Hey all, I’ve reached out to support but not getting any help.
I’m a new user. When I sync the memory everything is Speaker 0 and Speaker 1. It doesn’t recognize my voice though I went through that in the set-up process. When I change Speaker 1 to a name (like “jim”) it doesn’t update all the Speaker 1 mentions. I don’t have any speaker tagging capability. Can someone help me out here?
hey @Dave1 We do our best in terms of matching speakers with previous meetings but it is still an ongoing process to make it accurate. We recommend having a bunch of recordings with names tagged, so we are able to create a speaker profile which makes the process a lot more accurate.
We are conducting a lot more internal research on improving matching lemme know if I can help out in any other way!
If also seems like you are on the free unlimited plan which does not include speaker splitting please subscribe to pro for the same!
I came out here for some answers about speaker tagging myself. How can I play back the audio recording so I can hear the voice and then tag the speaker appropriately?
For context I’m coming from Otter for recording, auto speaker tagging, summaries. I’m very supportive of Pocket, but before making a bold claim like “industry best” go take a look at somebody like Otter who’s been doing this for years. They do an exceptional job of auto tagging speakers from Voice profiles they’ve built. They also offer a feature to re-tag and rematch speakers.
So please don’t claim to be industry best quite yet, when you have some of those features and they’re consistent I’m open to that discussion.
you can currently download the audio from the summary page to playback
Thanks for the feedback, Please do try out our desktop app where speaker splitting and diarization tends to have better accuracy due to us having separate audio streams
And as always Speaker identification is a iterative process. Our ML Models learn as more speakers are tagged. You can currently retag/rename/edit users for each individual segment as well on our app
Understood. Thank you I’ll check out the desktop app.
I think what I’m asking for is the ability to playback the audio in the transcript so that I can easily update the speaker tagging. That’s become a pretty standard design pattern for software products like this. During playback, you can click on other parts of the transcript to jump around.
Please upvote here: Listen to audio on device
Thanks for the reply. Just so I understand, unless I’m on the subscription plan, it won’t recognize my voice or others?
hey @Dave1 yes! we wont be able to start creating voice profiles for users unless you are on a subscription + as mentioned above it takes tagging a few recordings before we are able to accurately recognize users
Can you elaborate on this? I am curious how this works…
Is it that the computer microphone is better/providing two sources?
Or is it recording on both the computer and the device at the same time, and then able to compare the two sources for tagging?
If answer #2, are there ways to optimize this (aka, putting the pocket in a different location than the computer?)
Hey @Gus so microphones on laptops provide two sources which allows for better transcription + the quality of mics is a lot higher which allows for a better speaker profile to be generated (that why transcription providers on laptop sometimes have higher speaker identification)
@mhirrrr Is v0.1.4 the latest version of the desktop app? Can you resend the link to download the latest version?
Hey! Could you help us with the windows version you are using so we could debug this better thanks!
@mhirrrr Windows 11 Home. Currently have Pocket_0.1.3_x64-setup installed.
Where can i get the desktop app?
hey! you can find the links here Pocket Desktop App! - #72 by satyamx64
@mhirrrr Can you please explain how to retag/rename/edit users for each individual segment as well on our app? I have attempted to do this both in the Android and Desktop app, and see no evidence that the rename took. Every speaker is still labeled with the default “Speaker0” etc label.
Not only has my speaker list been “removed" (no names under Speaker Management), but I’m also unable to add a speaker or change the speaker tags in a transcript.
Hey @p_odoi which version is this we haven’t pushed out any changes this weekend do let us know