使用 libsox 对 wav 文件进行下采样 [英] Downsampling a wav file using libsox

问题描述

我正在尝试使用 Sox 库 C 程序(http://sox.sourceforge.net/).从 PCM 到 mu 的转换工作正常.但是当我应用下采样效果时，输出文件的持续时间只是 i/p 文件的一半(见下文).我使用了帖子中提到的技术 如何在通过 sox C 库进行格式转换时更改采样率? 但这对我没有帮助.

I'm trying to convert a 16KHZ 16-bit Signed PCM encoded wave file to 8KHz 8-bit mu-encoded wav file using the Sox library C program (http://sox.sourceforge.net/). The conversion from PCM to mu is working fine. But when I apply the down-sampling effect the output file duration is just half of the i/p file(see below). I used the technique mentioned in the post How to change the samples rates when do the format conversion by sox C libraries? but it didn't help me.

当我执行下面的代码时，我看到一个警告

When I execute the below code I see a warning

wav: Premature EOF on .wav input file

输出:

Input File     : 'text2speech_0.wav'
Channels       : 1
Sample Rate    : 16000
Precision      : 16-bit
**Duration       : 00:00:06.24 = 99777 samples ~ 467.705 CDDA sectors**
File Size      : 200k
Bit Rate       : 256k
Sample Encoding: 16-bit Signed Integer PCM


Input File     : 'out_8k.wav'
Channels       : 1
Sample Rate    : 8000
Precision      : 14-bit
**Duration       : 00:00:03.12 = 24945 samples ~ 233.859 CDDA sectors**
File Size      : 49.9k
Bit Rate       : 128k
Sample Encoding: 8bit u-law

代码:

int main(int argc, char * argv[])
{
    static sox_format_t * in, * out; /* input and output files */
    sox_effects_chain_t * chain;
    sox_effect_t * e;
    char * args[10];
    assert(argc == 3);
    assert(sox_init() == SOX_SUCCESS);
    assert(in = sox_open_read(argv[1], NULL, NULL, NULL));

    assert(out = sox_open_write(argv[2], &in->signal, NULL, NULL, NULL, NULL));

    chain = sox_create_effects_chain(&in->encoding, &out->encoding);

    e = sox_create_effect(sox_find_effect("input"));
    args[0] = (char *)in, assert(sox_effect_options(e, 1, args) == SOX_SUCCESS);
    assert(sox_add_effect(chain, e, &in->signal, &in->signal) == SOX_SUCCESS);
    free(e);

    out->signal.rate = 8000;
    in->signal.rate = 16000;

   if (in->signal.rate != out->signal.rate) {
        e = sox_create_effect(sox_find_effect("rate"));
        args[0] = "16000", assert(sox_effect_options(e, 1, args) == SOX_SUCCESS);
        assert(sox_add_effect(chain, e, &in->signal, &out->signal) == SOX_SUCCESS);
        free(e);
    }

    if (in->signal.channels != out->signal.channels) {
        e = sox_create_effect(sox_find_effect("channels"));
        assert(sox_effect_options(e, 0, NULL) == SOX_SUCCESS);
        assert(sox_add_effect(chain, e, &in->signal, &out->signal) == SOX_SUCCESS);
        free(e);
    }

    e = sox_create_effect(sox_find_effect("output"));
    args[0] = (char *)out, assert(sox_effect_options(e, 1, args) == SOX_SUCCESS);
    assert(sox_add_effect(chain, e, &in->signal, &out->signal) == SOX_SUCCESS);
    free(e);

    sox_flow_effects(chain, NULL, NULL);

    sox_delete_effects_chain(chain);
    sox_close(out);
    sox_close(in);
    sox_quit();

    return 0;
}

编译&执行:

gcc -g -o example3 example3.c `pkg-config --cflags --libs sox`
./example3 text2speech_0.wav out_8k.wav

推荐答案

Bellow 是工作代码的差异.我认为犯的主要错误是在打开写入之前必须设置out"结构.

Bellow is a diff for a working code. I believe the main mistake made is that the "out" structure must set before opening it for writing.

+#include <sox.h>
+#include <assert.h>
+#include <stdlib.h>
+#include <string.h>
+
+ int main(int argc, char * argv[])
+ {
+     static sox_format_t * in, * out; /* input and output files */
+@@ -7,9 +13,15 @@
 assert(argc == 3);
 assert(sox_init() == SOX_SUCCESS);
 assert(in = sox_open_read(argv[1], NULL, NULL, NULL));
 +    out= (sox_format_t *) malloc(sizeof (sox_format_t));
 +    memcpy(out, in, sizeof (sox_format_t));
 +    out->encoding.encoding = SOX_ENCODING_ULAW;
 +    out->encoding.bits_per_sample=8;
 +    out->signal.rate = 8000;
 +    out->signal.precision = 8;
 +    out->signal.length = SOX_UNSPEC;
 +    assert(out = sox_open_write(argv[2], &out->signal, &out->encoding, NULL, NULL, NULL));

 -    assert(out = sox_open_write(argv[2-    assert(out = sox_open_write(argv[2],  &in->signal, NULL, NULL, NULL, NULL));
 -
 chain = sox_create_effects_chain(&in->encoding, &out->encoding);

 e = sox_create_effect(sox_find_effect("input"));
 @@ -17,16 +29,14 @@
 assert(sox_add_effect(chain, e, &in->signal, &in->signal) == SOX_SUCCESS);
 free(e);

 -    out->signal.rate = 8000;
 -    in->signal.rate = 16000;
 if (in->signal.rate != out->signal.rate) {
     e = sox_create_effect(sox_find_effect("rate"));
 -        args[0] = "16000", assert(sox_effect_options(e, 1, args) == SOX_SUCCESS);
 +        assert(sox_effect_options(e, 0, NULL) == SOX_SUCCESS);
 +         e->handler.flags |= SOX_EFF_LENGTH;
 +  
     assert(sox_add_effect(chain, e, &in->signal, &out->signal) == SOX_SUCCESS);
     free(e);

这篇关于使用 libsox 对 wav 文件进行下采样的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！