Loving the transcript feature that you all are working on. I would like to request the following.the ability to have the text synced to the clip and not just the sequence.The ability to select a string of text and have it convert to in and out pointsWhen exporting the transcript to a text file I would love to have the speaker as well as the timecode references to export with it. Or at least the option to turn them on or off when exportingThanks guys for your great work keep it up!

Godsanimator

Inspiring

Question

Transcript feature request

Forum|Forum|4 years ago
March 9, 2021
6 replies
612 views

Loving the transcript feature that you all are working on. I would like to request the following.

the ability to have the text synced to the clip and not just the sequence.
The ability to select a string of text and have it convert to in and out points
When exporting the transcript to a text file I would love to have the speaker as well as the timecode references to export with it. Or at least the option to turn them on or off when exporting

Thanks guys for your great work keep it up!

This topic has been closed for replies.

G

Gurgles

Participating Frequently

I agree with all of these suggestions. I myself use the transcript feature more than captioning/subtitles purely to assemble edits that are interview driven.

Here is what I'd like to see in addition.

Speech to text more refined. Right now everything that is transcribed is broken into only a single sentence or two. Works great for subtitles/captions. Not very good for assisting in piecing together edits. Look at what services such as Temi are doing and how it's speech to text is being used in a more robust way.
Please please please add timecode stamps when exporting a .txt file. As of right now the text file export feature isn't helpful when I have no idea where the text is in coorliation with the sequence's timecode. Just simply mirroring the layout in the text panel when exporting would be nice.
General improved optimization for faster editing within the text panel and better speech recognition in the AI. Especially for longer interviews (1+ hours) I'd say it's about 85% accurate in speech recognition.
Finally, for the love all that is beautiful in the world. Please add a highlighter. Not being able to mark or make note of things said makes it harder to go back and find the important content.

I love the feature so far and can't wait to see how Adobe continues to improve this. Already saving me a lot of money and time.

Stan Jones

Community Expert

> Please please please add timecode stamps when exporting a .txt file. As of right now the text file export feature isn't helpful when I have no idea where the text is in coorliation with the sequence's timecode. Just simply mirroring the layout in the text panel when exporting would be nice.

This was added in the Release version 22.0.0. The .txt export includes speaker and timecode. Formatting can really mess up this being useful, but there is also now a .csv export. You can open in a spreadsheet and pick the information you want.

Stan

G

Gurgles

Participating Frequently

Hi Stan,

I can confirm that is indeed the case. I had stopped using the feature for a while prior to v22's release. Happy to see it there now.

I'd still love to see some better optimization and flexibility for the text panel. The panel seems to really slow down when transcribed sequences are longer than 40 minutes or so. They way text is broken out now is by every 1 - 3 sentences and often by just a couple words. That's hard to easily read and scrub through when you have an hour or more of an interview. And it eats up a fair amount of time trying to merge all of the text into a more legible format. It would be nice to have larger paragraphs of text when premiere initially transcribes the sequence for those of us who aren't using it for subtitles or captioning purposes.

Still really like the feature. It's helping a lot.

Exile Edit

Participant

Another really helpful feature for the transcript would be the ability to search for a phrase rather than just a sequence.

For instance instead of just being able to search the word "I've" you could seach for "I've been". This would dramatically cut down on the results and make finding what you need much simpler.

Would also be great to be able to filter results by speaker.

Thanks for considering!!!

Exile Edit

Participant

Typo above, should read: Another really helpful feature for the transcript would be the ability to search for a phrase rather than just a single word.

Exile Edit

Participant

I have been using the feature as well and these are the EXACT OPTIONS I would love to see implemented.

For my workflow I am using the transcript more than the captions -- although the captions can be useful.

But I am using the transcript, to find sections, then jump to the portion of the interview/clip

The above suggestions would be very helpful for the following:

• Using the transcript to select sections of dialogue for "line breakdowns" or stringouts - this could help automate the assistant's workflow

• If the transcript followed along with the source clip, then as you cut into a new seuqence, the transcript could auto-populate along with the clips, so you wouldnt need to generate new transcripts for each sequence.

• If the names of the speaker were tagged along with the timecode of the transcript then it would make it really easy to share transcripts with clients etc.

A

Amy5CC4

Participant

I also would love to be able to export the speakers in the transcript. Why have the ability to edit and tag the speakers, if it can't be exported? For long conversations, my team might have to go back to using a paid service, just so that a transcription makes sense.

Otherwise, I'm finding that this is a great feature, really timely.

Thanks

D

defaultfuz35ghb3aul

Participant

this is what i was searching it can be useful to have an export text with all the timecode and speakers not only the text. And the ability to change the timecode in the transcript

Ken Ochalek

Participant

Agreed on all counts, especially the third bullet.

I work with a market research organization, and the ability to have speaker-labelled text transcripts of over one hundred ~2 hour focus groups per year would be earth-shattering.

It's outside the scope of this thread, but the ability to run Sensei on a local system would also be amazing.

R Neil Haugen

Legend

All great and needed suggestions!

Neil

Everyone's mileage always varies ...

Sign up

To post, reply, or follow discussions, please sign in with your Adobe ID.

Sign in to Adobe Community

To post, reply, or follow discussions, please sign in with your Adobe ID.

Scanning file for viruses.

This file cannot be downloaded