1. When requesting the API for Alibaba Cloud Intelligent Speech Interaction – Text-to-Speech, the supported audio encoding formats are PCM/WAV/MP3. The default value is pcm. To keep the response body as small as possible, we need to verify the response body sizes of these three formats.
2. When requesting the API in Postman, if the response format is mp3, its size is 187.62 KB. See Figure 1.
3. When the response format is pcm, its size is 1.48 MB. See Figure 2.
4. When the response format is wav, its size is 1.47 MB. See Figure 3.
5. I found that the size of the mp3 format is only about 12% of the other two formats. Because this API is an interface on my own server (which then passes parameters through to the Alibaba Cloud API), I suspected that the mp3 transmission might have been compressed. However, upon closer inspection of the response headers, the headers are completely identical, especially the Transfer-Encoding value, which is chunked for all of them. This indicates that no compression was applied. The actual size of the mp3 format is indeed only about 12% of the other two formats.