FreeRDP/libfreerdp/codec/dsp.c

/**
 * FreeRDP: A Remote Desktop Protocol Implementation
 * Digital Sound Processing
 *
 * Copyright 2010-2011 Vic Lee
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#ifdef HAVE_CONFIG_H
#include "config.h"
#endif

#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include <winpr/crt.h>

#include <freerdp/types.h>
#include <freerdp/log.h>
#include <freerdp/codec/dsp.h>

#if !defined(WITH_DSP_FFMPEG)
#if defined(WITH_GSM)
#include <gsm/gsm.h>
#endif

#if defined(WITH_LAME)
#include <lame/lame.h>
#endif

#if defined(WITH_FAAD2)
#include <neaacdec.h>
#endif

#if defined(WITH_FAAC)
#include <faac.h>
#endif

#if defined(WITH_SOXR)
#include <soxr.h>
#endif

#else
#include "dsp_ffmpeg.h"
#endif

#define TAG FREERDP_TAG("dsp")

#if !defined(WITH_DSP_FFMPEG)

union _ADPCM
{
	struct
	{
		INT16 last_sample[2];
		INT16 last_step[2];
	} ima;
	struct
	{
		BYTE predictor[2];
		INT32 delta[2];
		INT32 sample1[2];
		INT32 sample2[2];
	} ms;
};
typedef union _ADPCM ADPCM;

struct _FREERDP_DSP_CONTEXT
{
	BOOL encoder;

	ADPCM adpcm;
	AUDIO_FORMAT format;

	wStream* buffer;
	wStream* resample;

#if defined(WITH_GSM)
	gsm gsm;
#endif
#if defined(WITH_LAME)
	lame_t lame;
	hip_t hip;
#endif
#if defined(WITH_FAAD2)
	NeAACDecHandle faad;
	BOOL faadSetup;
#endif

#if defined(WITH_FAAC)
	faacEncHandle faac;
	unsigned long faacInputSamples;
	unsigned long faacMaxOutputBytes;
#endif

#if defined(WITH_SOXR)
	soxr_t sox;
#endif
};

static BOOL freerdp_dsp_channel_mix(FREERDP_DSP_CONTEXT* context,
                                    const BYTE* src, size_t size,
                                    const AUDIO_FORMAT* srcFormat,
                                    const BYTE** data, size_t* length)
{
	UINT32 bpp;
	size_t samples;
	size_t x, y;

	if (!context || !data || !length)
		return FALSE;

	if (srcFormat->wFormatTag != WAVE_FORMAT_PCM)
		return FALSE;

	bpp = srcFormat->wBitsPerSample > 8 ? 2 : 1;
	samples = size / bpp;

	if (context->format.nChannels == srcFormat->nChannels)
	{
		*data = src;
		*length = size;
		return TRUE;
	}

	Stream_SetPosition(context->buffer, 0);

	/* Destination has more channels than source */
	if (context->format.nChannels > srcFormat->nChannels)
	{
		switch (srcFormat->nChannels)
		{
			case 1:
				if (!Stream_EnsureCapacity(context->buffer, size * 2))
					return FALSE;

				for (x = 0; x < samples; x++)
				{
					for (y = 0; y < bpp; y++)
						Stream_Write_UINT8(context->buffer, src[x * bpp + y]);

					for (y = 0; y < bpp; y++)
						Stream_Write_UINT8(context->buffer, src[x * bpp + y]);
				}

				Stream_SealLength(context->buffer);
				*data = Stream_Buffer(context->buffer);
				*length = Stream_Length(context->buffer);
				return TRUE;

			case 2: /* We only support stereo, so we can not handle this case. */
			default: /* Unsupported number of channels */
				return FALSE;
		}
	}

	/* Destination has less channels than source */
	switch (srcFormat->nChannels)
	{
		case 2:
			if (!Stream_EnsureCapacity(context->buffer, size / 2))
				return FALSE;

			/* Simply drop second channel.
			 * TODO: Calculate average */
			for (x = 0; x < samples; x++)
			{
				for (y = 0; y < bpp; y++)
					Stream_Write_UINT8(context->buffer, src[2 * x * bpp + y]);
			}

			Stream_SealLength(context->buffer);
			*data = Stream_Buffer(context->buffer);
			*length = Stream_Length(context->buffer);
			return TRUE;

		case 1: /* Invalid, do we want to use a 0 channel sound? */
		default: /* Unsupported number of channels */
			return FALSE;
	}

	return FALSE;
}

/**
 * Microsoft Multimedia Standards Update
 * http://download.microsoft.com/download/9/8/6/9863C72A-A3AA-4DDB-B1BA-CA8D17EFD2D4/RIFFNEW.pdf
 */

static BOOL freerdp_dsp_resample(FREERDP_DSP_CONTEXT* context,
                                 const BYTE* src, size_t size,
                                 const AUDIO_FORMAT* srcFormat,
                                 const BYTE** data, size_t* length)
{
#if defined(WITH_SOXR)
	soxr_error_t error;
	size_t idone, odone;
#endif
	size_t sframes, rframes;
	size_t rsize;
	size_t j;
	size_t sbytes, rbytes;
	size_t srcBytesPerFrame, dstBytesPerFrame;
	size_t srcChannels, dstChannels;
	AUDIO_FORMAT format;

	if (srcFormat->wFormatTag != WAVE_FORMAT_PCM)
	{
		WLog_ERR(TAG, "%s requires %s for sample input, got %s", __FUNCTION__,
		         audio_format_get_tag_string(WAVE_FORMAT_PCM),
		         audio_format_get_tag_string(srcFormat->wFormatTag));
		return FALSE;
	}

	srcChannels = srcFormat->nChannels;
	dstChannels = context->format.nChannels;
	srcBytesPerFrame = (srcFormat->wBitsPerSample > 8) ? 2 : 1;
	dstBytesPerFrame = (context->format.wBitsPerSample > 8) ? 2 : 1;
	/* We want to ignore differences of source and destination format. */
	format = *srcFormat;
	format.wFormatTag = WAVE_FORMAT_UNKNOWN;

	if (audio_format_compatible(&format, &context->format))
		return TRUE;

#if defined(WITH_SOXR)
	sbytes = srcChannels * srcBytesPerFrame;
	sframes = size / sbytes;
	rbytes = dstBytesPerFrame * dstChannels;
	/* Integer rounding correct division */
	rframes = (sframes * context->format.nSamplesPerSec + (srcFormat->nSamplesPerSec + 1) / 2) /
	          srcFormat->nSamplesPerSec;
	rsize = rframes * rbytes;

	if (!Stream_EnsureCapacity(context->resample, rsize))
		return FALSE;

	error = soxr_process(context->sox, src, sframes, &idone,
	                     Stream_Buffer(context->resample),
	                     Stream_Capacity(context->resample) / rbytes, &odone);
	Stream_SetLength(context->resample, odone * rbytes);
	*data = Stream_Buffer(context->resample);
	*length = Stream_Length(context->resample);
	return (error == 0) ? TRUE : FALSE;
#else
	WLog_ERR(TAG, "Missing resample support, recompile -DWITH_SOXR=ON or -DWITH_DSP_FFMPEG=ON");
	return FALSE;
#endif
}

/**
 * Microsoft IMA ADPCM specification:
 *
 * http://wiki.multimedia.cx/index.php?title=Microsoft_IMA_ADPCM
 * http://wiki.multimedia.cx/index.php?title=IMA_ADPCM
 */

static const INT16 ima_step_index_table[] =
{
	-1, -1, -1, -1, 2, 4, 6, 8,
	    -1, -1, -1, -1, 2, 4, 6, 8
    };

static const INT16 ima_step_size_table[] =
{
	7, 8, 9, 10, 11, 12, 13, 14, 16, 17,
	19, 21, 23, 25, 28, 31, 34, 37, 41, 45,
	50, 55, 60, 66, 73, 80, 88, 97, 107, 118,
	130, 143, 157, 173, 190, 209, 230, 253, 279, 307,
	337, 371, 408, 449, 494, 544, 598, 658, 724, 796,
	876, 963, 1060, 1166, 1282, 1411, 1552, 1707, 1878, 2066,
	2272, 2499, 2749, 3024, 3327, 3660, 4026, 4428, 4871, 5358,
	5894, 6484, 7132, 7845, 8630, 9493, 10442, 11487, 12635, 13899,
	15289, 16818, 18500, 20350, 22385, 24623, 27086, 29794, 32767
};

static UINT16 dsp_decode_ima_adpcm_sample(ADPCM* adpcm,
        unsigned int channel, BYTE sample)
{
	INT32 ss;
	INT32 d;
	ss = ima_step_size_table[adpcm->ima.last_step[channel]];
	d = (ss >> 3);

	if (sample & 1)
		d += (ss >> 2);

	if (sample & 2)
		d += (ss >> 1);

	if (sample & 4)
		d += ss;

	if (sample & 8)
		d = -d;

	d += adpcm->ima.last_sample[channel];

	if (d < -32768)
		d = -32768;
	else if (d > 32767)
		d = 32767;

	adpcm->ima.last_sample[channel] = (INT16) d;
	adpcm->ima.last_step[channel] += ima_step_index_table[sample];

	if (adpcm->ima.last_step[channel] < 0)
		adpcm->ima.last_step[channel] = 0;
	else if (adpcm->ima.last_step[channel] > 88)
		adpcm->ima.last_step[channel] = 88;

	return (UINT16) d;
}

static BOOL freerdp_dsp_decode_ima_adpcm(FREERDP_DSP_CONTEXT* context,
        const BYTE* src, size_t size, wStream* out)
{
	BYTE* dst;
	BYTE sample;
	UINT16 decoded;
	UINT32 out_size = size * 4;
	UINT32 channel;
	const UINT32 block_size = context->format.nBlockAlign;
	const UINT32 channels = context->format.nChannels;
	int i;

	if (!Stream_EnsureCapacity(out, out_size))
		return FALSE;

	dst = Stream_Pointer(out);

	while (size > 0)
	{
		if (size % block_size == 0)
		{
			context->adpcm.ima.last_sample[0] = (INT16)(((UINT16)(*src)) | (((UINT16)(*(src + 1))) << 8));
			context->adpcm.ima.last_step[0] = (INT16)(*(src + 2));
			src += 4;
			size -= 4;
			out_size -= 16;

			if (channels > 1)
			{
				context->adpcm.ima.last_sample[1] = (INT16)(((UINT16)(*src)) | (((UINT16)(*(src + 1))) << 8));
				context->adpcm.ima.last_step[1] = (INT16)(*(src + 2));
				src += 4;
				size -= 4;
				out_size -= 16;
			}
		}

		if (channels > 1)
		{
			for (i = 0; i < 8; i++)
			{
				channel = (i < 4 ? 0 : 1);
				sample = ((*src) & 0x0f);
				decoded = dsp_decode_ima_adpcm_sample(&context->adpcm, channel, sample);
				dst[((i & 3) << 3) + (channel << 1)] = (decoded & 0xFF);
				dst[((i & 3) << 3) + (channel << 1) + 1] = (decoded >> 8);
				sample = ((*src) >> 4);
				decoded = dsp_decode_ima_adpcm_sample(&context->adpcm, channel, sample);
				dst[((i & 3) << 3) + (channel << 1) + 4] = (decoded & 0xFF);
				dst[((i & 3) << 3) + (channel << 1) + 5] = (decoded >> 8);
				src++;
			}

			dst += 32;
			size -= 8;
		}
		else
		{
			sample = ((*src) & 0x0f);
			decoded = dsp_decode_ima_adpcm_sample(&context->adpcm, 0, sample);
			*dst++ = (decoded & 0xFF);
			*dst++ = (decoded >> 8);
			sample = ((*src) >> 4);
			decoded = dsp_decode_ima_adpcm_sample(&context->adpcm, 0, sample);
			*dst++ = (decoded & 0xFF);
			*dst++ = (decoded >> 8);
			src++;
			size--;
		}
	}

	Stream_SetPointer(out, dst);
	return TRUE;
}

#if defined(WITH_GSM)
static BOOL freerdp_dsp_decode_gsm610(FREERDP_DSP_CONTEXT* context,
                                      const BYTE* src, size_t size, wStream* out)
{
	size_t offset = 0;

	while (offset < size)
	{
		int rc;
		gsm_signal gsmBlockBuffer[160] = { 0 };
		rc = gsm_decode(context->gsm, (gsm_byte*) &src[offset], gsmBlockBuffer);

		if (rc < 0)
			return FALSE;

		if ((offset % 65) == 0)
			offset += 33;
		else
			offset += 32;

		if (!Stream_EnsureRemainingCapacity(out, sizeof(gsmBlockBuffer)))
			return FALSE;

		Stream_Write(out, (void*) gsmBlockBuffer, sizeof(gsmBlockBuffer));
	}

	return TRUE;
}

static BOOL freerdp_dsp_encode_gsm610(FREERDP_DSP_CONTEXT* context,
                                      const BYTE* src, size_t size, wStream* out)
{
	size_t offset = 0;

	while (offset < size)
	{
		gsm_signal* signal = (gsm_signal*)&src[offset];

		if (!Stream_EnsureRemainingCapacity(out, sizeof(gsm_frame)))
			return FALSE;

		gsm_encode(context->gsm, signal, Stream_Pointer(out));

		if ((offset % 65) == 0)
			Stream_Seek(out, 33);
		else
			Stream_Seek(out, 32);

		offset += 160;
	}

	return TRUE;
}
#endif

#if defined(WITH_LAME)
static BOOL freerdp_dsp_decode_mp3(FREERDP_DSP_CONTEXT* context,
                                   const BYTE* src, size_t size, wStream* out)
{
	int rc, x;
	short* pcm_l;
	short* pcm_r;
	size_t buffer_size;

	if (!context || !src || !out)
		return FALSE;

	buffer_size = 2 * context->format.nChannels * context->format.nSamplesPerSec;

	if (!Stream_EnsureCapacity(context->buffer, 2 * buffer_size))
		return FALSE;

	pcm_l = (short*)Stream_Buffer(context->buffer);
	pcm_r = (short*)Stream_Buffer(context->buffer) + buffer_size;
	rc = hip_decode(context->hip, (unsigned char*)/* API is not modifying content */src,
	                size, pcm_l, pcm_r);

	if (rc <= 0)
		return FALSE;

	if (!Stream_EnsureRemainingCapacity(out, rc * context->format.nChannels * 2))
		return FALSE;

	for (x = 0; x < rc; x++)
	{
		Stream_Write_UINT16(out, pcm_l[x]);
		Stream_Write_UINT16(out, pcm_r[x]);
	}

	return TRUE;
}

static BOOL freerdp_dsp_encode_mp3(FREERDP_DSP_CONTEXT* context,
                                   const BYTE* src, size_t size, wStream* out)
{
	size_t samples_per_channel;
	int rc;

	if (!context || !src || !out)
		return FALSE;

	samples_per_channel = size / context->format.nChannels / context->format.wBitsPerSample / 8;

	/* Ensure worst case buffer size for mp3 stream taken from LAME header */
	if (!Stream_EnsureRemainingCapacity(out, 1.25 * samples_per_channel + 7200))
		return FALSE;

	samples_per_channel = size / 2 /* size of a sample */ / context->format.nChannels;
	rc = lame_encode_buffer_interleaved(context->lame, (short*)src, samples_per_channel,
	                                    Stream_Pointer(out), Stream_GetRemainingCapacity(out));

	if (rc < 0)
		return FALSE;

	Stream_Seek(out, rc);
	return TRUE;
}
#endif

#if defined(WITH_FAAC)
static BOOL freerdp_dsp_encode_faac(FREERDP_DSP_CONTEXT* context,
                                    const BYTE* src, size_t size, wStream* out)
{
	int16_t* inSamples = (int16_t*)src;
	int32_t* outSamples;
	unsigned int bpp;
	unsigned int nrSamples, x;
	int rc;

	if (!context || !src || !out)
		return FALSE;

	bpp = context->format.wBitsPerSample / 8 * context->format.nChannels;
	nrSamples = size / bpp;

	if (!Stream_EnsureCapacity(context->buffer,
	                           nrSamples * sizeof(int32_t) * context->format.nChannels))
		return FALSE;

	if (!Stream_EnsureRemainingCapacity(out, context->faacMaxOutputBytes))
		return FALSE;

	outSamples = (int32_t*)Stream_Buffer(context->buffer);

	for (x = 0; x < nrSamples * context->format.nChannels; x++)
		outSamples[x] = inSamples[x];

	rc = faacEncEncode(context->faac, outSamples, nrSamples * context->format.nChannels,
	                   Stream_Pointer(out), Stream_GetRemainingCapacity(out));

	if (rc < 0)
		return FALSE;
	else if (rc > 0)
		Stream_Seek(out, rc);

	return TRUE;
}
#endif

#if defined(WITH_FAAD2)
static BOOL freerdp_dsp_decode_faad(FREERDP_DSP_CONTEXT* context,
                                    const BYTE* src, size_t size, wStream* out)
{
	NeAACDecFrameInfo info;
	void* output;
	size_t offset = 0;

	if (!context || !src || !out)
		return FALSE;

	if (!context->faadSetup)
	{
		unsigned long samplerate;
		unsigned char channels;
		char err = NeAACDecInit(context->faad, /* API is not modifying content */(unsigned char*)src,
		                        size, &samplerate, &channels);

		if (err != 0)
			return FALSE;

		if (channels != context->format.nChannels)
			return FALSE;

		if (samplerate != context->format.nSamplesPerSec)
			return FALSE;

		context->faadSetup = TRUE;
	}

	while (offset < size)
	{
		size_t outSize;
		void* sample_buffer;
		outSize = context->format.nSamplesPerSec * context->format.nChannels *
		          context->format.wBitsPerSample / 8;

		if (!Stream_EnsureRemainingCapacity(out, outSize))
			return FALSE;

		sample_buffer = Stream_Pointer(out);
		output = NeAACDecDecode2(context->faad, &info, (unsigned char*)&src[offset], size - offset,
		                         &sample_buffer, Stream_GetRemainingCapacity(out));

		if (info.error != 0)
			return FALSE;

		offset += info.bytesconsumed;

		if (info.samples == 0)
			continue;

		Stream_Seek(out, info.samples * context->format.wBitsPerSample / 8);
	}

	return TRUE;
}

#endif

/**
 * 0     1     2     3
 * 2 0   6 4   10 8  14 12   <left>
 *
 * 4     5     6     7
 * 3 1   7 5   11 9  15 13   <right>
 */
static const struct
{
	BYTE byte_num;
	BYTE byte_shift;
} ima_stereo_encode_map[] =
{
	{ 0, 0 },
	{ 4, 0 },
	{ 0, 4 },
	{ 4, 4 },
	{ 1, 0 },
	{ 5, 0 },
	{ 1, 4 },
	{ 5, 4 },
	{ 2, 0 },
	{ 6, 0 },
	{ 2, 4 },
	{ 6, 4 },
	{ 3, 0 },
	{ 7, 0 },
	{ 3, 4 },
	{ 7, 4 }
};

static BYTE dsp_encode_ima_adpcm_sample(ADPCM* adpcm, int channel, INT16 sample)
{
	INT32 e;
	INT32 d;
	INT32 ss;
	BYTE enc;
	INT32 diff;
	ss = ima_step_size_table[adpcm->ima.last_step[channel]];
	d = e = sample - adpcm->ima.last_sample[channel];
	diff = ss >> 3;
	enc = 0;

	if (e < 0)
	{
		enc = 8;
		e = -e;
	}

	if (e >= ss)
	{
		enc |= 4;
		e -= ss;
	}

	ss >>= 1;

	if (e >= ss)
	{
		enc |= 2;
		e -= ss;
	}

	ss >>= 1;

	if (e >= ss)
	{
		enc |= 1;
		e -= ss;
	}

	if (d < 0)
		diff = d + e - diff;
	else
		diff = d - e + diff;

	diff += adpcm->ima.last_sample[channel];

	if (diff < -32768)
		diff = -32768;
	else if (diff > 32767)
		diff = 32767;

	adpcm->ima.last_sample[channel] = (INT16) diff;
	adpcm->ima.last_step[channel] += ima_step_index_table[enc];

	if (adpcm->ima.last_step[channel] < 0)
		adpcm->ima.last_step[channel] = 0;
	else if (adpcm->ima.last_step[channel] > 88)
		adpcm->ima.last_step[channel] = 88;

	return enc;
}

static BOOL freerdp_dsp_encode_ima_adpcm(FREERDP_DSP_CONTEXT* context,
        const BYTE* src, size_t size, wStream* out)
{
	int i;
	BYTE* dst;
	BYTE* start;
	INT16 sample;
	BYTE encoded;
	UINT32 out_size;
	size_t align;
	out_size = size / 2;

	if (!Stream_EnsureRemainingCapacity(out, size))
		return FALSE;

	start = dst = Stream_Pointer(out);
	align = (context->format.nChannels > 1) ? 32 : 4;

	while (size > align)
	{
		if ((dst - start) % context->format.nBlockAlign == 0)
		{
			*dst++ = context->adpcm.ima.last_sample[0] & 0xFF;
			*dst++ = (context->adpcm.ima.last_sample[0] >> 8) & 0xFF;
			*dst++ = (BYTE) context->adpcm.ima.last_step[0];
			*dst++ = 0;

			if (context->format.nChannels > 1)
			{
				*dst++ = context->adpcm.ima.last_sample[1] & 0xFF;
				*dst++ = (context->adpcm.ima.last_sample[1] >> 8) & 0xFF;
				*dst++ = (BYTE) context->adpcm.ima.last_step[1];
				*dst++ = 0;
			}
		}

		if (context->format.nChannels > 1)
		{
			ZeroMemory(dst, 8);

			for (i = 0; i < 16; i++)
			{
				sample = (INT16)(((UINT16)(*src)) | (((UINT16)(*(src + 1))) << 8));
				src += 2;
				encoded = dsp_encode_ima_adpcm_sample(&context->adpcm, i % 2, sample);
				dst[ima_stereo_encode_map[i].byte_num] |= encoded << ima_stereo_encode_map[i].byte_shift;
			}

			dst += 8;
			size -= 32;
		}
		else
		{
			sample = (INT16)(((UINT16)(*src)) | (((UINT16)(*(src + 1))) << 8));
			src += 2;
			encoded = dsp_encode_ima_adpcm_sample(&context->adpcm, 0, sample);
			sample = (INT16)(((UINT16)(*src)) | (((UINT16)(*(src + 1))) << 8));
			src += 2;
			encoded |= dsp_encode_ima_adpcm_sample(&context->adpcm, 0, sample) << 4;
			*dst++ = encoded;
			size -= 4;
		}
	}

	Stream_SetPointer(out, dst);
	return TRUE;
}

/**
 * Microsoft ADPCM Specification:
 *
 * http://wiki.multimedia.cx/index.php?title=Microsoft_ADPCM
 */

static const INT32 ms_adpcm_adaptation_table[] =
{
	230, 230, 230, 230, 307, 409, 512, 614,
	768, 614, 512, 409, 307, 230, 230, 230
};

static const INT32 ms_adpcm_coeffs1[7] =
{
	256, 512, 0, 192, 240, 460, 392
};

static const INT32 ms_adpcm_coeffs2[7] =
{
	0, -256, 0, 64, 0, -208, -232
};

static INLINE INT16 freerdp_dsp_decode_ms_adpcm_sample(ADPCM* adpcm, BYTE sample, int channel)
{
	INT8 nibble;
	INT32 presample;
	nibble = (sample & 0x08 ? (INT8) sample - 16 : sample);
	presample = ((adpcm->ms.sample1[channel] * ms_adpcm_coeffs1[adpcm->ms.predictor[channel]]) +
	             (adpcm->ms.sample2[channel] * ms_adpcm_coeffs2[adpcm->ms.predictor[channel]])) / 256;
	presample += nibble * adpcm->ms.delta[channel];

	if (presample > 32767)
		presample = 32767;
	else if (presample < -32768)
		presample = -32768;

	adpcm->ms.sample2[channel] = adpcm->ms.sample1[channel];
	adpcm->ms.sample1[channel] = presample;
	adpcm->ms.delta[channel] = adpcm->ms.delta[channel] * ms_adpcm_adaptation_table[sample] / 256;

	if (adpcm->ms.delta[channel] < 16)
		adpcm->ms.delta[channel] = 16;

	return (INT16) presample;
}

static BOOL freerdp_dsp_decode_ms_adpcm(FREERDP_DSP_CONTEXT* context,
                                        const BYTE* src, size_t size, wStream* out)
{
	BYTE* dst;
	BYTE sample;
	const UINT32 out_size = size * 4;
	const UINT32 channels = context->format.nChannels;
	const UINT32 block_size = context->format.nBlockAlign;

	if (!Stream_EnsureCapacity(out, out_size))
		return FALSE;

	dst = Stream_Pointer(out);

	while (size > 0)
	{
		if (size % block_size == 0)
		{
			if (channels > 1)
			{
				context->adpcm.ms.predictor[0] = *src++;
				context->adpcm.ms.predictor[1] = *src++;
				context->adpcm.ms.delta[0] = *((INT16*) src);
				src += 2;
				context->adpcm.ms.delta[1] = *((INT16*) src);
				src += 2;
				context->adpcm.ms.sample1[0] = *((INT16*) src);
				src += 2;
				context->adpcm.ms.sample1[1] = *((INT16*) src);
				src += 2;
				context->adpcm.ms.sample2[0] = *((INT16*) src);
				src += 2;
				context->adpcm.ms.sample2[1] = *((INT16*) src);
				src += 2;
				size -= 14;
				*((INT16*) dst) = context->adpcm.ms.sample2[0];
				dst += 2;
				*((INT16*) dst) = context->adpcm.ms.sample2[1];
				dst += 2;
				*((INT16*) dst) = context->adpcm.ms.sample1[0];
				dst += 2;
				*((INT16*) dst) = context->adpcm.ms.sample1[1];
				dst += 2;
			}
			else
			{
				context->adpcm.ms.predictor[0] = *src++;
				context->adpcm.ms.delta[0] = *((INT16*) src);
				src += 2;
				context->adpcm.ms.sample1[0] = *((INT16*) src);
				src += 2;
				context->adpcm.ms.sample2[0] = *((INT16*) src);
				src += 2;
				size -= 7;
				*((INT16*) dst) = context->adpcm.ms.sample2[0];
				dst += 2;
				*((INT16*) dst) = context->adpcm.ms.sample1[0];
				dst += 2;
			}
		}

		if (channels > 1)
		{
			sample = *src++;
			size--;
			*((INT16*) dst) = freerdp_dsp_decode_ms_adpcm_sample(&context->adpcm, sample >> 4, 0);
			dst += 2;
			*((INT16*) dst) = freerdp_dsp_decode_ms_adpcm_sample(&context->adpcm, sample & 0x0F, 1);
			dst += 2;
			sample = *src++;
			size--;
			*((INT16*) dst) = freerdp_dsp_decode_ms_adpcm_sample(&context->adpcm, sample >> 4, 0);
			dst += 2;
			*((INT16*) dst) = freerdp_dsp_decode_ms_adpcm_sample(&context->adpcm, sample & 0x0F, 1);
			dst += 2;
		}
		else
		{
			sample = *src++;
			size--;
			*((INT16*) dst) = freerdp_dsp_decode_ms_adpcm_sample(&context->adpcm, sample >> 4, 0);
			dst += 2;
			*((INT16*) dst) = freerdp_dsp_decode_ms_adpcm_sample(&context->adpcm, sample & 0x0F, 0);
			dst += 2;
		}
	}

	Stream_SetPointer(out, dst);
	return TRUE;
}

static BYTE freerdp_dsp_encode_ms_adpcm_sample(ADPCM* adpcm, INT32 sample, int channel)
{
	INT32 presample;
	INT32 errordelta;
	presample = ((adpcm->ms.sample1[channel] * ms_adpcm_coeffs1[adpcm->ms.predictor[channel]]) +
	             (adpcm->ms.sample2[channel] * ms_adpcm_coeffs2[adpcm->ms.predictor[channel]])) / 256;
	errordelta = (sample - presample) / adpcm->ms.delta[channel];

	if ((sample - presample) % adpcm->ms.delta[channel] > adpcm->ms.delta[channel] / 2)
		errordelta++;

	if (errordelta > 7)
		errordelta = 7;
	else if (errordelta < -8)
		errordelta = -8;

	presample += adpcm->ms.delta[channel] * errordelta;

	if (presample > 32767)
		presample = 32767;
	else if (presample < -32768)
		presample = -32768;

	adpcm->ms.sample2[channel] = adpcm->ms.sample1[channel];
	adpcm->ms.sample1[channel] = presample;
	adpcm->ms.delta[channel] = adpcm->ms.delta[channel] * ms_adpcm_adaptation_table[(((
	                               BYTE) errordelta) & 0x0F)] / 256;

	if (adpcm->ms.delta[channel] < 16)
		adpcm->ms.delta[channel] = 16;

	return ((BYTE) errordelta) & 0x0F;
}

static BOOL freerdp_dsp_encode_ms_adpcm(FREERDP_DSP_CONTEXT* context, const BYTE* src, size_t size,
                                        wStream* out)
{
	BYTE* dst;
	BYTE* start;
	INT32 sample;
	UINT32 out_size;
	const size_t step = 8 + ((context->format.nChannels > 1) ? 4 : 0);
	out_size = size / 2;

	if (!Stream_EnsureRemainingCapacity(out, size))
		return FALSE;

	start = dst = Stream_Pointer(out);

	if (context->adpcm.ms.delta[0] < 16)
		context->adpcm.ms.delta[0] = 16;

	if (context->adpcm.ms.delta[1] < 16)
		context->adpcm.ms.delta[1] = 16;

	while (size >= step)
	{
		if ((dst - start) % context->format.nBlockAlign == 0)
		{
			if (context->format.nChannels > 1)
			{
				*dst++ = context->adpcm.ms.predictor[0];
				*dst++ = context->adpcm.ms.predictor[1];
				*dst++ = (BYTE)(context->adpcm.ms.delta[0] & 0xFF);
				*dst++ = (BYTE)((context->adpcm.ms.delta[0] >> 8) & 0xFF);
				*dst++ = (BYTE)(context->adpcm.ms.delta[1] & 0xFF);
				*dst++ = (BYTE)((context->adpcm.ms.delta[1] >> 8) & 0xFF);
				context->adpcm.ms.sample1[0] = *((INT16*)(src + 4));
				context->adpcm.ms.sample1[1] = *((INT16*)(src + 6));
				context->adpcm.ms.sample2[0] = *((INT16*)(src + 0));
				context->adpcm.ms.sample2[1] = *((INT16*)(src + 2));
				*((INT16*)(dst + 0)) = (INT16) context->adpcm.ms.sample1[0];
				*((INT16*)(dst + 2)) = (INT16) context->adpcm.ms.sample1[1];
				*((INT16*)(dst + 4)) = (INT16) context->adpcm.ms.sample2[0];
				*((INT16*)(dst + 6)) = (INT16) context->adpcm.ms.sample2[1];
				dst += 8;
				src += 8;
				size -= 8;
			}
			else
			{
				*dst++ = context->adpcm.ms.predictor[0];
				*dst++ = (BYTE)(context->adpcm.ms.delta[0] & 0xFF);
				*dst++ = (BYTE)((context->adpcm.ms.delta[0] >> 8) & 0xFF);
				context->adpcm.ms.sample1[0] = *((INT16*)(src + 2));
				context->adpcm.ms.sample2[0] = *((INT16*)(src + 0));
				*((INT16*)(dst + 0)) = (INT16) context->adpcm.ms.sample1[0];
				*((INT16*)(dst + 2)) = (INT16) context->adpcm.ms.sample2[0];
				dst += 4;
				src += 4;
				size -= 4;
			}
		}

		sample = *((INT16*) src);
		src += 2;
		*dst = freerdp_dsp_encode_ms_adpcm_sample(&context->adpcm, sample, 0) << 4;
		sample = *((INT16*) src);
		src += 2;
		*dst += freerdp_dsp_encode_ms_adpcm_sample(&context->adpcm, sample,
		        context->format.nChannels > 1 ? 1 : 0);
		dst++;
		size -= 4;
	}

	Stream_SetPointer(out, dst);
	return TRUE;
}

#endif

FREERDP_DSP_CONTEXT* freerdp_dsp_context_new(BOOL encoder)
{
#if defined(WITH_DSP_FFMPEG)
	return freerdp_dsp_ffmpeg_context_new(encoder);
#else
	FREERDP_DSP_CONTEXT* context = calloc(1, sizeof(FREERDP_DSP_CONTEXT));

	if (!context)
		return NULL;

	context->buffer = Stream_New(NULL, 4096);

	if (!context->buffer)
		goto fail;

	context->resample = Stream_New(NULL, 4096);

	if (!context->resample)
		goto fail;

	context->encoder = encoder;
#if defined(WITH_GSM)
	context->gsm = gsm_create();

	if (!context->gsm)
		goto fail;

	{
		int rc;
		int val = 1;
		rc = gsm_option(context->gsm, GSM_OPT_WAV49, &val);

		if (rc < 0)
			goto fail;
	}
#endif
#if defined(WITH_LAME)

	if (encoder)
	{
		context->lame = lame_init();

		if (!context->lame)
			goto fail;
	}
	else
	{
		context->hip = hip_decode_init();

		if (!context->hip)
			goto fail;
	}

#endif
#if defined(WITH_FAAD2)

	if (!encoder)
	{
		context->faad = NeAACDecOpen();

		if (!context->faad)
			goto fail;
	}

#endif
	return context;
fail:
	freerdp_dsp_context_free(context);
	return NULL;
#endif
}

void freerdp_dsp_context_free(FREERDP_DSP_CONTEXT* context)
{
#if defined(WITH_DSP_FFMPEG)
	freerdp_dsp_ffmpeg_context_free(context);
#else

	if (context)
	{
		Stream_Free(context->buffer, TRUE);
		Stream_Free(context->resample, TRUE);
#if defined(WITH_GSM)
		gsm_destroy(context->gsm);
#endif
#if defined(WITH_LAME)

		if (context->encoder)
			lame_close(context->lame);
		else
			hip_decode_exit(context->hip);

#endif
#if defined(WITH_FAAD2)

		if (!context->encoder)
			NeAACDecClose(context->faad);

#endif
#if defined (WITH_FAAC)

		if (context->faac)
			faacEncClose(context->faac);

#endif
#if defined(WITH_SOXR)
		soxr_delete(context->sox);
#endif
		free(context);
	}

#endif
}

BOOL freerdp_dsp_encode(FREERDP_DSP_CONTEXT* context, const AUDIO_FORMAT* srcFormat,
                        const BYTE* data, size_t length, wStream* out)
{
#if defined(WITH_DSP_FFMPEG)
	return freerdp_dsp_ffmpeg_encode(context, srcFormat, data, length, out);
#else
	const BYTE* resampleData;
	size_t resampleLength;
	AUDIO_FORMAT format = *srcFormat;

	if (!context || !context->encoder || !srcFormat || !data || !out)
		return FALSE;

	if (!freerdp_dsp_channel_mix(context, data, length, srcFormat, &resampleData, &resampleLength))
		return FALSE;

	format.nChannels = context->format.nChannels;

	if (!freerdp_dsp_resample(context, resampleData, resampleLength, &format, &data, &length))
		return FALSE;

	switch (context->format.wFormatTag)
	{
		case WAVE_FORMAT_PCM:
			if (!Stream_EnsureRemainingCapacity(out, length))
				return FALSE;

			Stream_Write(out, data, length);
			return TRUE;

		case WAVE_FORMAT_ADPCM:
			return freerdp_dsp_encode_ms_adpcm(context, data, length, out);

		case WAVE_FORMAT_DVI_ADPCM:
			return freerdp_dsp_encode_ima_adpcm(context, data, length, out);
#if defined(WITH_GSM)

		case WAVE_FORMAT_GSM610:
			return freerdp_dsp_encode_gsm610(context, data, length, out);
#endif
#if defined(WITH_LAME)

		case WAVE_FORMAT_MPEGLAYER3:
			return freerdp_dsp_encode_mp3(context, data, length, out);
#endif
#if defined(WITH_FAAC)

		case WAVE_FORMAT_AAC_MS:
			return freerdp_dsp_encode_faac(context, data, length, out);
#endif

		default:
			return FALSE;
	}

	return FALSE;
#endif
}

BOOL freerdp_dsp_decode(FREERDP_DSP_CONTEXT* context, const AUDIO_FORMAT* srcFormat,
                        const BYTE* data, size_t length, wStream* out)
{
#if defined(WITH_DSP_FFMPEG)
	return freerdp_dsp_ffmpeg_decode(context, srcFormat, data, length, out);
#else

	if (!context || context->encoder || !srcFormat || !data || !out)
		return FALSE;

	switch (context->format.wFormatTag)
	{
		case WAVE_FORMAT_PCM:
			if (!Stream_EnsureRemainingCapacity(out, length))
				return FALSE;

			Stream_Write(out, data, length);
			return TRUE;

		case WAVE_FORMAT_ADPCM:
			return freerdp_dsp_decode_ms_adpcm(context, data, length, out);

		case WAVE_FORMAT_DVI_ADPCM:
			return freerdp_dsp_decode_ima_adpcm(context, data, length, out);
#if defined(WITH_GSM)

		case WAVE_FORMAT_GSM610:
			return freerdp_dsp_decode_gsm610(context, data, length, out);
#endif
#if defined(WITH_LAME)

		case WAVE_FORMAT_MPEGLAYER3:
			return freerdp_dsp_decode_mp3(context, data, length, out);
#endif
#if defined(WITH_FAAD2)

		case WAVE_FORMAT_AAC_MS:
			return freerdp_dsp_decode_faad(context, data, length, out);
#endif

		default:
			return FALSE;
	}

	return FALSE;
#endif
}

BOOL freerdp_dsp_supports_format(const AUDIO_FORMAT* format, BOOL encode)
{
#if defined(WITH_DSP_FFMPEG)
	return freerdp_dsp_ffmpeg_supports_format(format, encode);
#else

	switch (format->wFormatTag)
	{
		case WAVE_FORMAT_PCM:
			return TRUE;
#if defined(WITH_DSP_EXPERIMENTAL)

		case WAVE_FORMAT_ADPCM:
		case WAVE_FORMAT_DVI_ADPCM:
			return TRUE;
#endif
#if defined(WITH_GSM)

		case WAVE_FORMAT_GSM610:
#if defined(WITH_DSP_EXPERIMENTAL)
			return TRUE;
#else
			return !encode;
#endif
#endif
#if defined(WITH_LAME)

		case WAVE_FORMAT_MPEGLAYER3:
#if defined(WITH_DSP_EXPERIMENTAL)
			return TRUE;
#else
			return !encode;
#endif
#endif

		case WAVE_FORMAT_AAC_MS:
#if defined(WITH_FAAD2)
			if (!encode)
				return TRUE;

#endif
#if defined(WITH_FAAC) && defined(WITH_DSP_EXPERIMENTAL)

			if (encode)
				return TRUE;

#endif

		default:
			return FALSE;
	}

	return FALSE;
#endif
}

BOOL freerdp_dsp_context_reset(FREERDP_DSP_CONTEXT* context, const AUDIO_FORMAT* targetFormat)
{
#if defined(WITH_DSP_FFMPEG)
	return freerdp_dsp_ffmpeg_context_reset(context, targetFormat);
#else

	if (!context || !targetFormat)
		return FALSE;

	context->format = *targetFormat;
#if defined(WITH_FAAD2)
	context->faadSetup = FALSE;
#endif
#if defined(WITH_FAAC)

	if (context->encoder)
	{
		faacEncConfigurationPtr cfg;

		if (context->faac)
			faacEncClose(context->faac);

		context->faac = faacEncOpen(targetFormat->nSamplesPerSec, targetFormat->nChannels,
		                            &context->faacInputSamples, &context->faacMaxOutputBytes);

		if (!context->faac)
			return FALSE;

		cfg = faacEncGetCurrentConfiguration(context->faac);
		cfg->bitRate = 10000;
		faacEncSetConfiguration(context->faac, cfg);
	}

#endif
#if defined(WITH_SOXR)
	{
		soxr_io_spec_t iospec = soxr_io_spec(SOXR_INT16, SOXR_INT16);
		soxr_error_t error;
		soxr_delete(context->sox);
		context->sox = soxr_create(context->format.nSamplesPerSec, targetFormat->nSamplesPerSec,
		                           targetFormat->nChannels, &error, &iospec, NULL, NULL);

		if (!context->sox || (error != 0))
			return FALSE;
	}
#endif
	return TRUE;
#endif
}