After analysis, Selects works out who is speaking in your footage and matches each speaker to a face on camera and an audio track. It gets most of this right on its own, but it is worth a look before you cut. Speaker setup feeds camera switching, speaker labels, and audio routing across every draft in the project, so a wrong match here shows up everywhere.

Speaker settings live in the [Speaker profile] panel. Click [Speakers] at the bottom of the floating toolbar, on the right of the timeline, to open it.

The panel groups speakers by source file, so a project with three synced cameras and a separate audio recorder lists each file with its own speakers underneath. A file that has not finished analyzing is not listed yet, and its speakers appear once analysis finishes.

Each speaker gets one row with three columns: [Name], [Profile], and [Audio]. Click the number and colored badge on the left of a row to jump the playhead to that speaker's next line, which is the fastest way to confirm you have the right person before you change anything.

Selects picks up speaker names from the conversation during analysis, and it does not always get them right. Click a name to open [Rename speaker], then choose how far the change should reach.

Click [Save] to apply it. The name flows through to the transcript, the timeline, and your handoff.

The [Profile] column shows the face Selects matched to that speaker, pulled from your camera footage. Camera switching follows whichever face is assigned here, so this column is the one that decides who the cut goes to.

A row reading [Off camera] means Selects could not match that speaker to any camera. That is normal for an interviewer, a producer, or anyone working behind the lens. Click the column to open the picker, choose the face that belongs to this speaker, and click [Save].

The [Audio] column shows the track assigned to each speaker, with a waveform and the track name underneath. Selects picks the best available track during analysis.

Click the play button on the row to preview what is on that track. To change it, click the column to open [Main audio track for (speaker name)], preview each option, choose the right one, and click [Done]. A row reading [No audio] means nothing was assigned, and the same picker is where you fix it.

If analysis missed somebody, click [Add speaker] at the bottom of that file's list. Give them a name, assign a profile and an audio track, then click [Save].

Adding a speaker creates the profile only. You still need to find their lines in the transcript and reassign those segments to the new speaker, because Selects will not go back and re-diarize the file for you.

To remove one, right-click their row and click [Delete]. If the file has other speakers, Selects asks which one to merge their lines into, so no dialogue ends up without a speaker.

[Speaker audio settings] decides whether each speaker's assigned tracks play all the way through or only while they are talking. Open the [Apps] tab at the top of the project panel, beside [Project], and click [Speaker audio settings], then pick one of two modes.

Click [Apply] to set it on the draft you have open. [Review speakers] at the bottom of the same panel opens the [Speaker profile] panel if you spot something that needs fixing first.

Color-coding your audio tracks makes a busy multi-cam timeline much easier to read. There are two ways to set a color, and both do the same thing.

Eleven colors are available: Pink, Red, Orange, Yellow, Lime, Green, Cyan, Blue, Violet, Fuchsia, and Stone. Cyan is the default.

The color shows in the audio picker so you can tell tracks apart at a glance, and it carries through to Premiere on handoff as a label color. It is optional, but it earns its keep on projects with several speakers or a lot of audio tracks.

Camera switching happens in chat. Ask for it in the [Assistant] panel and chat asks a few things before it starts: how you want the edit to feel, how often to go wide, how often to cut to a reaction, and, when someone appears on more than one camera, which of their angles is the main one. Answer in your own words, for example "stay wide during the intro" or "use camera 2 more", and chat works to that for the whole edit.

To change the angle on part of the edit yourself, right-click a clip on the timeline and pick a different [Video track] or [Audio track]. To move a line to a different speaker, select it in the transcript and press Alt with a number from 1 to 9, where the number is that speaker's row in the [Speaker profile] panel. The line also switches to that speaker's audio track. On the Final Cut and DaVinci keymaps this is Ctrl with a number instead.

Learn more in Edit with chat.

Review speaker settings as soon as your project opens, before you generate a draft. Speaker, camera, and audio assignments feed every draft you make from that footage, so a wrong match at the start turns into rework on everything downstream. It takes a minute to check and saves considerably more than that.

The camera switching settings panel was removed, and camera switching now runs entirely through chat. Ask for it in the [Assistant] panel and describe how you want the cameras to cut. The [Speaker profile] panel is still where you assign each speaker to a face and an audio track, which is what switching follows.

A speaker shows as [Off camera] when Selects could not match their voice to a face in any of your camera tracks. This is expected for interviewers, producers, and anyone who never appears on screen. If they are on camera and Selects missed it, click the Profile column and assign the right face by hand.

Renaming a speaker with [Rename globally] updates that person everywhere in the project, including drafts you have already made. [Rename just in this source file] keeps the change on that one file. Pick the scope in the rename dialog, so you are never guessing which one you got.

If analysis missed a speaker, add them yourself with [Add speaker] at the bottom of that file's list. Bear in mind this creates the speaker profile only, so their lines stay attributed to whoever the transcript assigned them to until you reassign those segments in the transcript.

You do not need camera switching on single-camera footage, but the Speaker profile panel is still worth a look. Speaker names appear throughout your transcript and travel with your handoff, and the audio assignment decides which track each speaker plays from, which matters whenever you recorded audio separately.