I am having it stream. though it seems the voice model can only produce 24Khz voice, I need to down-sample the voice to 8Khz. The 24Khz is fine, but when sampled down to 8Khz the voice is noisy.
I also have tried non stream and still get the same result when down sampling with ffmpg.
Sample is attached.
emma-sse-8k.wav
I am having it stream. though it seems the voice model can only produce 24Khz voice, I need to down-sample the voice to 8Khz. The 24Khz is fine, but when sampled down to 8Khz the voice is noisy.
I also have tried non stream and still get the same result when down sampling with ffmpg.
Sample is attached.
emma-sse-8k.wav