Integrate with FFmpeg decode filter audio demo (#118)

Signed-off-by: begeekmyfriend <begeekmyfriend@gmail.com>
Co-authored-by: begeekmyfriend <begeekmyfriend@gmail.com>

ffmpeg-examples/Makefile

@@ -0,0 +1,52 @@
+CC=g++
+GDB ?= FALSE
+
+# use pkg-config for getting CFLAGS and LDLIBS
+SHARED_LIBS=libavdevice                          \
+            libavformat                          \
+            libavfilter                          \
+            libavcodec                           \
+            libswresample                        \
+            libswscale                           \
+            libavutil
+
+ifeq ($(GDB), TRUE)
+	OPTFLAG += -g
+endif
+
+CFLAGS := $(shell pkg-config --cflags $(SHARED_LIBS)) -I.. -Wall -std=c++11 -fopenmp ${OPTFLAG}
+LDLIBS := $(shell pkg-config --libs $(SHARED_LIBS)) -L../build/lib -lsherpa-ncnn-c-api -lsherpa-ncnn-core -lkaldi-native-fbank-core -lncnn -lportaudio -lpthread -lm
+
+#Get libavutil version and extract major, minor and micro
+LIBAVUTIL_VERSION := $(shell pkg-config --modversion libavutil)
+LIBAVUTIL_MAJOR := $(shell echo "$(LIBAVUTIL_VERSION)" | awk -F. '{print $$1}')
+LIBAVUTIL_MINOR := $(shell echo "$(LIBAVUTIL_VERSION)" | awk -F. '{print $$2}')
+LIBAVUTIL_MICRO := $(shell echo "$(LIBAVUTIL_VERSION)" | awk -F. '{print $$3}')
+#Check if libavutil version is 57.28.100 or above
+FFMPEG_51_AND_ABOVE = $(shell echo "$(LIBAVUTIL_MAJOR) $(LIBAVUTIL_MINOR) $(LIBAVUTIL_MICRO)" | awk '{if ($$1 > 57 || ($$1 == 57 && $$2 > 28) || ($$1 == 57 && $$2 == 28 && $$3 >= 100)) print "TRUE"; else print "FALSE"}')
+ifeq ($(FFMPEG_51_AND_ABOVE), FALSE)
+	$(error FFmpeg version should be n5.1 or above!)
+endif
+
+EXAMPLES=sherpa-ncnn-ffmpeg
+
+OBJS=$(addsuffix .o,$(EXAMPLES))
+
+.phony: all clean
+
+all: $(EXAMPLES)
+	@echo $(EXAMPLES)
+	$(RM) $(OBJS)
+
+$(EXAMPLES): $(OBJS)
+	$(CC) $(addsuffix .o,$@) $(CFLAGS) $(LDLIBS) -o $@
+
+%.o : %.c
+	${CC} ${CFLAGS} -c -o $@ $<
+
+clean:
+	$(RM) $(EXAMPLES) $(OBJS)
+    
+build_info:
+	@echo "libavutil version: $(LIBAVUTIL_VERSION)"
+	@echo "Supported examples: $(EXAMPLES)"

ffmpeg-examples/sherpa-ncnn-ffmpeg.c

@@ -0,0 +1,402 @@
+/**
+ * Copyright (c)  2023  Xiaomi Corporation (authors: Fangjun Kuang)
+ *
+ * See LICENSE for clarification regarding multiple authors
+ *
+ * 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.
+ */
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "sherpa-ncnn/c-api/c-api.h"
+
+
+/*
+ * Copyright (c) 2010 Nicolas George
+ * Copyright (c) 2011 Stefano Sabatini
+ * Copyright (c) 2012 Clément Bœsch
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+/**
+ * @file audio decoding and filtering usage example
+ * @example decode_filter_audio.c
+ *
+ * Demux, decode and filter audio input file, generate a raw audio
+ * file to be played with ffplay.
+ */
+
+#include <unistd.h>
+extern "C" {
+#include <libavcodec/avcodec.h>
+#include <libavformat/avformat.h>
+#include <libavfilter/buffersink.h>
+#include <libavfilter/buffersrc.h>
+#include <libavutil/channel_layout.h>
+#include <libavutil/opt.h>
+}
+
+static const char *filter_descr = "aresample=16000,aformat=sample_fmts=s16:channel_layouts=mono";
+
+static AVFormatContext *fmt_ctx;
+static AVCodecContext *dec_ctx;
+AVFilterContext *buffersink_ctx;
+AVFilterContext *buffersrc_ctx;
+AVFilterGraph *filter_graph;
+static int audio_stream_index = -1;
+
+static int open_input_file(const char *filename)
+{
+    const AVCodec *dec;
+    int ret;
+
+    if ((ret = avformat_open_input(&fmt_ctx, filename, NULL, NULL)) < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot open input file %s\n", filename);
+        return ret;
+    }
+
+    if ((ret = avformat_find_stream_info(fmt_ctx, NULL)) < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot find stream information\n");
+        return ret;
+    }
+
+    /* select the audio stream */
+    ret = av_find_best_stream(fmt_ctx, AVMEDIA_TYPE_AUDIO, -1, -1, &dec, 0);
+    if (ret < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot find an audio stream in the input file\n");
+        return ret;
+    }
+    audio_stream_index = ret;
+
+    /* create decoding context */
+    dec_ctx = avcodec_alloc_context3(dec);
+    if (!dec_ctx)
+        return AVERROR(ENOMEM);
+    avcodec_parameters_to_context(dec_ctx, fmt_ctx->streams[audio_stream_index]->codecpar);
+
+    /* init the audio decoder */
+    if ((ret = avcodec_open2(dec_ctx, dec, NULL)) < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot open audio decoder\n");
+        return ret;
+    }
+
+    return 0;
+}
+
+static int init_filters(const char *filters_descr)
+{
+    char args[512];
+    int ret = 0;
+    const AVFilter *abuffersrc  = avfilter_get_by_name("abuffer");
+    const AVFilter *abuffersink = avfilter_get_by_name("abuffersink");
+    AVFilterInOut *outputs = avfilter_inout_alloc();
+    AVFilterInOut *inputs  = avfilter_inout_alloc();
+    static const enum AVSampleFormat out_sample_fmts[] = { AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE };
+    static const int out_sample_rates[] = { 16000, -1 };
+    const AVFilterLink *outlink;
+    AVRational time_base = fmt_ctx->streams[audio_stream_index]->time_base;
+
+    filter_graph = avfilter_graph_alloc();
+    if (!outputs || !inputs || !filter_graph) {
+        ret = AVERROR(ENOMEM);
+        goto end;
+    }
+
+    /* buffer audio source: the decoded frames from the decoder will be inserted here. */
+    if (dec_ctx->ch_layout.order == AV_CHANNEL_ORDER_UNSPEC)
+        av_channel_layout_default(&dec_ctx->ch_layout, dec_ctx->ch_layout.nb_channels);
+    ret = snprintf(args, sizeof(args),
+            "time_base=%d/%d:sample_rate=%d:sample_fmt=%s:channel_layout=",
+             time_base.num, time_base.den, dec_ctx->sample_rate,
+             av_get_sample_fmt_name(dec_ctx->sample_fmt));
+    av_channel_layout_describe(&dec_ctx->ch_layout, args + ret, sizeof(args) - ret);
+    ret = avfilter_graph_create_filter(&buffersrc_ctx, abuffersrc, "in",
+                                       args, NULL, filter_graph);
+    if (ret < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot create audio buffer source\n");
+        goto end;
+    }
+
+    /* buffer audio sink: to terminate the filter chain. */
+    ret = avfilter_graph_create_filter(&buffersink_ctx, abuffersink, "out",
+                                       NULL, NULL, filter_graph);
+    if (ret < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot create audio buffer sink\n");
+        goto end;
+    }
+
+    ret = av_opt_set_int_list(buffersink_ctx, "sample_fmts", out_sample_fmts, -1,
+                              AV_OPT_SEARCH_CHILDREN);
+    if (ret < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot set output sample format\n");
+        goto end;
+    }
+
+    ret = av_opt_set(buffersink_ctx, "ch_layouts", "mono",
+                              AV_OPT_SEARCH_CHILDREN);
+    if (ret < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot set output channel layout\n");
+        goto end;
+    }
+
+    ret = av_opt_set_int_list(buffersink_ctx, "sample_rates", out_sample_rates, -1,
+                              AV_OPT_SEARCH_CHILDREN);
+    if (ret < 0) {
+        av_log(NULL, AV_LOG_ERROR, "Cannot set output sample rate\n");
+        goto end;
+    }
+
+    /*
+     * Set the endpoints for the filter graph. The filter_graph will
+     * be linked to the graph described by filters_descr.
+     */
+
+    /*
+     * The buffer source output must be connected to the input pad of
+     * the first filter described by filters_descr; since the first
+     * filter input label is not specified, it is set to "in" by
+     * default.
+     */
+    outputs->name       = av_strdup("in");
+    outputs->filter_ctx = buffersrc_ctx;
+    outputs->pad_idx    = 0;
+    outputs->next       = NULL;
+
+    /*
+     * The buffer sink input must be connected to the output pad of
+     * the last filter described by filters_descr; since the last
+     * filter output label is not specified, it is set to "out" by
+     * default.
+     */
+    inputs->name       = av_strdup("out");
+    inputs->filter_ctx = buffersink_ctx;
+    inputs->pad_idx    = 0;
+    inputs->next       = NULL;
+
+    if ((ret = avfilter_graph_parse_ptr(filter_graph, filters_descr,
+                                        &inputs, &outputs, NULL)) < 0)
+        goto end;
+
+    if ((ret = avfilter_graph_config(filter_graph, NULL)) < 0)
+        goto end;
+
+    /* Print summary of the sink buffer
+     * Note: args buffer is reused to store channel layout string */
+    outlink = buffersink_ctx->inputs[0];
+    av_channel_layout_describe(&outlink->ch_layout, args, sizeof(args));
+    av_log(NULL, AV_LOG_INFO, "Output: srate:%dHz fmt:%s chlayout:%s\n",
+           (int)outlink->sample_rate,
+           (char *)av_x_if_null(av_get_sample_fmt_name((AVSampleFormat)outlink->format), "?"),
+           args);
+
+end:
+    avfilter_inout_free(&inputs);
+    avfilter_inout_free(&outputs);
+
+    return ret;
+}
+
+static void sherpa_decode_frame(const AVFrame *frame, SherpaNcnnRecognizer *recognizer)
+{
+#define N 3200  // 100s. Sample rate is fixed to 16 kHz
+    static float samples[N];
+    static int nb_samples = 0;
+    const int16_t *p = (int16_t*)frame->data[0];
+
+    if (frame->nb_samples + nb_samples > N) {
+        AcceptWaveform(recognizer, 16000, samples, nb_samples);
+        Decode(recognizer);
+        if (IsEndpoint(recognizer)) {
+            SherpaNcnnResult *r = GetResult(recognizer);
+            if (strlen(r->text)) {
+                fprintf(stderr, "%s\n", r->text);
+            }
+            DestroyResult(r);
+        }
+        nb_samples = 0;
+    }
+
+    for (int i = 0; i < frame->nb_samples; i++) {
+        samples[nb_samples++] = p[i] / 32768.;
+    }
+}
+
+static inline char *__av_err2str(int errnum)
+{
+    static char str[AV_ERROR_MAX_STRING_SIZE];
+    memset(str, 0, sizeof(str));
+    return av_make_error_string(str, AV_ERROR_MAX_STRING_SIZE, errnum);
+}
+
+int main(int argc, char **argv)
+{
+    int ret;
+    int num_threads = 4;
+    AVPacket *packet = av_packet_alloc();
+    AVFrame *frame = av_frame_alloc();
+    AVFrame *filt_frame = av_frame_alloc();
+    const char *kUsage =
+        "\n"
+        "Usage:\n"
+        "  ./bin/decode-file-c-api \\\n"
+        "    /path/to/tokens.txt \\\n"
+        "    /path/to/encoder.ncnn.param \\\n"
+        "    /path/to/encoder.ncnn.bin \\\n"
+        "    /path/to/decoder.ncnn.param \\\n"
+        "    /path/to/decoder.ncnn.bin \\\n"
+        "    /path/to/joiner.ncnn.param \\\n"
+        "    /path/to/joiner.ncnn.bin \\\n"
+        "    /path/to/foo.wav [<num_threads> [decode_method, can be "
+        "greedy_search/modified_beam_search]]"
+        "\n\n"
+        "Please refer to \n"
+        "https://k2-fsa.github.io/sherpa/ncnn/pretrained_models/index.html\n"
+        "for a list of pre-trained models to download.\n";
+
+
+    if (!packet || !frame || !filt_frame) {
+        fprintf(stderr, "Could not allocate frame or packet\n");
+        exit(1);
+    }
+
+    if (argc < 9 || argc > 11) {
+        fprintf(stderr, "%s\n", kUsage);
+        return -1;
+    }
+
+    SherpaNcnnModelConfig model_config;
+    model_config.tokens = argv[1];
+    model_config.encoder_param = argv[2];
+    model_config.encoder_bin = argv[3];
+    model_config.decoder_param = argv[4];
+    model_config.decoder_bin = argv[5];
+    model_config.joiner_param = argv[6];
+    model_config.joiner_bin = argv[7];
+  
+    if (argc >= 10 && atoi(argv[9]) > 0) {
+        num_threads = atoi(argv[9]);
+    }
+    model_config.num_threads = num_threads;
+    model_config.use_vulkan_compute = 0;
+  
+    SherpaNcnnDecoderConfig decoder_config;
+    decoder_config.decoding_method = "greedy_search";
+  
+    if (argc == 11) {
+        decoder_config.decoding_method = argv[10];
+    }
+    decoder_config.num_active_paths = 4;
+    decoder_config.enable_endpoint = 1;
+    decoder_config.rule1_min_trailing_silence = 2.4;
+    decoder_config.rule2_min_trailing_silence = 1.2;
+    decoder_config.rule3_min_utterance_length = 300;
+  
+    SherpaNcnnRecognizer *recognizer =
+        CreateRecognizer(&model_config, &decoder_config);
+
+
+    if ((ret = open_input_file(argv[8])) < 0)
+        exit(1);
+
+    if ((ret = init_filters(filter_descr)) < 0)
+        exit(1);
+
+    /* read all packets */
+    while (1) {
+        if ((ret = av_read_frame(fmt_ctx, packet)) < 0)
+            break;
+
+        if (packet->stream_index == audio_stream_index) {
+            ret = avcodec_send_packet(dec_ctx, packet);
+            if (ret < 0) {
+                av_log(NULL, AV_LOG_ERROR, "Error while sending a packet to the decoder\n");
+                break;
+            }
+
+            while (ret >= 0) {
+                ret = avcodec_receive_frame(dec_ctx, frame);
+                if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF) {
+                    break;
+                } else if (ret < 0) {
+                    av_log(NULL, AV_LOG_ERROR, "Error while receiving a frame from the decoder\n");
+                    exit(1);
+                }
+
+                if (ret >= 0) {
+                    /* push the audio data from decoded frame into the filtergraph */
+                    if (av_buffersrc_add_frame_flags(buffersrc_ctx, frame, AV_BUFFERSRC_FLAG_KEEP_REF) < 0) {
+                        av_log(NULL, AV_LOG_ERROR, "Error while feeding the audio filtergraph\n");
+                        break;
+                    }
+
+                    /* pull filtered audio from the filtergraph */
+                    while (1) {
+                        ret = av_buffersink_get_frame(buffersink_ctx, filt_frame);
+                        if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF)
+                            break;
+                        if (ret < 0)
+                            exit(1);
+                        sherpa_decode_frame(filt_frame, recognizer);
+                        av_frame_unref(filt_frame);
+                    }
+                    av_frame_unref(frame);
+                }
+            }
+        }
+        av_packet_unref(packet);
+    }
+
+    // add some tail padding
+    float tail_paddings[4800] = {0};  // 0.3 seconds at 16 kHz sample rate
+    AcceptWaveform(recognizer, 16000, tail_paddings, 4800);
+
+    InputFinished(recognizer);
+
+    Decode(recognizer);
+    SherpaNcnnResult *r = GetResult(recognizer);
+    fprintf(stderr, "%s\n", r->text);
+    DestroyResult(r);
+
+    DestroyRecognizer(recognizer);
+    avfilter_graph_free(&filter_graph);
+    avcodec_free_context(&dec_ctx);
+    avformat_close_input(&fmt_ctx);
+    av_packet_free(&packet);
+    av_frame_free(&frame);
+    av_frame_free(&filt_frame);
+
+    if (ret < 0 && ret != AVERROR_EOF) {
+        fprintf(stderr, "Error occurred: %s\n", __av_err2str(ret));
+        exit(1);
+    }
+
+    return 0;
+}