smart_format affecting year transcription in Speech to Text

[hello-world-bfree]

Similar yet seemingly different from discussion-1293.

When a date is spoken, i.e., "722 BC" (read as "seven twenty-two BC) will be transcribed as "July")

[deepgram-community[bot]]

Thanks for asking your question. Please be sure to reply with as much detail as possible so the community can assist you efficiently.
_{Consider joining our Discord community for more opportunity to engage with your fellow Deepgram users. You can earn points which can be redeemed for cool stuff by being active in our communities!}

[deepgram-community[bot]]

It looks like we're missing some important information to help debug your issue. Would you mind providing us with the following details in a reply?

• The programming language you are working in (e.g. JavaScript, Python).
• The deepgram product you are using (e.g Speech to Text, Agent API)
• A request ID that triggered your error or issue.

[hello-world-bfree]

• Python
• Speech to Text
• fd35ddf1-1e65-4be8-9af4-a67d7edc48fe & 133dd1ac-9306-4bff-a05f-d728fae486d4

[deepgram-community[bot]]

That BC date might be a bit tricky for our models to get right. I see you did try smart format as well on one of the requests, which was something I was going to suggest.

[00:05:47.77 - 00:05:53.85] [Speaker 0] [Channel 0]: Capernaum is also where the exile began in the Old Testament in July. <--- I assume this should be 722 BC
Keyterms might be helpful in this case: https://developers.deepgram.com/docs/keyterm

something like:

  --request POST \
  --header 'Authorization: Token YOUR_DEEPGRAM_API_KEY' \
  --header 'Content-Type: audio/wav' \
  --data-binary @youraudio.wav \
  --url "<https://api.deepgram.com/v1/listen?model=nova-3&keyterm=BC>"```

> ![avatar](https://avatars.slack-edge.com/2023-02-20/4855338233600_fad9174023ee3f640ce1_32.png) This message was sent by John Vajda from Deepgram, via our community automation.

[deepgram-community[bot]]

We've verified this behavior specific to certain subsets of BC/BCE/AD years when using Smart Formatting specifically. We are tracking this work and should have this resolved over the next few weeks.

This message was sent by nick kaimakis from Deepgram, via our community automation.

[hello-world-bfree]

Hello!👋 Just checking in on this issue. I know you'd mentioned it'd be a little tricky, but wanted to see if there's been any updates or movement on this issue?

Thanks so much!

[deepgram-community[bot]]

we haven't been able to prioritize this yet, but we did just recently launch a new feature that might help you address this tangentially -- are you using streaming or batch STT? /what model?

[hello-world-bfree]

Using batch currently and nova-3

[hello-world-bfree]

Has there been any traction on this? It's still an issue unfortunately

[deepgram-community[bot]]

this is still not yet natively supported via formatting, but I recommend using our entity detection feature in combination with smart formatting to do your own custom downstream formatting postprocessing for edge cases such as these

This message was sent by nick kaimakis from Deepgram, via our community automation.