IBM Data and AI

Welcome to the IBM Data and AI Ideas Portal for Clients!

We welcome and appreciate your feedback on IBM Data and AI Products to help make them even better than they are today!
Before you submit an idea, please perform a search first as a similar idea may have already been reported in the portal. If a related idea is not yet listed, please create a new idea and include with it a description which includes expected behavior as well as why having this feature would improve the service and how it would address your use case.
IBM Employees:
Clients:
  • Our team welcomes any feedback and suggestions you have for improving our offerings / products! This forum allows us to connect your offering / product improvement ideas with IBM product and engineering teams.

  • If you have not registered on this portal please click on the following link and register. To complete registration you will need to open the email you will receive from Aha to confirm your identity. http://ibm.biz/IBM-Data-and-AI-Portal-Register

Additional Information:
  • The shorter URL for this site is: https://ibm.biz/IBM-Data-and-AI-Ideas

  • To view our roadmaps: http://ibm.biz/Data-and-AI-Roadmaps

  • Reminder: This is not the place to submit defects or support needs, please use normal support channel for these cases

  • Please do not use the Ideas Portal for reporting bugs - we ask that you report bugs or issues with the product by contacting IBM support.

Utterance segmentation made sensitive to speaker identity (features)

Utterance segmentation appears to be entirely independent of (the features used for) speaker labeling.  Specifically, it was noticed that even though the speaker labeling correctly identifies that a new speaker (very clear because it goes from male voice to female voice) after (what is transcribed as) `seven's` and before (what is transcribed as) `to` in `three sixty eight reduce speed to two one zero then descend and maintain three thousand advise seven's to one zero thousand three thousand US air six eighty six ` - that entire utterance is chunked together - this leads to errors because the 'to' should be transcribed as 'two' (and would have been had it been an utterance initial word.

  • Avatar32.5fb70cce7410889e661286fd7f1897de Guest
  • Jan 9 2019
  • Already exists
Who would benefit from this IDEA? As a customer transcribing mulit-speaker speech, I want the transcritiption to be as accurate and correctly segmented as possible
  • Attach files
  • Admin
    Marco Noel commented
    14 Nov, 2019 08:45pm

    This is the "speaker_labels" feature of STT.

NOTICE TO EU RESIDENTS: per EU Data Protection Policy, if you wish to remove your personal information from the IBM ideas portal, please login to the ideas portal using your previously registered information then change your email to "anonymous@euprivacy.out" and first name to "anonymous" and last name to "anonymous". This will ensure that IBM will not send any emails to you about all idea submissions