Labels you never have to fix
Each track is transcribed on its own and labeled with the name you gave it, so Host, Guest and Producer are right in every paragraph from the first read.
Reviewed by David Dat Nguyen, CXO Last updated
Multitrack audio
Tracks in.
One transcript.
Upload the isolated tracks from your recorder or your editing session, name each one, and Sonix transcribes them together into a single transcript where the speaker labels are already right.
No speaker detection to correct, because the label comes from the track. Built for podcasts, remote interviews and any shoot with a mic on every person.
30 free minutes on every new account, no credit card.See pricing
Two to five isolated tracks
Track name becomes the speaker
Merged into one transcript
The labels come from the mics
Three things multitrack upload does for a multi-mic recording.
One transcript in time order
The tracks are merged by their timecodes into a single conversation, with each paragraph where it was spoken, ready to edit and export like any Sonix transcript.
Cleaner words from cleaner audio
An isolated track has one voice and no room noise from the other mics, which is the audio a transcription engine hears best.
Anywhere there is a mic on every person
How three kinds of teams use multitrack.
Podcasters
The recorder already keeps the hosts apart; so does the transcript.
- Upload the host and guest tracks from a remote recording session
- Name the tracks once; the show notes and transcript carry the right names
- Export DOCX for the episode page and SRT for the video version
Filmmakers
Lavs and booms, transcribed as the people they were on.
- Upload each lavalier track from the interview with the subject's name
- Bring the merged transcript into Premiere or Final Cut by speaker
- Keep crew talk on its own track and out of the subject's paragraphs
Radio stations
Studio and remote guests on separate channels, one script out.
- Transcribe the presenter and the phone line as separate tracks
- Label the tracks by name for a readable rundown or web article
- Archive one searchable transcript per show with speakers attached
How multitrack works
The settings and limits, as they are in the product today.
- Separate tracks at upload2 to 5 files
- Choose "Separate tracks" on the upload form and add two to five files. The form will not continue with a single track.
- A name per trackOne per track
- Name each track before you transcribe. That name is the speaker label on every paragraph from that track, so no voice detection runs.
- Isolated and equal lengthEqual length
- Keep every track the same length and fully isolated. If one mic bleeds into another, Sonix hears it as an echo and the labels suffer.
- Merged by timeBy timecode
- Each track is transcribed and the results are interleaved by timecode into one transcript, which you edit and export like any other.
- BillingOne recording
- A multitrack upload is billed once, as one recording, not once per track.
- The APIspeaker_labels
- Send the multitrack flag with the media and a speaker_labels list, and the API applies one label per channel.
Multitrack, answered
How would a podcaster use multitrack for an episode?
Export the host and guest tracks separately from your recording tool, upload both with Separate tracks selected, and name them. The transcript comes back as one conversation with Host and Guest on the right paragraphs, ready for the episode page as DOCX or for the video edition as SRT.
How many tracks can I upload?
Two to five; the form will not continue with one. Each track should be a separate, isolated recording of the same session, the same length, so the merge by timecode lines up.
Does it cost more than a single file?
No. A multitrack upload is billed once, as one recording, not once per track.
What if my tracks have bleed from the other mics?
Sonix hears the bleed as an echo of the other speaker and may transcribe it on the wrong track. Use the isolated tracks from the recorder or mute the other mics in your editor before exporting the stems. For a single mixed file, upload it normally and let speaker identification label the voices.
Can a film crew keep the crew mic out of the transcript?
Yes. Only upload the tracks you want transcribed, or upload the crew track under its own name and delete those paragraphs in the editor. Each track's words stay on their own labeled paragraphs.
Pages around this one
The feature this page belongs to, and its neighbors.
Everything else Sonix does
with the same recording
Get started
Upload your tracks
Bring the stems from your last session and read them as one conversation. 30 free minutes, no credit card.
4.98
10xfaster than real time
30free minutes to start
