kcat/openal-soft
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
openal-soft/Alc/panning.c

747 lines
28 KiB
C
Raw Normal View History

Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
/**
* 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
Update COPYING to the latest https://www.gnu.org/licenses/old-licenses/lgpl-2.0.txt to fix the FSF' address Fix the FSF' address in the source
2014-08-18 14:11:03 +02:00
* Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
* 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"
Mix to multichannel for effects This mixes to a 4-channel first-order ambisonics buffer. With ACN ordering and N3D scaling, this makes it easy to remain compatible with effects that only care about mono input since channel 0 is an unattenuated mono signal.
2016-01-28 00:02:46 -08:00
#include "alAuxEffectSlot.h"
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
#include "alu.h"
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
#include "bool.h"
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
#include "ambdec.h"
#include "bformatdec.h"
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
Separate calculating ambisonic coefficients from the panning gains
2016-01-25 06:11:51 -08:00
extern inline void CalcXYZCoeffs(ALfloat x, ALfloat y, ALfloat z, ALfloat coeffs[MAX_AMBI_COEFFS]);
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
#define ZERO_ORDER_SCALE 0.0f
#define FIRST_ORDER_SCALE 1.0f
#define SECOND_ORDER_SCALE (1.0f / 1.22474f)
#define THIRD_ORDER_SCALE (1.0f / 1.30657f)
Minor reformating
2015-08-22 07:23:43 -07:00
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
static const ALuint FuMa2ACN[MAX_AMBI_COEFFS] = {
0, /* W */
3, /* X */
1, /* Y */
2, /* Z */
6, /* R */
7, /* S */
5, /* T */
8, /* U */
4, /* V */
12, /* K */
13, /* L */
11, /* M */
14, /* N */
10, /* O */
15, /* P */
9, /* Q */
};
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
/* NOTE: These are scale factors as applied to Ambisonics content. Decoder
* coefficients should be divided by these values to get proper N3D scalings.
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
*/
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
static const ALfloat UnitScale[MAX_AMBI_COEFFS] = {
1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f,
1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f, 1.0f
};
static const ALfloat SN3D2N3DScale[MAX_AMBI_COEFFS] = {
1.000000000f, /* ACN 0 (W), sqrt(1) */
1.732050808f, /* ACN 1 (Y), sqrt(3) */
1.732050808f, /* ACN 2 (Z), sqrt(3) */
1.732050808f, /* ACN 3 (X), sqrt(3) */
2.236067978f, /* ACN 4 (V), sqrt(5) */
2.236067978f, /* ACN 5 (T), sqrt(5) */
2.236067978f, /* ACN 6 (R), sqrt(5) */
2.236067978f, /* ACN 7 (S), sqrt(5) */
2.236067978f, /* ACN 8 (U), sqrt(5) */
2.645751311f, /* ACN 9 (Q), sqrt(7) */
2.645751311f, /* ACN 10 (O), sqrt(7) */
2.645751311f, /* ACN 11 (M), sqrt(7) */
2.645751311f, /* ACN 12 (K), sqrt(7) */
2.645751311f, /* ACN 13 (L), sqrt(7) */
2.645751311f, /* ACN 14 (N), sqrt(7) */
2.645751311f, /* ACN 15 (P), sqrt(7) */
};
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
static const ALfloat FuMa2N3DScale[MAX_AMBI_COEFFS] = {
Use more accurate floating point literals
2015-11-06 10:27:28 -08:00
1.414213562f, /* ACN 0 (W), sqrt(2) */
1.732050808f, /* ACN 1 (Y), sqrt(3) */
1.732050808f, /* ACN 2 (Z), sqrt(3) */
1.732050808f, /* ACN 3 (X), sqrt(3) */
1.936491673f, /* ACN 4 (V), sqrt(15)/2 */
1.936491673f, /* ACN 5 (T), sqrt(15)/2 */
2.236067978f, /* ACN 6 (R), sqrt(5) */
1.936491673f, /* ACN 7 (S), sqrt(15)/2 */
1.936491673f, /* ACN 8 (U), sqrt(15)/2 */
2.091650066f, /* ACN 9 (Q), sqrt(35/8) */
1.972026594f, /* ACN 10 (O), sqrt(35)/3 */
2.231093404f, /* ACN 11 (M), sqrt(224/45) */
2.645751311f, /* ACN 12 (K), sqrt(7) */
2.231093404f, /* ACN 13 (L), sqrt(224/45) */
1.972026594f, /* ACN 14 (N), sqrt(35)/3 */
2.091650066f, /* ACN 15 (P), sqrt(35/8) */
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
};
Separate calculating ambisonic coefficients from the panning gains
2016-01-25 06:11:51 -08:00
void CalcDirectionCoeffs(const ALfloat dir[3], ALfloat coeffs[MAX_AMBI_COEFFS])
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 07:33:13 -07:00
{
/* Convert from OpenAL coords to Ambisonics. */
Store default speaker configurations in a struct
2014-10-02 20:25:30 -07:00
ALfloat x = -dir[2];
ALfloat y = -dir[0];
ALfloat z = dir[1];
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
/* Zeroth-order */
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
coeffs[0] = 1.0f; /* ACN 0 = 1 */
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
/* First-order */
Use more accurate floating point literals
2015-11-06 10:27:28 -08:00
coeffs[1] = 1.732050808f * y; /* ACN 1 = sqrt(3) * Y */
coeffs[2] = 1.732050808f * z; /* ACN 2 = sqrt(3) * Z */
coeffs[3] = 1.732050808f * x; /* ACN 3 = sqrt(3) * X */
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
/* Second-order */
Use more accurate floating point literals
2015-11-06 10:27:28 -08:00
coeffs[4] = 3.872983346f * x * y; /* ACN 4 = sqrt(15) * X * Y */
coeffs[5] = 3.872983346f * y * z; /* ACN 5 = sqrt(15) * Y * Z */
coeffs[6] = 1.118033989f * (3.0f*z*z - 1.0f); /* ACN 6 = sqrt(5)/2 * (3*Z*Z - 1) */
coeffs[7] = 3.872983346f * x * z; /* ACN 7 = sqrt(15) * X * Z */
coeffs[8] = 1.936491673f * (x*x - y*y); /* ACN 8 = sqrt(15)/2 * (X*X - Y*Y) */
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
/* Third-order */
Use more accurate floating point literals
2015-11-06 10:27:28 -08:00
coeffs[9] = 2.091650066f * y * (3.0f*x*x - y*y); /* ACN 9 = sqrt(35/8) * Y * (3*X*X - Y*Y) */
coeffs[10] = 10.246950766f * z * x * y; /* ACN 10 = sqrt(105) * Z * X * Y */
coeffs[11] = 1.620185175f * y * (5.0f*z*z - 1.0f); /* ACN 11 = sqrt(21/8) * Y * (5*Z*Z - 1) */
coeffs[12] = 1.322875656f * z * (5.0f*z*z - 3.0f); /* ACN 12 = sqrt(7)/2 * Z * (5*Z*Z - 3) */
coeffs[13] = 1.620185175f * x * (5.0f*z*z - 1.0f); /* ACN 13 = sqrt(21/8) * X * (5*Z*Z - 1) */
coeffs[14] = 5.123475383f * z * (x*x - y*y); /* ACN 14 = sqrt(105)/2 * Z * (X*X - Y*Y) */
coeffs[15] = 2.091650066f * x * (x*x - 3.0f*y*y); /* ACN 15 = sqrt(35/8) * X * (X*X - 3*Y*Y) */
Separate calculating ambisonic coefficients from the panning gains
2016-01-25 06:11:51 -08:00
}
void CalcAngleCoeffs(ALfloat angle, ALfloat elevation, ALfloat coeffs[MAX_AMBI_COEFFS])
{
ALfloat dir[3] = {
sinf(angle) * cosf(elevation),
sinf(elevation),
-cosf(angle) * cosf(elevation)
};
CalcDirectionCoeffs(dir, coeffs);
}
void ComputeAmbientGains(const ChannelConfig *chancoeffs, ALuint numchans, ALfloat ingain, ALfloat gains[MAX_OUTPUT_CHANNELS])
{
ALuint i;
for(i = 0;i < numchans;i++)
{
// The W coefficients are based on a mathematical average of the
// output. The square root of the base average provides for a more
// perceptual average volume, better suited to non-directional gains.
gains[i] = sqrtf(chancoeffs[i][0]) * ingain;
}
for(;i < MAX_OUTPUT_CHANNELS;i++)
gains[i] = 0.0f;
}
void ComputePanningGains(const ChannelConfig *chancoeffs, ALuint numchans, const ALfloat coeffs[MAX_AMBI_COEFFS], ALfloat ingain, ALfloat gains[MAX_OUTPUT_CHANNELS])
{
ALuint i, j;
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 07:33:13 -07:00
Separate calculating ambisonic coefficients from the panning gains
2016-01-25 06:11:51 -08:00
for(i = 0;i < numchans;i++)
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 07:33:13 -07:00
{
Minor reformating
2015-08-22 07:23:43 -07:00
float gain = 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 07:33:13 -07:00
for(j = 0;j < MAX_AMBI_COEFFS;j++)
Separate calculating ambisonic coefficients from the panning gains
2016-01-25 06:11:51 -08:00
gain += chancoeffs[i][j]*coeffs[j];
Minor reformating
2015-08-22 07:23:43 -07:00
gains[i] = gain * ingain;
Slightly increase the ambient gain volume
2014-11-15 04:11:22 -08:00
}
Loop over the gain values only once
2015-09-24 12:36:02 -07:00
for(;i < MAX_OUTPUT_CHANNELS;i++)
gains[i] = 0.0f;
Use a method to set omni-directional channel gains
2014-11-04 03:33:35 -08:00
}
Rename ComputeBFormatGains to ComputeFirstOrderGains
2016-01-31 09:00:23 -08:00
void ComputeFirstOrderGains(const ChannelConfig *chancoeffs, ALuint numchans, const ALfloat mtx[4], ALfloat ingain, ALfloat gains[MAX_OUTPUT_CHANNELS])
Add preliminary AL_EXT_BFORMAT support Currently missing the AL_ORIENTATION source property. Gain stepping also does not work.
2014-10-31 17:18:45 -07:00
{
Support B-Format source rotation with AL_ORIENTATION
2014-10-31 22:43:13 -07:00
ALuint i, j;
Add preliminary AL_EXT_BFORMAT support Currently missing the AL_ORIENTATION source property. Gain stepping also does not work.
2014-10-31 17:18:45 -07:00
Separate calculating ambisonic coefficients from the panning gains
2016-01-25 06:11:51 -08:00
for(i = 0;i < numchans;i++)
Add preliminary AL_EXT_BFORMAT support Currently missing the AL_ORIENTATION source property. Gain stepping also does not work.
2014-10-31 17:18:45 -07:00
{
Loop over the gain values only once
2015-09-24 12:36:02 -07:00
float gain = 0.0f;
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
for(j = 0;j < 4;j++)
Separate calculating ambisonic coefficients from the panning gains
2016-01-25 06:11:51 -08:00
gain += chancoeffs[i][j] * mtx[j];
Loop over the gain values only once
2015-09-24 12:36:02 -07:00
gains[i] = gain * ingain;
Add preliminary AL_EXT_BFORMAT support Currently missing the AL_ORIENTATION source property. Gain stepping also does not work.
2014-10-31 17:18:45 -07:00
}
Loop over the gain values only once
2015-09-24 12:36:02 -07:00
for(;i < MAX_OUTPUT_CHANNELS;i++)
gains[i] = 0.0f;
Add preliminary AL_EXT_BFORMAT support Currently missing the AL_ORIENTATION source property. Gain stepping also does not work.
2014-10-31 17:18:45 -07:00
}
Use ComputeAngleGains for multi-channel sources and remove the unused lookup table stuff
2012-04-29 05:04:46 -07:00
Mark a function as const
2014-11-15 23:20:38 -08:00
DECL_CONST static inline const char *GetLabelFromChannel(enum Channel channel)
Add a method to convert channel enums to a label string
2014-11-15 00:19:56 -08:00
{
switch(channel)
{
case FrontLeft: return "front-left";
case FrontRight: return "front-right";
case FrontCenter: return "front-center";
case LFE: return "lfe";
case BackLeft: return "back-left";
case BackRight: return "back-right";
case BackCenter: return "back-center";
case SideLeft: return "side-left";
case SideRight: return "side-right";
Use a different method for HRTF mixing This new method mixes sources normally into a 14-channel buffer with the channels placed all around the listener. HRTF is then applied to the channels given their positions and written to a 2-channel buffer, which gets written out to the device. This method has the benefit that HRTF processing becomes more scalable. The costly HRTF filters are applied to the 14-channel buffer after the mix is done, turning it into a post-process with a fixed overhead. Mixing sources is done with normal non-HRTF methods, so increasing the number of playing sources only incurs normal mixing costs. Another benefit is that it improves B-Format playback since the soundfield gets mixed into speakers covering all three dimensions, which then get filtered based on their locations. The main downside to this is that the spatial resolution of the HRTF dataset does not play a big role anymore. However, the hope is that with ambisonics- based panning, the perceptual position of panned sounds will still be good. It is also an option to increase the number of virtual channels for systems that can handle it, or maybe even decrease it for weaker systems.
2014-11-22 04:20:17 -08:00
Use an 8-channel cube for HRTF's virtual format. There were phase issues caused by applying HRTF directly to the B-Format channels, since the HRIR delays were all averaged which removed the inter-aural time-delay, which in turn removed significant spatial information.
2016-02-20 00:53:01 -08:00
case UpperFrontLeft: return "upper-front-left";
case UpperFrontRight: return "upper-front-right";
case UpperBackLeft: return "upper-back-left";
case UpperBackRight: return "upper-back-right";
case LowerFrontLeft: return "lower-front-left";
case LowerFrontRight: return "lower-front-right";
case LowerBackLeft: return "lower-back-left";
case LowerBackRight: return "lower-back-right";
Rename the BFormat channels to Aux
2016-03-16 06:49:35 -07:00
case Aux0: return "aux-0";
case Aux1: return "aux-1";
case Aux2: return "aux-2";
case Aux3: return "aux-3";
Support B-Format output with the wave file writer
2014-11-25 22:20:00 -08:00
Add a method to convert channel enums to a label string
2014-11-15 00:19:56 -08:00
case InvalidChannel: break;
}
return "(unknown)";
}
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
DECL_CONST static const char *GetChannelLayoutName(enum DevFmtChannels chans)
{
switch(chans)
{
case DevFmtMono: return "mono";
case DevFmtStereo: return "stereo";
case DevFmtQuad: return "quad";
case DevFmtX51: return "surround51";
case DevFmtX51Rear: return "surround51rear";
case DevFmtX61: return "surround61";
case DevFmtX71: return "surround71";
case DevFmtBFormat3D:
break;
}
return NULL;
}
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
typedef struct ChannelMap {
enum Channel ChanName;
ChannelConfig Config;
} ChannelMap;
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
static void SetChannelMap(const enum Channel *devchans, ChannelConfig *ambicoeffs,
const ChannelMap *chanmap, size_t count, ALuint *outcount,
ALboolean isfuma)
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
{
Use ALuint instead of size_t for a loop iterator
2015-11-06 02:21:24 -08:00
size_t j, k;
ALuint i;
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
for(i = 0;i < MAX_OUTPUT_CHANNELS && devchans[i] != InvalidChannel;i++)
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
{
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
if(devchans[i] == LFE)
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
{
for(j = 0;j < MAX_AMBI_COEFFS;j++)
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
ambicoeffs[i][j] = 0.0f;
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
continue;
}
for(j = 0;j < count;j++)
{
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
if(devchans[i] != chanmap[j].ChanName)
continue;
if(isfuma)
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
{
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
/* Reformat FuMa -> ACN/N3D */
for(k = 0;k < MAX_AMBI_COEFFS;++k)
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
{
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
ALuint acn = FuMa2ACN[k];
ambicoeffs[i][acn] = chanmap[j].Config[k] / FuMa2N3DScale[acn];
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
}
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
else
{
for(k = 0;k < MAX_AMBI_COEFFS;++k)
ambicoeffs[i][k] = chanmap[j].Config[k];
}
break;
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}
if(j == count)
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
ERR("Failed to match %s channel (%u) in channel map\n", GetLabelFromChannel(devchans[i]), i);
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}
Avoid passing the device to SetChannelMap
2016-01-27 04:44:21 -08:00
*outcount = i;
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
static bool MakeSpeakerMap(ALCdevice *device, const AmbDecConf *conf, ALuint speakermap[MAX_OUTPUT_CHANNELS])
{
ALuint i;
for(i = 0;i < conf->NumSpeakers;i++)
{
int c = -1;
/* NOTE: AmbDec does not define any standard speaker names, however
* for this to work we have to by able to find the output channel
* the speaker definition corresponds to. Therefore, OpenAL Soft
* requires these channel labels to be recognized:
*
* LF = Front left
* RF = Front right
* LS = Side left
* RS = Side right
* LB = Back left
* RB = Back right
* CE = Front center
* CB = Back center
*
* Additionally, surround51 will acknowledge back speakers for side
* channels, and surround51rear will acknowledge side speakers for
* back channels, to avoid issues with an ambdec expecting 5.1 to
* use the side channels when the device is configured for back,
* and vice-versa.
*/
if(al_string_cmp_cstr(conf->Speakers[i].Name, "LF") == 0)
c = GetChannelIdxByName(device->RealOut, FrontLeft);
else if(al_string_cmp_cstr(conf->Speakers[i].Name, "RF") == 0)
c = GetChannelIdxByName(device->RealOut, FrontRight);
else if(al_string_cmp_cstr(conf->Speakers[i].Name, "CE") == 0)
c = GetChannelIdxByName(device->RealOut, FrontCenter);
else if(al_string_cmp_cstr(conf->Speakers[i].Name, "LS") == 0)
{
if(device->FmtChans == DevFmtX51Rear)
c = GetChannelIdxByName(device->RealOut, BackLeft);
else
c = GetChannelIdxByName(device->RealOut, SideLeft);
}
else if(al_string_cmp_cstr(conf->Speakers[i].Name, "RS") == 0)
{
if(device->FmtChans == DevFmtX51Rear)
c = GetChannelIdxByName(device->RealOut, BackRight);
else
c = GetChannelIdxByName(device->RealOut, SideRight);
}
else if(al_string_cmp_cstr(conf->Speakers[i].Name, "LB") == 0)
{
if(device->FmtChans == DevFmtX51)
c = GetChannelIdxByName(device->RealOut, SideLeft);
else
c = GetChannelIdxByName(device->RealOut, BackLeft);
}
else if(al_string_cmp_cstr(conf->Speakers[i].Name, "RB") == 0)
{
if(device->FmtChans == DevFmtX51)
c = GetChannelIdxByName(device->RealOut, SideRight);
else
c = GetChannelIdxByName(device->RealOut, BackRight);
}
else if(al_string_cmp_cstr(conf->Speakers[i].Name, "CB") == 0)
c = GetChannelIdxByName(device->RealOut, BackCenter);
else
{
ERR("AmbDec speaker label \"%s\" not recognized\n",
al_string_get_cstr(conf->Speakers[i].Name));
return false;
}
if(c == -1)
{
ERR("Failed to lookup AmbDec speaker label %s\n",
al_string_get_cstr(conf->Speakers[i].Name));
return false;
}
speakermap[i] = c;
}
return true;
}
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
static bool LoadChannelSetup(ALCdevice *device)
{
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
ChannelMap chanmap[MAX_OUTPUT_CHANNELS];
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
ALuint speakermap[MAX_OUTPUT_CHANNELS];
const ALfloat *coeff_scale = UnitScale;
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
const char *layout = NULL;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ALfloat ambiscale = 1.0f;
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
const char *fname;
AmbDecConf conf;
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
size_t i;
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
layout = GetChannelLayoutName(device->FmtChans);
if(!layout) return false;
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if(!ConfigValueStr(al_string_get_cstr(device->DeviceName), "layouts", layout, &fname))
return false;
ambdec_init(&conf);
if(!ambdec_load(&conf, fname))
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
{
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
ERR("Failed to load layout file %s\n", fname);
goto fail;
}
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if(conf.FreqBands != 1)
{
ERR("AmbDec layout file must be single-band (freq_bands = %u)\n", conf.FreqBands);
goto fail;
}
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if(!MakeSpeakerMap(device, &conf, speakermap))
goto fail;
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if(conf.ChanMask > 0x1ff)
ambiscale = THIRD_ORDER_SCALE;
else if(conf.ChanMask > 0xf)
ambiscale = SECOND_ORDER_SCALE;
else if(conf.ChanMask > 0x1)
ambiscale = FIRST_ORDER_SCALE;
else
ambiscale = 0.0f;
Use ACN ordering for ambisonics coefficients arrays Note that it still uses FuMa scalings internally. Coefficients loaded from config files specify if they're FuMa (in both ordering and scaling) or N3D, and will get reordered or rescaled as needed.
2015-08-28 10:58:30 -07:00
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if(conf.CoeffScale == ADS_SN3D)
coeff_scale = SN3D2N3DScale;
else if(conf.CoeffScale == ADS_FuMa)
coeff_scale = FuMa2N3DScale;
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
for(i = 0;i < conf.NumSpeakers;i++)
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
{
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
ALuint chan = speakermap[i];
ALuint j, k = 0;
for(j = 0;j < MAX_AMBI_COEFFS;j++)
chanmap[i].Config[j] = 0.0f;
chanmap[i].ChanName = device->RealOut.ChannelName[chan];
for(j = 0;j < 1;j++)
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
{
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if((conf.ChanMask&(1<<j)))
chanmap[i].Config[j] = conf.HFMatrix[i][k++] / coeff_scale[j] * conf.HFOrderGain[0];
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
}
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
for(;j < 4;j++)
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
{
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if((conf.ChanMask&(1<<j)))
chanmap[i].Config[j] = conf.HFMatrix[i][k++] / coeff_scale[j] * conf.HFOrderGain[1];
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
}
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
for(;j < 9;j++)
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
{
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
if((conf.ChanMask&(1<<j)))
chanmap[i].Config[j] = conf.HFMatrix[i][k++] / coeff_scale[j] * conf.HFOrderGain[2];
}
for(;j < 16;j++)
{
if((conf.ChanMask&(1<<j)))
chanmap[i].Config[j] = conf.HFMatrix[i][k++] / coeff_scale[j] * conf.HFOrderGain[3];
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
}
}
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
SetChannelMap(device->Dry.ChannelName, device->Dry.AmbiCoeffs, chanmap, conf.NumSpeakers,
&device->Dry.NumChannels, AL_FALSE);
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
device->Dry.AmbiScale = ambiscale;
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
ambdec_deinit(&conf);
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
return true;
Replace the custom layout options with an ambdec loader The ambisonic layout options were never very good for clarity or flexibility. Hopefully using ambdec files will prove to be better.
2016-03-15 08:00:03 -07:00
fail:
ambdec_deinit(&conf);
return false;
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
}
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
ALvoid aluInitPanning(ALCdevice *device)
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
{
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
/* NOTE: These decoder coefficients are using FuMa channel ordering and
* normalization, since that's what was produced by the Ambisonic Decoder
* Toolbox. SetChannelMap will convert them to N3D.
*/
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
static const ChannelMap MonoCfg[1] = {
Use more accurate floating point literals
2015-11-06 10:27:28 -08:00
{ FrontCenter, { 1.414213562f } },
Store default speaker configurations in a struct
2014-10-02 20:25:30 -07:00
}, StereoCfg[2] = {
Use more accurate floating point literals
2015-11-06 10:27:28 -08:00
{ FrontLeft, { 0.707106781f, 0.0f, 0.5f, 0.0f } },
{ FrontRight, { 0.707106781f, 0.0f, -0.5f, 0.0f } },
Store default speaker configurations in a struct
2014-10-02 20:25:30 -07:00
}, QuadCfg[4] = {
Use second-order coefficients for 5.1 Since the panning only drives 5 speakers, it shouldn't use more than 5 ambisonic channels (9 total in second order, minus 4 for unused height).
2016-03-13 03:32:32 -07:00
{ FrontLeft, { 0.353553f, 0.306184f, 0.306184f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, 0.117186f } },
{ FrontRight, { 0.353553f, 0.306184f, -0.306184f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, -0.117186f } },
{ BackLeft, { 0.353553f, -0.306184f, 0.306184f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, -0.117186f } },
{ BackRight, { 0.353553f, -0.306184f, -0.306184f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, 0.117186f } },
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}, X51SideCfg[5] = {
Use second-order coefficients for 5.1 Since the panning only drives 5 speakers, it shouldn't use more than 5 ambisonic channels (9 total in second order, minus 4 for unused height).
2016-03-13 03:32:32 -07:00
{ FrontLeft, { 0.208954f, 0.199518f, 0.223424f, 0.0f, 0.0f, 0.0f, 0.0f, -0.012543f, 0.144260f } },
{ FrontRight, { 0.208950f, 0.199514f, -0.223425f, 0.0f, 0.0f, 0.0f, 0.0f, -0.012544f, -0.144258f } },
{ FrontCenter, { 0.109403f, 0.168250f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.100431f, -0.000001f } },
{ SideLeft, { 0.470934f, -0.346484f, 0.327504f, 0.0f, 0.0f, 0.0f, 0.0f, -0.022188f, -0.041113f } },
{ SideRight, { 0.470936f, -0.346480f, -0.327507f, 0.0f, 0.0f, 0.0f, 0.0f, -0.022186f, 0.041114f } },
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}, X51RearCfg[5] = {
Use second-order coefficients for 5.1 Since the panning only drives 5 speakers, it shouldn't use more than 5 ambisonic channels (9 total in second order, minus 4 for unused height).
2016-03-13 03:32:32 -07:00
{ FrontLeft, { 0.208954f, 0.199518f, 0.223424f, 0.0f, 0.0f, 0.0f, 0.0f, -0.012543f, 0.144260f } },
{ FrontRight, { 0.208950f, 0.199514f, -0.223425f, 0.0f, 0.0f, 0.0f, 0.0f, -0.012544f, -0.144258f } },
{ FrontCenter, { 0.109403f, 0.168250f, -0.000002f, 0.0f, 0.0f, 0.0f, 0.0f, 0.100431f, -0.000001f } },
{ BackLeft, { 0.470934f, -0.346484f, 0.327504f, 0.0f, 0.0f, 0.0f, 0.0f, -0.022188f, -0.041113f } },
{ BackRight, { 0.470936f, -0.346480f, -0.327507f, 0.0f, 0.0f, 0.0f, 0.0f, -0.022186f, 0.041114f } },
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}, X61Cfg[6] = {
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
{ FrontLeft, { 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 } },
{ FrontRight, { 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 } },
{ FrontCenter, { 0.109403f, 0.179490f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.142031f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.072024f, 0.000000f } },
{ BackCenter, { 0.353556f, -0.461940f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.165723f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, 0.000000f } },
{ SideLeft, { 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 } },
{ SideRight, { 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 } },
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
}, X71Cfg[7] = {
Use N3D scaling instead of FuMa
2015-09-23 15:03:53 -07:00
{ FrontLeft, { 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 } },
{ FrontRight, { 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 } },
{ FrontCenter, { 0.109403f, 0.179490f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.142031f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.072024f, 0.000000f } },
{ BackLeft, { 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 } },
{ BackRight, { 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 } },
{ SideLeft, { 0.224739f, 0.000000f, 0.340644f, 0.0f, 0.0f, 0.0f, 0.0f, -0.210697f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, -0.065795f } },
{ SideRight, { 0.224739f, 0.000000f, -0.340644f, 0.0f, 0.0f, 0.0f, 0.0f, -0.210697f, 0.000000f, 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 0.000000f, 0.065795f } },
Use an 8-channel cube for HRTF's virtual format. There were phase issues caused by applying HRTF directly to the B-Format channels, since the HRIR delays were all averaged which removed the inter-aural time-delay, which in turn removed significant spatial information.
2016-02-20 00:53:01 -08:00
}, Cube8Cfg[8] = {
{ UpperFrontLeft, { 0.176776695f, 0.072168784f, 0.072168784f, 0.072168784f } },
{ UpperFrontRight, { 0.176776695f, 0.072168784f, -0.072168784f, 0.072168784f } },
{ UpperBackLeft, { 0.176776695f, -0.072168784f, 0.072168784f, 0.072168784f } },
{ UpperBackRight, { 0.176776695f, -0.072168784f, -0.072168784f, 0.072168784f } },
{ LowerFrontLeft, { 0.176776695f, 0.072168784f, 0.072168784f, -0.072168784f } },
{ LowerFrontRight, { 0.176776695f, 0.072168784f, -0.072168784f, -0.072168784f } },
{ LowerBackLeft, { 0.176776695f, -0.072168784f, 0.072168784f, -0.072168784f } },
{ LowerBackRight, { 0.176776695f, -0.072168784f, -0.072168784f, -0.072168784f } },
Use 2-channel UHJ for stereo output
2016-02-26 16:09:06 -08:00
}, BFormat2D[3] = {
Rename the BFormat channels to Aux
2016-03-16 06:49:35 -07:00
{ Aux0, { 1.0f, 0.0f, 0.0f, 0.0f } },
{ Aux1, { 0.0f, 1.0f, 0.0f, 0.0f } },
{ Aux2, { 0.0f, 0.0f, 1.0f, 0.0f } },
Support B-Format output with the wave file writer
2014-11-25 22:20:00 -08:00
}, BFormat3D[4] = {
Rename the BFormat channels to Aux
2016-03-16 06:49:35 -07:00
{ Aux0, { 1.0f, 0.0f, 0.0f, 0.0f } },
{ Aux1, { 0.0f, 1.0f, 0.0f, 0.0f } },
{ Aux2, { 0.0f, 0.0f, 1.0f, 0.0f } },
{ Aux3, { 0.0f, 0.0f, 0.0f, 1.0f } },
Store default speaker configurations in a struct
2014-10-02 20:25:30 -07:00
};
Initialize a couple variables mingw complains about
2014-11-06 23:05:21 -08:00
const ChannelMap *chanmap = NULL;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ALfloat ambiscale = 1.0f;
Initialize a couple variables mingw complains about
2014-11-06 23:05:21 -08:00
size_t count = 0;
Use 2-channel UHJ for stereo output
2016-02-26 16:09:06 -08:00
ALuint i;
Store default speaker configurations in a struct
2014-10-02 20:25:30 -07:00
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
device->Dry.AmbiScale = 1.0f;
memset(device->Dry.AmbiCoeffs, 0, sizeof(device->Dry.AmbiCoeffs));
device->Dry.NumChannels = 0;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
Use a different method for HRTF mixing This new method mixes sources normally into a 14-channel buffer with the channels placed all around the listener. HRTF is then applied to the channels given their positions and written to a 2-channel buffer, which gets written out to the device. This method has the benefit that HRTF processing becomes more scalable. The costly HRTF filters are applied to the 14-channel buffer after the mix is done, turning it into a post-process with a fixed overhead. Mixing sources is done with normal non-HRTF methods, so increasing the number of playing sources only incurs normal mixing costs. Another benefit is that it improves B-Format playback since the soundfield gets mixed into speakers covering all three dimensions, which then get filtered based on their locations. The main downside to this is that the spatial resolution of the HRTF dataset does not play a big role anymore. However, the hope is that with ambisonics- based panning, the perceptual position of panned sounds will still be good. It is also an option to increase the number of virtual channels for systems that can handle it, or maybe even decrease it for weaker systems.
2014-11-22 04:20:17 -08:00
if(device->Hrtf)
{
Use an 8-channel cube for HRTF's virtual format. There were phase issues caused by applying HRTF directly to the B-Format channels, since the HRIR delays were all averaged which removed the inter-aural time-delay, which in turn removed significant spatial information.
2016-02-20 00:53:01 -08:00
static const struct {
enum Channel Channel;
ALfloat Angle;
ALfloat Elevation;
} CubeInfo[8] = {
{ UpperFrontLeft, DEG2RAD( -45.0f), DEG2RAD( 45.0f) },
{ UpperFrontRight, DEG2RAD( 45.0f), DEG2RAD( 45.0f) },
{ UpperBackLeft, DEG2RAD(-135.0f), DEG2RAD( 45.0f) },
{ UpperBackRight, DEG2RAD( 135.0f), DEG2RAD( 45.0f) },
{ LowerFrontLeft, DEG2RAD( -45.0f), DEG2RAD(-45.0f) },
{ LowerFrontRight, DEG2RAD( 45.0f), DEG2RAD(-45.0f) },
{ LowerBackLeft, DEG2RAD(-135.0f), DEG2RAD(-45.0f) },
{ LowerBackRight, DEG2RAD( 135.0f), DEG2RAD(-45.0f) },
};
Use a different method for HRTF mixing This new method mixes sources normally into a 14-channel buffer with the channels placed all around the listener. HRTF is then applied to the channels given their positions and written to a 2-channel buffer, which gets written out to the device. This method has the benefit that HRTF processing becomes more scalable. The costly HRTF filters are applied to the 14-channel buffer after the mix is done, turning it into a post-process with a fixed overhead. Mixing sources is done with normal non-HRTF methods, so increasing the number of playing sources only incurs normal mixing costs. Another benefit is that it improves B-Format playback since the soundfield gets mixed into speakers covering all three dimensions, which then get filtered based on their locations. The main downside to this is that the spatial resolution of the HRTF dataset does not play a big role anymore. However, the hope is that with ambisonics- based panning, the perceptual position of panned sounds will still be good. It is also an option to increase the number of virtual channels for systems that can handle it, or maybe even decrease it for weaker systems.
2014-11-22 04:20:17 -08:00
Use an 8-channel cube for HRTF's virtual format. There were phase issues caused by applying HRTF directly to the B-Format channels, since the HRIR delays were all averaged which removed the inter-aural time-delay, which in turn removed significant spatial information.
2016-02-20 00:53:01 -08:00
count = COUNTOF(Cube8Cfg);
chanmap = Cube8Cfg;
ambiscale = FIRST_ORDER_SCALE;
Use a different method for HRTF mixing This new method mixes sources normally into a 14-channel buffer with the channels placed all around the listener. HRTF is then applied to the channels given their positions and written to a 2-channel buffer, which gets written out to the device. This method has the benefit that HRTF processing becomes more scalable. The costly HRTF filters are applied to the 14-channel buffer after the mix is done, turning it into a post-process with a fixed overhead. Mixing sources is done with normal non-HRTF methods, so increasing the number of playing sources only incurs normal mixing costs. Another benefit is that it improves B-Format playback since the soundfield gets mixed into speakers covering all three dimensions, which then get filtered based on their locations. The main downside to this is that the spatial resolution of the HRTF dataset does not play a big role anymore. However, the hope is that with ambisonics- based panning, the perceptual position of panned sounds will still be good. It is also an option to increase the number of virtual channels for systems that can handle it, or maybe even decrease it for weaker systems.
2014-11-22 04:20:17 -08:00
for(i = 0;i < count;i++)
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
device->Dry.ChannelName[i] = chanmap[i].ChanName;
Use B-Format for HRTF's virtual output format This adds the ability to directly decode B-Format with HRTF, though only first- order (WXYZ) for now. Second- and third-order would be easilly doable, however we'd need to be able to up-mix first-order content (from the BFORMAT2D and BFORMAT3D buffer formats) since it would be inappropriate to decode lower-order content with a higher-order decoder.
2015-02-09 15:59:29 -08:00
for(;i < MAX_OUTPUT_CHANNELS;i++)
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
device->Dry.ChannelName[i] = InvalidChannel;
SetChannelMap(device->Dry.ChannelName, device->Dry.AmbiCoeffs, chanmap, count,
&device->Dry.NumChannels, AL_TRUE);
device->Dry.AmbiScale = ambiscale;
Calculate HRTF coefficients for all B-Format channels at once It's possible to calculate HRTF coefficients for full third-order ambisonics now, but it's still not possible to use them here without upmixing first-order content.
2015-02-10 11:22:28 -08:00
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
for(i = 0;i < device->Dry.NumChannels;i++)
Use B-Format for HRTF's virtual output format This adds the ability to directly decode B-Format with HRTF, though only first- order (WXYZ) for now. Second- and third-order would be easilly doable, however we'd need to be able to up-mix first-order content (from the BFORMAT2D and BFORMAT3D buffer formats) since it would be inappropriate to decode lower-order content with a higher-order decoder.
2015-02-09 15:59:29 -08:00
{
Generalize GetChannelIdxByName
2016-03-10 14:29:44 -08:00
int chan = GetChannelIdxByName(device->Dry, CubeInfo[i].Channel);
Use an 8-channel cube for HRTF's virtual format. There were phase issues caused by applying HRTF directly to the B-Format channels, since the HRIR delays were all averaged which removed the inter-aural time-delay, which in turn removed significant spatial information.
2016-02-20 00:53:01 -08:00
GetLerpedHrtfCoeffs(device->Hrtf, CubeInfo[i].Elevation, CubeInfo[i].Angle, 1.0f, 1.0f,
device->Hrtf_Params[chan].Coeffs, device->Hrtf_Params[chan].Delay);
Use B-Format for HRTF's virtual output format This adds the ability to directly decode B-Format with HRTF, though only first- order (WXYZ) for now. Second- and third-order would be easilly doable, however we'd need to be able to up-mix first-order content (from the BFORMAT2D and BFORMAT3D buffer formats) since it would be inappropriate to decode lower-order content with a higher-order decoder.
2015-02-09 15:59:29 -08:00
}
Use a different method for HRTF mixing This new method mixes sources normally into a 14-channel buffer with the channels placed all around the listener. HRTF is then applied to the channels given their positions and written to a 2-channel buffer, which gets written out to the device. This method has the benefit that HRTF processing becomes more scalable. The costly HRTF filters are applied to the 14-channel buffer after the mix is done, turning it into a post-process with a fixed overhead. Mixing sources is done with normal non-HRTF methods, so increasing the number of playing sources only incurs normal mixing costs. Another benefit is that it improves B-Format playback since the soundfield gets mixed into speakers covering all three dimensions, which then get filtered based on their locations. The main downside to this is that the spatial resolution of the HRTF dataset does not play a big role anymore. However, the hope is that with ambisonics- based panning, the perceptual position of panned sounds will still be good. It is also an option to increase the number of virtual channels for systems that can handle it, or maybe even decrease it for weaker systems.
2014-11-22 04:20:17 -08:00
return;
}
Use 2-channel UHJ for stereo output
2016-02-26 16:09:06 -08:00
if(device->Uhj_Encoder)
{
count = COUNTOF(BFormat2D);
chanmap = BFormat2D;
ambiscale = FIRST_ORDER_SCALE;
for(i = 0;i < count;i++)
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
device->Dry.ChannelName[i] = chanmap[i].ChanName;
Use 2-channel UHJ for stereo output
2016-02-26 16:09:06 -08:00
for(;i < MAX_OUTPUT_CHANNELS;i++)
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
device->Dry.ChannelName[i] = InvalidChannel;
SetChannelMap(device->Dry.ChannelName, device->Dry.AmbiCoeffs, chanmap, count,
&device->Dry.NumChannels, AL_TRUE);
device->Dry.AmbiScale = ambiscale;
Use 2-channel UHJ for stereo output
2016-02-26 16:09:06 -08:00
return;
}
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
if(device->AmbiDecoder)
{
/* NOTE: This is ACN/N3D ordering and scaling, rather than FuMa. */
static const ChannelMap Ambi3D[4] = {
Rename the BFormat channels to Aux
2016-03-16 06:49:35 -07:00
{ Aux0, { 1.0f, 0.0f, 0.0f, 0.0f } },
{ Aux1, { 0.0f, 1.0f, 0.0f, 0.0f } },
{ Aux2, { 0.0f, 0.0f, 1.0f, 0.0f } },
{ Aux3, { 0.0f, 0.0f, 0.0f, 1.0f } },
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
};
ALuint speakermap[MAX_OUTPUT_CHANNELS];
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
const char *fname = "";
const char *layout;
AmbDecConf conf;
ambdec_init(&conf);
layout = GetChannelLayoutName(device->FmtChans);
if(!layout) goto ambi_fail;
if(!ConfigValueStr(al_string_get_cstr(device->DeviceName), "ambisonics", layout, &fname))
goto ambi_fail;
if(!ambdec_load(&conf, fname))
{
ERR("Failed to load %s\n", fname);
goto ambi_fail;
}
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
if(conf.ChanMask > 0xffff)
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
{
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
ERR("Unsupported channel mask 0x%04x (max 0xffff)\n", conf.ChanMask);
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
goto ambi_fail;
}
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
if(conf.ChanMask > 0xf)
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
{
ERR("Only first-order is supported for HQ decoding (mask 0x%04x, max 0xf)\n",
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
conf.ChanMask);
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
goto ambi_fail;
}
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
if(!MakeSpeakerMap(device, &conf, speakermap))
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
goto ambi_fail;
count = COUNTOF(Ambi3D);
chanmap = Ambi3D;
ambiscale = FIRST_ORDER_SCALE;
for(i = 0;i < count;i++)
device->Dry.ChannelName[i] = chanmap[i].ChanName;
for(;i < MAX_OUTPUT_CHANNELS;i++)
device->Dry.ChannelName[i] = InvalidChannel;
SetChannelMap(device->Dry.ChannelName, device->Dry.AmbiCoeffs, chanmap, count,
&device->Dry.NumChannels, AL_FALSE);
device->Dry.AmbiScale = ambiscale;
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
TRACE("Enabling %s-band ambisonic decoder\n", (conf.FreqBands==1)?"single":"dual");
bformatdec_reset(device->AmbiDecoder, &conf, count, device->Frequency, speakermap);
ambdec_deinit(&conf);
Add a specific output for first-order sources
2016-03-22 17:52:20 -07:00
if(bformatdec_getOrder(device->AmbiDecoder) >= 2)
{
memset(device->FOAOut.AmbiCoeffs, 0, sizeof(device->FOAOut.AmbiCoeffs));
device->FOAOut.AmbiCoeffs[0][0] = 1.0f;
device->FOAOut.AmbiCoeffs[1][1] = 1.0f;
device->FOAOut.AmbiCoeffs[2][2] = 1.0f;
device->FOAOut.AmbiCoeffs[3][3] = 1.0f;
}
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
return;
ambi_fail:
Add config options to enable the hq ambisonic decoder
2016-03-16 01:36:57 -07:00
ambdec_deinit(&conf);
Add a dual-band ambisonic decoder This uses a virtual B-Format buffer for mixing, and then uses a dual-band decoder for improved positional quality. This currently only works with first- order output since first-order input (from the AL_EXT_BFROMAT extension) would not sound correct when fed through a second- or third-order decoder. This also does not currently implement near-field compensation since near-field rendering effects are not implemented.
2016-03-15 05:08:05 -07:00
bformatdec_free(device->AmbiDecoder);
device->AmbiDecoder = NULL;
}
Use a different method for HRTF mixing This new method mixes sources normally into a 14-channel buffer with the channels placed all around the listener. HRTF is then applied to the channels given their positions and written to a 2-channel buffer, which gets written out to the device. This method has the benefit that HRTF processing becomes more scalable. The costly HRTF filters are applied to the 14-channel buffer after the mix is done, turning it into a post-process with a fixed overhead. Mixing sources is done with normal non-HRTF methods, so increasing the number of playing sources only incurs normal mixing costs. Another benefit is that it improves B-Format playback since the soundfield gets mixed into speakers covering all three dimensions, which then get filtered based on their locations. The main downside to this is that the spatial resolution of the HRTF dataset does not play a big role anymore. However, the hope is that with ambisonics- based panning, the perceptual position of panned sounds will still be good. It is also an option to increase the number of virtual channels for systems that can handle it, or maybe even decrease it for weaker systems.
2014-11-22 04:20:17 -08:00
Keep track of the real output's channel names
2016-03-10 01:04:28 -08:00
for(i = 0;i < MAX_OUTPUT_CHANNELS;i++)
device->Dry.ChannelName[i] = device->RealOut.ChannelName[i];
Add the ability to use custom output channel coefficients I'm not sure exactly how I want to do this yet, but this is a good starting point.
2014-11-10 17:53:42 -08:00
if(LoadChannelSetup(device))
return;
Combine some fields into a struct
2014-09-10 17:53:01 -07:00
switch(device->FmtChans)
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
{
Separate device format into 'channel config' and 'sample type' components
2010-12-04 19:50:00 -08:00
case DevFmtMono:
Set gains using the device channel index
2014-11-05 02:54:11 -08:00
count = COUNTOF(MonoCfg);
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
chanmap = MonoCfg;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ambiscale = ZERO_ORDER_SCALE;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
break;
Separate device format into 'channel config' and 'sample type' components
2010-12-04 19:50:00 -08:00
case DevFmtStereo:
Set gains using the device channel index
2014-11-05 02:54:11 -08:00
count = COUNTOF(StereoCfg);
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
chanmap = StereoCfg;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ambiscale = FIRST_ORDER_SCALE;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
break;
Separate device format into 'channel config' and 'sample type' components
2010-12-04 19:50:00 -08:00
case DevFmtQuad:
Set gains using the device channel index
2014-11-05 02:54:11 -08:00
count = COUNTOF(QuadCfg);
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
chanmap = QuadCfg;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ambiscale = SECOND_ORDER_SCALE;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
break;
Separate device format into 'channel config' and 'sample type' components
2010-12-04 19:50:00 -08:00
case DevFmtX51:
Set gains using the device channel index
2014-11-05 02:54:11 -08:00
count = COUNTOF(X51SideCfg);
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
chanmap = X51SideCfg;
Use second-order coefficients for 5.1 Since the panning only drives 5 speakers, it shouldn't use more than 5 ambisonic channels (9 total in second order, minus 4 for unused height).
2016-03-13 03:32:32 -07:00
ambiscale = SECOND_ORDER_SCALE;
Add a 5.1 device format that uses side channels instead of back
2011-05-28 19:35:32 -07:00
break;
Fix 5.1 surround sound Apparently, 5.1 surround sound is supposed to use the "side" channels, not the back channels, and we've been wrong this whole time. That means the "5.1 Side" is actually the correct 5.1 setup, and using the back channels is anomalous. Additionally, this means the 5.1 buffer format should also use the the side channels instead of the back channels. A final note: the 5.1 mixing coefficients are changed so both use the original 5.1 surround sound set (with the surround channels at +/-110 degrees). So the only difference now between 5.1 "side" and 5.1 "back" is the channel labels.
2014-11-07 00:54:16 -08:00
case DevFmtX51Rear:
count = COUNTOF(X51RearCfg);
chanmap = X51RearCfg;
Use second-order coefficients for 5.1 Since the panning only drives 5 speakers, it shouldn't use more than 5 ambisonic channels (9 total in second order, minus 4 for unused height).
2016-03-13 03:32:32 -07:00
ambiscale = SECOND_ORDER_SCALE;
Fix 5.1 surround sound Apparently, 5.1 surround sound is supposed to use the "side" channels, not the back channels, and we've been wrong this whole time. That means the "5.1 Side" is actually the correct 5.1 setup, and using the back channels is anomalous. Additionally, this means the 5.1 buffer format should also use the the side channels instead of the back channels. A final note: the 5.1 mixing coefficients are changed so both use the original 5.1 surround sound set (with the surround channels at +/-110 degrees). So the only difference now between 5.1 "side" and 5.1 "back" is the channel labels.
2014-11-07 00:54:16 -08:00
break;
Separate device format into 'channel config' and 'sample type' components
2010-12-04 19:50:00 -08:00
case DevFmtX61:
Set gains using the device channel index
2014-11-05 02:54:11 -08:00
count = COUNTOF(X61Cfg);
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
chanmap = X61Cfg;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ambiscale = THIRD_ORDER_SCALE;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
break;
Separate device format into 'channel config' and 'sample type' components
2010-12-04 19:50:00 -08:00
case DevFmtX71:
Set gains using the device channel index
2014-11-05 02:54:11 -08:00
count = COUNTOF(X71Cfg);
Remove the channel name from ChannelConfig
2014-11-05 03:46:00 -08:00
chanmap = X71Cfg;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ambiscale = THIRD_ORDER_SCALE;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
break;
Support B-Format output with the wave file writer
2014-11-25 22:20:00 -08:00
case DevFmtBFormat3D:
count = COUNTOF(BFormat3D);
chanmap = BFormat3D;
Don't explicitly store first-order coefficients It seems a simple scaling on the coefficients will allow first-order content to work with second- and third-order coefficients, although obviously not with any improved locality. That may be something to look into for the future, but this is good enough for now.
2015-08-18 07:44:17 -07:00
ambiscale = 1.0f;
Support B-Format output with the wave file writer
2014-11-25 22:20:00 -08:00
break;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
}
Don't bother with LFE in the channel setup, set the values explicitly
2014-11-12 19:26:53 -08:00
Organize the dry buffer properties into a struct
2016-03-09 23:43:57 -08:00
SetChannelMap(device->Dry.ChannelName, device->Dry.AmbiCoeffs, chanmap, count,
&device->Dry.NumChannels, AL_TRUE);
device->Dry.AmbiScale = ambiscale;
Separate speaker/panning initialization into another source file
2010-08-03 00:21:36 -07:00
}
Mix to multichannel for effects This mixes to a 4-channel first-order ambisonics buffer. With ACN ordering and N3D scaling, this makes it easy to remain compatible with effects that only care about mono input since channel 0 is an unattenuated mono signal.
2016-01-28 00:02:46 -08:00
void aluInitEffectPanning(ALeffectslot *slot)
{
static const ChannelMap FirstOrderN3D[4] = {
Rename the BFormat channels to Aux
2016-03-16 06:49:35 -07:00
{ Aux0, { 1.0f, 0.0f, 0.0f, 0.0f } },
{ Aux1, { 0.0f, 1.0f, 0.0f, 0.0f } },
{ Aux2, { 0.0f, 0.0f, 1.0f, 0.0f } },
{ Aux3, { 0.0f, 0.0f, 0.0f, 1.0f } },
Mix to multichannel for effects This mixes to a 4-channel first-order ambisonics buffer. With ACN ordering and N3D scaling, this makes it easy to remain compatible with effects that only care about mono input since channel 0 is an unattenuated mono signal.
2016-01-28 00:02:46 -08:00
};
static const enum Channel AmbiChannels[MAX_OUTPUT_CHANNELS] = {
Rename the BFormat channels to Aux
2016-03-16 06:49:35 -07:00
Aux0, Aux1, Aux2, Aux3, InvalidChannel
Mix to multichannel for effects This mixes to a 4-channel first-order ambisonics buffer. With ACN ordering and N3D scaling, this makes it easy to remain compatible with effects that only care about mono input since channel 0 is an unattenuated mono signal.
2016-01-28 00:02:46 -08:00
};
memset(slot->AmbiCoeffs, 0, sizeof(slot->AmbiCoeffs));
slot->NumChannels = 0;
SetChannelMap(AmbiChannels, slot->AmbiCoeffs, FirstOrderN3D, COUNTOF(FirstOrderN3D),
&slot->NumChannels, AL_FALSE);
}
Reference in New Issue Copy Permalink