Transcription needs complete function please

  1. You should be able to manually add a transcription block

  2. if you do a transcription on a layer, you should be able to change the colour of the text after the fact to be compatible with other layers.

  3. moving the text should be an option

  4. breaking up the bubble should be an option if 2 people are speaking in that timeline

  5. my biggest issue: voices separate when it’s a man and a woman. It needs to be able to separate with 2 men and 2 men in the same vocal range.


1 Like

Hi @Marisa_Dery,

  1. Adding this to the wishlist.

  2. If you change the color of a layer, the color of its associated transcription will change to (if it doesn’t refresh, just move or zoom the spectrogram a bit)

  3. You can adjust sentence timings and individual word timings (zoom to reveal individual word timings) by grabbing the borders with the mouse:

  4. and 5. you should do Unmix > Multiple Voices first to separate the 2 speakers into 2 different layers, then apply Unmix > Transcription to each layer:

1 Like

The issue is that I have at least 2 people talking ( sometimes over each other). When I use the transcription feature, it only sees one voice. While I can add the extra sentence to the box, I’d like to break the box in two and make it voice 1 purple and voice 2 green. Maybe have a right click function when you highlight to say “ apply transcription segment to green”. This way you can apply it to another layer.

@Marisa_Dery Here I have recorded a demo with 2 people talking over each over, then I unmix the 2 voices into 2 layers, then I transcribe each layer (so each individual) separately, each with its own color. And show the resulting transcription in the transcript dialog.
Hope that answers your question !

1 Like

Video is great, thanks. My issue is that the 2 voices are just too close in pitch / fundamental frequency so it just won’t unmix them. I have to carve them out. I’ve been at it for a while :(. I will create various layers and do what you suggest. Thanks

@Marisa_Dery Ah, I see. Expect improvements in that area with the upcoming SpectraLayers 11 :slight_smile:
Also, I just sent you a PM about your use case.


@Marisa_Dery can you send me the 2 voices you’re trying to unmix. Message me and I’ll see what I can do.

Unfortunately I can’t because it is for a current case so it’s confidential. Thank you for offering though…