No ACELP.KELVIN encoder!
2024-01-29 14:11:20
Find a FFmpeg version that supports ACELP.KELVIN encoding, if you can't find one, create a ACELP.KELVIN encoder. Input should be 8 kHz, mono/stereo, 16-bit PCM WAV and output should be in ACELP.KELVIN WAV. To make a ACELP.KELVIN encoder: take the decoder sources and modifiy for encoding. Post as a EXE that works with Windows 7 (compressed as a ZIP file) and as source code (also compressed as a ZIP file) One quote that drives me crazy, because apparently human speech only requires mono, for some reason:Take a sample input audio source is encoded in “acelp.kelvin” which is not supported by STT API To determine how the audio source is encoded, run ffmpeg -i input.wav output will shown the encodingStream #0:0: Audio: acelp.kelvin (5[1][0][0] / 0x0135), 8000 Hz, 2 channels, s16p, 17 kb/s In the bolded paragraph line of text, he's literally talking about a STEREO ACELP.KELVIN file, which you guess, human speech only requires mono, for some reason. What reason? Because mono sounds very good for human speech! Stereo also sounds very good for human speech, but check this out, running MediaInfo on this sample file: "format-0x135-Roligt_samtal.wav" reveals the following:General Complete name : E:\Sample files\format-0x135-Roligt_samtal.wav Format : Wave Format settings : WaveFormatEx File size : 175 KiB Duration : 2 min 44 s Overall bit rate : 8 723 b/s Audio Format : KELVIN Codec ID : 135 Codec ID/Hint : Sipro Duration : 2 min 44 s Bit rate : 8 720 b/s Channel(s) : 1 channel Sampling rate : 8 000 Hz Stream size : 175 KiB (100%) - In this file, is MONO. but in the file at the quote is STEREO. Many wishes and wises for coding a ACELP.KELVIN encoder.MOD edited: paragraph > line of text