AV_SAMPLE_FMT_FLT是FFmpeg中的一种音频采样格式，表示每个采样点使用32位浮点数表示。

下面是使用FFmpeg库进行音频转换的示例代码，将音频从AV_SAMPLE_FMT_S16（有符号16位整数）转换为AV_SAMPLE_FMT_FLT（32位浮点数）：

#include 
#include 
#include 
#include 
#include 
#include 
#include 
#include 
#include 

int main() {
    // 输入音频参数
    AVSampleFormat inSampleFormat = AV_SAMPLE_FMT_S16;
    int inSampleRate = 44100;
    uint64_t inChannelLayout = AV_CH_LAYOUT_STEREO;

    // 输出音频参数
    AVSampleFormat outSampleFormat = AV_SAMPLE_FMT_FLT;
    int outSampleRate = 44100;
    uint64_t outChannelLayout = AV_CH_LAYOUT_STEREO;

    // 初始化输入和输出音频上下文
    SwrContext *swrContext = swr_alloc_set_opts(NULL,
                                                outChannelLayout, outSampleFormat, outSampleRate,
                                                inChannelLayout, inSampleFormat, inSampleRate,
                                                0, NULL);
    if (!swrContext) {
        fprintf(stderr, "Failed to allocate SwrContext\n");
        return -1;
    }

    // 初始化SwrContext
    if (swr_init(swrContext) < 0) {
        fprintf(stderr, "Failed to initialize SwrContext\n");
        swr_free(&swrContext);
        return -1;
    }

    // 输入音频数据
    int inNbSamples = 1024;
    uint8_t **inData = (uint8_t **) calloc(inChannelLayout, sizeof(uint8_t *));
    inData[0] = (uint8_t *) malloc(av_samples_get_buffer_size(NULL, inChannelLayout, inNbSamples, inSampleFormat, 0));
    memset(inData[0], 0, av_samples_get_buffer_size(NULL, inChannelLayout, inNbSamples, inSampleFormat, 0));

    // 输出音频数据
    int outNbSamples = av_rescale_rnd(inNbSamples, outSampleRate, inSampleRate, AV_ROUND_UP);
    uint8_t **outData = (uint8_t **) calloc(outChannelLayout, sizeof(uint8_t *));
    outData[0] = (uint8_t *) malloc(av_samples_get_buffer_size(NULL, outChannelLayout, outNbSamples, outSampleFormat, 0));
    memset(outData[0], 0, av_samples_get_buffer_size(NULL, outChannelLayout, outNbSamples, outSampleFormat, 0));

    // 进行音频转换
    int convertResult = swr_convert(swrContext, outData, outNbSamples, (const uint8_t **) inData, inNbSamples);
    if (convertResult < 0) {
        fprintf(stderr, "Failed to convert audio\n");
        swr_free(&swrContext);
        free(inData[0]);
        free(inData);
        free(outData[0]);
        free(outData);
        return -1;
    }

    // 释放资源
    swr_free(&swrContext);
    free(inData[0]);
    free(inData);
    free(outData[0]);
    free(outData);

    return 0;
}

上述代码中，我们首先设置输入音频的采样格式、采样率和声道布局，然后设置输出音频的采样格式、采样率和声道布局。接着，我们分配输入和输出音频数据的内存，并将其初始化。

然后，我们使用swr_alloc_set_opts函数初始化SwrContext，并使用swr_init函数初始化SwrContext。接下来，我们进行音频转换，使用swr_convert函数将输入音频数据转换为输出音频数据。

最后，我们释放SwrContext和音频数据的内存。

请注意，以上代码只是演示了如何将音频从AV_SAMPLE_FMT_S16转换为AV_SAMPLE_FMT_FLT，实际使用时可能需要根据实际情况进行适当的修改。