Skip to content

Fix format converter issues found by paqa_dither #1051

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Fix format converter issues found by paqa_dither #1051

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
102 changes: 58 additions & 44 deletions src/common/pa_converters.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -58,12 +58,14 @@
Int32_To_Int16_Dither, Int32_To_Int8_Dither, Int16_To_Int32
*/

#include <math.h>

#include "pa_converters.h"
#include "pa_dither.h"
#include "pa_endianness.h"
#include "pa_types.h"

#define PA_FLOAT_TO_INT32(x) ((PaInt32) floorf(x))

PaSampleFormat PaUtil_SelectClosestAvailableFormat(
PaSampleFormat availableFormats, PaSampleFormat format )
Expand Down Expand Up @@ -320,13 +322,18 @@ PaUtilConverterTable paConverters = {
#define PA_CLIP_( val, min, max )\
{ val = ((val) < (min)) ? (min) : (((val) > (max)) ? (max) : (val)); }


static const float const_1_div_128_ = 1.0f / 128.0f; /* 8 bit multiplier */

static const float const_1_div_32768_ = 1.0f / 32768.f; /* 16 bit multiplier */

static const double const_1_div_2147483648_ = 1.0 / 2147483648.0; /* 32 bit multiplier */

/* Define half bit constant to cancel the bias caused by truncation when right shifting. */
static const int const_half_bit_16 = 0x4000;

/* Define half bit constant to cancel the bias caused by truncation when calling floorf. */
static const float const_half_bit_float = 0.5f;

/* -------------------------------------------------------------------------- */

static void Float32_To_Int32(
Expand Down Expand Up @@ -365,7 +372,7 @@ static void Float32_To_Int32_Dither(
double dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator );
/* use smaller scaler to prevent overflow when we add the dither */
double dithered = ((double)*src * (2147483646.0)) + dither;
*dest = (PaInt32) dithered;
*dest = PA_FLOAT_TO_INT32( dithered );

src += sourceStride;
dest += destinationStride;
Expand All @@ -388,7 +395,7 @@ static void Float32_To_Int32_Clip(
/* REVIEW */
double scaled = (double)*src * 0x7FFFFFFF;
PA_CLIP_( scaled, -2147483648., 2147483647. );
*dest = (PaInt32) scaled;
*dest = PA_FLOAT_TO_INT32( scaled );

src += sourceStride;
dest += destinationStride;
Expand All @@ -412,7 +419,7 @@ static void Float32_To_Int32_DitherClip(
/* use smaller scaler to prevent overflow when we add the dither */
double dithered = ((double)*src * (2147483646.0)) + dither;
PA_CLIP_( dithered, -2147483648., 2147483647. );
*dest = (PaInt32) dithered;
*dest = PA_FLOAT_TO_INT32( dithered );

src += sourceStride;
dest += destinationStride;
Expand All @@ -436,7 +443,7 @@ static void Float32_To_Int24(
{
/* convert to 32 bit and drop the low 8 bits */
double scaled = (double)(*src) * 2147483647.0;
temp = (PaInt32) scaled;
temp = PA_FLOAT_TO_INT32( scaled );

#if defined(PA_LITTLE_ENDIAN)
dest[0] = (unsigned char)(temp >> 8);
Expand Down Expand Up @@ -472,7 +479,7 @@ static void Float32_To_Int24_Dither(
/* use smaller scaler to prevent overflow when we add the dither */
double dithered = ((double)*src * (2147483646.0)) + dither;

temp = (PaInt32) dithered;
temp = PA_FLOAT_TO_INT32( dithered );

#if defined(PA_LITTLE_ENDIAN)
dest[0] = (unsigned char)(temp >> 8);
Expand Down Expand Up @@ -507,7 +514,7 @@ static void Float32_To_Int24_Clip(
/* convert to 32 bit and drop the low 8 bits */
double scaled = (double)*src * 0x7FFFFFFF;
PA_CLIP_( scaled, -2147483648., 2147483647. );
temp = (PaInt32) scaled;
temp = PA_FLOAT_TO_INT32( scaled );

#if defined(PA_LITTLE_ENDIAN)
dest[0] = (unsigned char)(temp >> 8);
Expand Down Expand Up @@ -544,7 +551,7 @@ static void Float32_To_Int24_DitherClip(
double dithered = ((double)*src * (2147483646.0)) + dither;
PA_CLIP_( dithered, -2147483648., 2147483647. );

temp = (PaInt32) dithered;
temp = PA_FLOAT_TO_INT32( dithered );

#if defined(PA_LITTLE_ENDIAN)
dest[0] = (unsigned char)(temp >> 8);
Expand Down Expand Up @@ -574,7 +581,7 @@ static void Float32_To_Int16(

while( count-- )
{
short samp = (short) (*src * (32767.0f));
short samp = (short) floorf(*src * (32767.0f));
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe use PA_FLOAT_TO_INT32 here.

*dest = samp;

src += sourceStride;
Expand All @@ -594,12 +601,11 @@ static void Float32_To_Int16_Dither(

while( count-- )
{

float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator );
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator ) + const_half_bit_float;
/* use smaller scaler to prevent overflow when we add the dither */
float dithered = (*src * (32766.0f)) + dither;

*dest = (PaInt16) dithered;
*dest = (PaInt16) PA_FLOAT_TO_INT32( dithered );

src += sourceStride;
dest += destinationStride;
Expand All @@ -619,7 +625,7 @@ static void Float32_To_Int16_Clip(

while( count-- )
{
long samp = (PaInt32) (*src * (32767.0f));
long samp = PA_FLOAT_TO_INT32( *src * 32767.0f );

PA_CLIP_( samp, -0x8000, 0x7FFF );
*dest = (PaInt16) samp;
Expand All @@ -638,15 +644,14 @@ static void Float32_To_Int16_DitherClip(
{
float *src = (float*)sourceBuffer;
PaInt16 *dest = (PaInt16*)destinationBuffer;
(void)ditherGenerator; /* unused parameter */

while( count-- )
{

float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator );
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator ) + const_half_bit_float;
/* use smaller scaler to prevent overflow when we add the dither */
float dithered = (*src * (32766.0f)) + dither;
PaInt32 samp = (PaInt32) dithered;
PaInt32 samp = PA_FLOAT_TO_INT32(dithered);
PA_CLIP_( samp, -0x8000, 0x7FFF );
*dest = (PaInt16) samp;

Expand Down Expand Up @@ -688,10 +693,10 @@ static void Float32_To_Int8_Dither(

while( count-- )
{
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator );
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator ) + const_half_bit_float;
/* use smaller scaler to prevent overflow when we add the dither */
float dithered = (*src * (126.0f)) + dither;
PaInt32 samp = (PaInt32) dithered;
PaInt32 samp = PA_FLOAT_TO_INT32( dithered );
*dest = (signed char) samp;

src += sourceStride;
Expand Down Expand Up @@ -734,10 +739,10 @@ static void Float32_To_Int8_DitherClip(

while( count-- )
{
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator );
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator ) + const_half_bit_float;
/* use smaller scaler to prevent overflow when we add the dither */
float dithered = (*src * (126.0f)) + dither;
PaInt32 samp = (PaInt32) dithered;
PaInt32 samp = PA_FLOAT_TO_INT32( dithered );
PA_CLIP_( samp, -0x80, 0x7F );
*dest = (signed char) samp;

Expand Down Expand Up @@ -779,7 +784,7 @@ static void Float32_To_UInt8_Dither(

while( count-- )
{
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator );
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator ) + const_half_bit_float;
/* use smaller scaler to prevent overflow when we add the dither */
float dithered = (*src * (126.0f)) + dither;
PaInt32 samp = (PaInt32) dithered;
Expand Down Expand Up @@ -825,7 +830,7 @@ static void Float32_To_UInt8_DitherClip(

while( count-- )
{
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator );
float dither = PaUtil_GenerateFloatTriangularDither( ditherGenerator ) + const_half_bit_float;
/* use smaller scaler to prevent overflow when we add the dither */
float dithered = (*src * (126.0f)) + dither;
PaInt32 samp = 128 + (PaInt32) dithered;
Expand Down Expand Up @@ -934,19 +939,12 @@ static void Int32_To_Int16_Dither(

while( count-- )
{
/* REVIEW */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator );
#if 1
*dest = (PaInt16) ((((*src)>>1) + dither) >> 15);
#else
/* EXPERIMENTAL force clip after dither. see ticket #112
Clip the intermediate value because adding the dither could cause
a numeric wraparound when shifting.
*/
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16;
PaInt32 temp = (((*src)>>1) + dither);
/* Clip the intermediate value because adding the dither could cause
* a numeric wraparound when shifting. */
PA_CLIP_(temp, (PaInt32) 0xC0000000, (PaInt32) 0x3FFFFFFF);
*dest = (PaInt16) (temp >> 15);
#endif

src += sourceStride;
dest += destinationStride;
Expand Down Expand Up @@ -987,10 +985,14 @@ static void Int32_To_Int8_Dither(
while( count-- )
{
/* increase dither scale to 24-bit value so that it would not be truncated completely when applied */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) << 8;
dither = (PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16) << 8;

/* apply dither, truncate resulting 32-bit value to 8-bit */
*dest = (signed char) ((((*src) >> 1) + dither) >> 23);
PaInt32 temp = (((*src)>>1) + dither);
/* Clip the intermediate value because adding the dither could cause
* a numeric wraparound when shifting. */
PA_CLIP_(temp, (PaInt32) 0xC0000000, (PaInt32) 0x3FFFFFFF);
*dest = (signed char) (temp >> 23);

src += sourceStride;
dest += destinationStride;
Expand Down Expand Up @@ -1032,10 +1034,14 @@ static void Int32_To_UInt8_Dither(
while( count-- )
{
/* increase dither scale to 24-bit value so that it would not be truncated completely when applied */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) << 8;
dither = (PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16) << 8;

/* apply dither, truncate resulting 32-bit value to 8-bit and convert to unsigned */
*dest = (unsigned char) ((((src[0] >> 1) + dither) >> 23) + 128);
PaInt32 temp = (((*src) >> 1) + dither);
/* Clip the intermediate value because adding the dither could cause
* a numeric wraparound when shifting. */
PA_CLIP_(temp, (PaInt32) 0xC0000000, (PaInt32) 0x3FFFFFFF);
*dest = (unsigned char) (temp >> 23) + 128;

src += sourceStride;
dest += destinationStride;
Expand Down Expand Up @@ -1168,7 +1174,7 @@ static void Int24_To_Int16_Dither(
#endif

/* REVIEW */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator );
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16;
*dest = (PaInt16) (((temp >> 1) + dither) >> 15);

src += sourceStride * 3;
Expand Down Expand Up @@ -1232,7 +1238,7 @@ static void Int24_To_Int8_Dither(
#endif

/* increase dither scale to 24-bit value so that it would not be truncated completely when applied */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) << 8;
dither = (PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16) << 8;

/* apply dither, truncate resulting 32-bit value to 8-bit */
*dest = (signed char) (((temp >> 1) + dither) >> 23);
Expand Down Expand Up @@ -1298,7 +1304,7 @@ static void Int24_To_UInt8_Dither(
#endif

/* increase dither scale to 24-bit value so that it would not be truncated completely when applied */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) << 8;
dither = (PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16) << 8;

/* apply dither, truncate resulting 32-bit value to 8-bit and convert to unsigned */
*dest = (unsigned char) ((((temp >> 1) + dither) >> 23) + 128);
Expand Down Expand Up @@ -1423,10 +1429,14 @@ static void Int16_To_Int8_Dither(
temp = ((PaInt32)src[0]) << 16;

/* increase dither scale to 24-bit value so that it would not be truncated completely when applied */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) << 8;
dither = (PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16) << 8;

/* apply dither, truncate resulting 32-bit value to 8-bit */
*dest = (signed char) (((temp >> 1) + dither) >> 23);
temp = ((temp >> 1) + dither);
/* Clip the intermediate value because adding the dither could cause
* a numeric wraparound when shifting. */
PA_CLIP_(temp, (PaInt32) 0xC0000000, (PaInt32) 0x3FFFFFFF);
*dest = (signed char) (temp >> 23);

src += sourceStride;
dest += destinationStride;
Expand Down Expand Up @@ -1471,10 +1481,14 @@ static void Int16_To_UInt8_Dither(
temp = ((PaInt32)src[0]) << 16;

/* increase dither scale to 24-bit value so that it would not be truncated completely when applied */
dither = PaUtil_Generate16BitTriangularDither( ditherGenerator ) << 8;
dither = (PaUtil_Generate16BitTriangularDither( ditherGenerator ) + const_half_bit_16) << 8;

/* apply dither, truncate resulting 32-bit value to 8-bit and convert to unsigned */
*dest = (unsigned char) ((((temp >> 1) + dither) >> 23) + 128);
/* apply dither, truncate resulting 32-bit value to 8-bit */
temp = ((temp >> 1) + dither);
/* Clip the intermediate value because adding the dither could cause
* a numeric wraparound when shifting. */
PA_CLIP_(temp, (PaInt32) 0xC0000000, (PaInt32) 0x3FFFFFFF);
*dest = (unsigned char) ((temp >> 23) + 128);

src += sourceStride;
dest += destinationStride;
Expand Down
Loading