFFS];
ALuint i, j;
/* Convert from OpenAL coords to Ambisonics. */
coeffs[0] = 0.7071f; /* sqrt(1.0 / 2.0) */
coeffs[1] = -dir[2]; /* X */
coeffs[2] = -dir[0]; /* Y */
coeffs[3] = dir[1]; /* Z */
coeffs[4] = 0.5f * (3.0f*dir[1]*dir[1] - 1.0f); /* 0.5 * (3*Z*Z - 1) */
coeffs[5] = 2.0f * dir[1] * -dir[2]; /* 2*Z*X */
coeffs[6] = 2.0f * -dir[0] * dir[1]; /* 2*Y*Z */
coeffs[7] = dir[2]*dir[2] - dir[0]*dir[0]; /* X*X - Y*Y */
coeffs[8] = 2.0f * -dir[2] * -dir[0]; /* 2*X*Y */
coeffs[9] = 0.5f * dir[1] * (5.0f*dir[1]*dir[1] - 3.0f); /* 0.5 * Z * (5*Z*Z - 3) */
coeffs[10] = 0.7262f * -dir[2] * (5.0f*dir[1]*dir[1] - 1.0f); /* sqrt(135.0 / 256.0) * X * (5*Z*Z - 1) */
coeffs[11] = 0.7262f * -dir[0] * (5.0f*dir[1]*dir[1] - 1.0f); /* sqrt(135.0 / 256.0) * Y * (5*Z*Z - 1) */
coeffs[12] = 2.5981f * dir[1] * (dir[2]*dir[2] - dir[0]*dir[0]); /* sqrt(27.0 / 4.0) * Z * (X*X - Y*Y) */
coeffs[13] = 5.1962f * -dir[2] * -dir[0] * dir[1]; /* sqrt(27) * X * Y * Z */
coeffs[14] = -dir[2] * (dir[2]*dir[2] - 3.0f*dir[0]*dir[0]); /* X * (X*X - 3*Y*Y) */
coeffs[15] = -dir[0] * (3.0f*dir[2]*dir[2] - dir[0]*dir[0]); /* Y * (3*X*X - Y*Y) */
for(i = 0;i < MaxChannels;i++)
gains[i] = 0.0f;
for(i = 0;i < device->NumSpeakers;i++)
{
enum Channel chan = device->Speaker[i].ChanName;
for(j = 0;j < MAX_AMBI_COEFFS;j++)
gains[chan] += device->Speaker[i].Coeff[j]*coeffs[j];
gains[chan] = maxf(gains[chan], 0.0f) * ingain;
}
}
2014-09-11 00:53:01 +00:00
ALvoid aluInitPanning(ALCdevice *device)
{
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
memset(device->Speaker, 0, sizeof(device->Speaker));
device->NumSpeakers = 0;
2014-09-11 00:53:01 +00:00
switch(device->FmtChans)
{
case DevFmtMono:
2014-09-11 00:53:01 +00:00
device->NumSpeakers = 1;
device->Speaker[0].ChanName = FrontCenter;
device->Speaker[0].Angle = DEG2RAD(0.0f);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
Set0thOrder(device->Speaker[0].Coeff, 1.4142f);
break;
case DevFmtStereo:
2014-09-11 00:53:01 +00:00
device->NumSpeakers = 2;
device->Speaker[0].ChanName = FrontLeft;
device->Speaker[1].ChanName = FrontRight;
device->Speaker[0].Angle = DEG2RAD(-90.0f);
device->Speaker[1].Angle = DEG2RAD( 90.0f);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
Set1stOrder(device->Speaker[0].Coeff, 0.7071f, -0.5f, 0.0f, 0.0f);
Set1stOrder(device->Speaker[1].Coeff, 0.7071f, 0.5f, 0.0f, 0.0f);
break;
case DevFmtQuad:
2014-09-11 00:53:01 +00:00
device->NumSpeakers = 4;
device->Speaker[0].ChanName = BackLeft;
device->Speaker[1].ChanName = FrontLeft;
device->Speaker[2].ChanName = FrontRight;
device->Speaker[3].ChanName = BackRight;
device->Speaker[0].Angle = DEG2RAD(-135.0f);
device->Speaker[1].Angle = DEG2RAD( -45.0f);
device->Speaker[2].Angle = DEG2RAD( 45.0f);
device->Speaker[3].Angle = DEG2RAD( 135.0f);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
Set2ndOrder(device->Speaker[0].Coeff, 0.353543f, -0.306192f, 0.306181f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.117193f);
Set2ndOrder(device->Speaker[1].Coeff, 0.353558f, 0.306181f, 0.306192f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.117183f);
Set2ndOrder(device->Speaker[2].Coeff, 0.353543f, 0.306181f, -0.306192f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.117193f);
Set2ndOrder(device->Speaker[3].Coeff, 0.353558f, -0.306192f, -0.306181f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.117183f);
break;
case DevFmtX51:
2014-09-11 00:53:01 +00:00
device->NumSpeakers = 5;
device->Speaker[0].ChanName = BackLeft;
device->Speaker[1].ChanName = FrontLeft;
device->Speaker[2].ChanName = FrontCenter;
device->Speaker[3].ChanName = FrontRight;
device->Speaker[4].ChanName = BackRight;
device->Speaker[0].Angle = DEG2RAD(-110.0f);
device->Speaker[1].Angle = DEG2RAD( -30.0f);
device->Speaker[2].Angle = DEG2RAD( 0.0f);
device->Speaker[3].Angle = DEG2RAD( 30.0f);
device->Speaker[4].Angle = DEG2RAD( 110.0f);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
Set3rdOrder(device->Speaker[0].Coeff, 0.470934f, -0.369630f, 0.349383f, 0.0f, 0.0f, 0.0f, 0.0f, -0.031379f, -0.058143f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.007116f, -0.043968f);
Set3rdOrder(device->Speaker[1].Coeff, 0.208954f, 0.212846f, 0.238350f, 0.0f, 0.0f, 0.0f, 0.0f, -0.017738f, 0.204014f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.051023f, 0.047490f);
Set3rdOrder(device->Speaker[2].Coeff, 0.109403f, 0.179490f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.142031f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.072024f, -0.000001f);
Set3rdOrder(device->Speaker[3].Coeff, 0.208950f, 0.212842f, -0.238350f, 0.0f, 0.0f, 0.0f, 0.0f, -0.017740f, -0.204011f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.051022f, -0.047489f);
Set3rdOrder(device->Speaker[4].Coeff, 0.470936f, -0.369626f, -0.349386f, 0.0f, 0.0f, 0.0f, 0.0f, -0.031375f, 0.058144f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.007119f, 0.043968f);
break;
case DevFmtX51Side:
2014-09-11 00:53:01 +00:00
device->NumSpeakers = 5;
device->Speaker[0].ChanName = SideLeft;
device->Speaker[1].ChanName = FrontLeft;
device->Speaker[2].ChanName = FrontCenter;
device->Speaker[3].ChanName = FrontRight;
device->Speaker[4].ChanName = SideRight;
device->Speaker[0].Angle = DEG2RAD(-90.0f);
device->Speaker[1].Angle = DEG2RAD(-30.0f);
device->Speaker[2].Angle = DEG2RAD( 0.0f);
device->Speaker[3].Angle = DEG2RAD( 30.0f);
device->Speaker[4].Angle = DEG2RAD( 90.0f);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
Set3rdOrder(device->Speaker[0].Coeff, 0.289151f, -0.081301f, 0.401292f, 0.0f, 0.0f, 0.0f, 0.0f, -0.188208f, -0.071420f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.010099f, -0.032897f);
Set3rdOrder(device->Speaker[1].Coeff, 0.167065f, 0.200583f, 0.172695f, 0.0f, 0.0f, 0.0f, 0.0f, 0.029855f, 0.186407f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.039241f, 0.068910f);
Set3rdOrder(device->Speaker[2].Coeff, 0.109403f, 0.179490f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.142031f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.072024f, -0.000001f);
Set3rdOrder(device->Speaker[3].Coeff, 0.167058f, 0.200580f, -0.172701f, 0.0f, 0.0f, 0.0f, 0.0f, 0.029846f, -0.186405f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.039241f, -0.068904f);
Set3rdOrder(device->Speaker[4].Coeff, 0.289157f, -0.081298f, -0.401295f, 0.0f, 0.0f, 0.0f, 0.0f, -0.188208f, 0.071419f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.010099f, 0.032897f);
break;
case DevFmtX61:
2014-09-11 00:53:01 +00:00
device->NumSpeakers = 6;
device->Speaker[0].ChanName = SideLeft;
device->Speaker[1].ChanName = FrontLeft;
device->Speaker[2].ChanName = FrontCenter;
device->Speaker[3].ChanName = FrontRight;
device->Speaker[4].ChanName = SideRight;
device->Speaker[5].ChanName = BackCenter;
device->Speaker[0].Angle = DEG2RAD(-90.0f);
device->Speaker[1].Angle = DEG2RAD(-30.0f);
device->Speaker[2].Angle = DEG2RAD( 0.0f);
device->Speaker[3].Angle = DEG2RAD( 30.0f);
device->Speaker[4].Angle = DEG2RAD( 90.0f);
device->Speaker[5].Angle = DEG2RAD(180.0f);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
Set3rdOrder(device->Speaker[0].Coeff, 0.289151f, -0.081301f, 0.401292f, 0.0f, 0.0f, 0.0f, 0.0f, -0.188208f, -0.071420f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.010099f, -0.032897f);
Set3rdOrder(device->Speaker[1].Coeff, 0.167065f, 0.200583f, 0.172695f, 0.0f, 0.0f, 0.0f, 0.0f, 0.029855f, 0.186407f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.039241f, 0.068910f);
Set3rdOrder(device->Speaker[2].Coeff, 0.109403f, 0.179490f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.142031f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.072024f, -0.000001f);
Set3rdOrder(device->Speaker[3].Coeff, 0.167058f, 0.200580f, -0.172701f, 0.0f, 0.0f, 0.0f, 0.0f, 0.029846f, -0.186405f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.039241f, -0.068904f);
Set3rdOrder(device->Speaker[4].Coeff, 0.289157f, -0.081298f, -0.401295f, 0.0f, 0.0f, 0.0f, 0.0f, -0.188208f, 0.071419f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.010099f, 0.032897f);
Set3rdOrder(device->Speaker[5].Coeff, 0.353556f, -0.461940f, -0.000006f, 0.0f, 0.0f, 0.0f, 0.0f, 0.165723f, -0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, 0.000005f);
break;
case DevFmtX71:
2014-09-11 00:53:01 +00:00
device->NumSpeakers = 7;
device->Speaker[0].ChanName = BackLeft;
device->Speaker[1].ChanName = SideLeft;
device->Speaker[2].ChanName = FrontLeft;
device->Speaker[3].ChanName = FrontCenter;
device->Speaker[4].ChanName = FrontRight;
device->Speaker[5].ChanName = SideRight;
device->Speaker[6].ChanName = BackRight;
device->Speaker[0].Angle = DEG2RAD(-150.0f);
device->Speaker[1].Angle = DEG2RAD( -90.0f);
device->Speaker[2].Angle = DEG2RAD( -30.0f);
device->Speaker[3].Angle = DEG2RAD( 0.0f);
device->Speaker[4].Angle = DEG2RAD( 30.0f);
device->Speaker[5].Angle = DEG2RAD( 90.0f);
device->Speaker[6].Angle = DEG2RAD( 150.0f);
Use an ambisonics-based panning method For mono sources, third-order ambisonics is utilized to generate panning gains. The general idea is that a panned mono sound can be encoded into b-format ambisonics as: w[i] = sample[i] * 0.7071; x[i] = sample[i] * dir[0]; y[i] = sample[i] * dir[1]; ... and subsequently rendered using: output[chan][i] = w[i] * w_coeffs[chan] + x[i] * x_coeffs[chan] + y[i] * y_coeffs[chan] + ...; By reordering the math, channel gains can be generated by doing: gain[chan] = 0.7071 * w_coeffs[chan] + dir[0] * x_coeffs[chan] + dir[1] * y_coeffs[chan] + ...; which then get applied as normal: output[chan][i] = sample[i] * gain[chan]; One of the reasons to use ambisonics for panning is that it provides arguably better reproduction for sounds emanating from between two speakers. As well, this makes it easier to pan in all 3 dimensions, with for instance a "3D7.1" or 8-channel cube speaker configuration by simply providing the necessary coefficients (this will need some work since some methods still use angle-based panpot, particularly multi-channel sources). Unfortunately, the math to reliably generate the coefficients for a given speaker configuration is too costly to do at run-time. They have to be pre- generated based on a pre-specified speaker arangement, which means the config options for tweaking speaker angles are no longer supportable. Eventually I hope to provide config options for custom coefficients, which can either be generated and written in manually, or via alsoft-config from user-specified speaker positions. The current default set of coefficients were generated using the MATLAB scripts (compatible with GNU Octave) from the excellent Ambisonic Decoder Toolbox, at https://bitbucket.org/ambidecodertoolbox/adt/
2014-09-30 14:33:13 +00:00
Set3rdOrder(device->Speaker[0].Coeff, 0.224752f, -0.295009f, 0.170325f, 0.0f, 0.0f, 0.0f, 0.0f, 0.105349f, -0.182473f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.000000f, 0.065799f);
Set3rdOrder(device->Speaker[1].Coeff, 0.224739f, 0.000002f, 0.340644f, 0.0f, 0.0f, 0.0f, 0.0f, -0.210697f, 0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.000000f, -0.065795f);
Set3rdOrder(device->Speaker[2].Coeff, 0.167065f, 0.200583f, 0.172695f, 0.0f, 0.0f, 0.0f, 0.0f, 0.029855f, 0.186407f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.039241f, 0.068910f);
Set3rdOrder(device->Speaker[3].Coeff, 0.109403f, 0.179490f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.142031f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.072024f, -0.000001f);
Set3rdOrder(device->Speaker[4].Coeff, 0.167058f, 0.200580f, -0.172701f, 0.0f, 0.0f, 0.0f, 0.0f, 0.029846f, -0.186405f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.039241f, -0.068904f);
Set3rdOrder(device->Speaker[5].Coeff, 0.224754f, 0.000004f, -0.340647f, 0.0f, 0.0f, 0.0f, 0.0f, -0.210697f, -0.000004f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.000000f, 0.065796f);
Set3rdOrder(device->Speaker[6].Coeff, 0.224739f, -0.295005f, -0.170331f, 0.0f, 0.0f, 0.0f, 0.0f, 0.105342f, 0.182470f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, -0.000000f, -0.065792f);
break;
}
}