- From: Deborah Dahl <Dahl@conversational-Technologies.com>
- Date: Wed, 6 Oct 2021 12:55:31 -0400
- To: <public-voiceinteraction@w3.org>
This question is mostly for Dirk, but if anyone else has thoughts or suggestions they would be welcome. This question came up in the call today about the "remote path" from the IPA service to the Provider Selection Service https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paArchitecture-1-2.htm#walkthrough Step 5 states "The IA Service forwards the received data simultaneously to the ASR in the local path and to the Provider Selection Service in the remote path." For the remote path, no ASR or NLU processing has been performed; the data is just audio and metadata. The first time that ASR or NLU is performed is at the "IPA Provider" step. However, at step 10, it states that "At this level, only the pure text is known and the used language." So the question is, where does the text come from, because there hasn't been any ASR yet?
Received on Wednesday, 6 October 2021 16:55:46 UTC