Evaluating Rehearsal Recording Quality and Synchronization

Hello! Your eight-minute rehearsal has given you something more useful than a simple microphone test: introductions from all three in-room speakers, ordinary conversation, overlap, a quiet interval, and a visible, audible clap. You also kept a take identifier and noted any interruptions.

Now the question is whether that take is good raw material for an editor. In about 40 minutes, you will review its sound and picture, check synchronization, and decide which issues require another test before recording day. You are evaluating the capture, not trying to produce a finished podcast.

Start with the files you would actually deliver

Confirm that the rehearsal uploads finished before you review it. Open the recorded material rather than relying on how the session looked or sounded live. If Riverside provides separate original tracks for the room and a remote guest, review those originals as well as any combined playback available. A live call can sound interrupted even when a locally recorded track is intact; conversely, a pleasant-sounding preview does not prove that every original track is usable.

Listen on headphones at a comfortable, steady playback volume. Keep the volume unchanged while comparing speakers: repeatedly turning it up for a quiet person can hide the very imbalance you are trying to detect. Review the room microphone as one track containing three voices. A remote guest, if present, has a separate track; the editor may adjust that guest independently but cannot independently raise just one of the three room voices captured by the shared microphone.

Transom’s Podcasting Basics, Part 3: Audio Levels and Processing explains why testing and correcting the recording setup is preferable to assuming that problems can be repaired later.

First pass: can you understand every voice?

Play the introductions at 0:45–2:00 without looking at the picture. Each person should be understandable at their normal speaking level. Compare the quietest and strongest voices, then check the ordinary conversation at 2:00–4:30: people may turn toward one another and change volume once the introductions are over. If you have a remote guest, compare that voice with the room track too.

Voice balance is about intelligibility, not identical-looking meter movements. A speaker who is modestly quieter may still be usable. A speaker who repeatedly disappears under the other two, or becomes hard to follow whenever they turn their head, needs a setup change. Because the room voices share a track, start with seating, microphone position, and how consistently people speak toward the mic. Raising the shared input gain would raise all three voices and the room noise, not selectively rescue the quiet speaker.

At 4:30–5:15, listen to the deliberate overlap. It is normal for two simultaneous voices to be harder to understand. Ask whether you can still follow enough of the exchange to edit it sensibly, and whether ordinary turn-taking immediately afterward becomes clear again. If overlap regularly makes a voice vanish, note that as both a capture and a conversation-management issue.

Next, listen for clipping during the loudest spoken moments, emphatic replies, or laughter. It may sound crunchy or harsh. If your playback tool displays a waveform, repeated flat-topped speech peaks can support what you hear; if you only have Riverside playback, use your ears and any available recording-level warnings. Do not judge the whole take by the synchronization clap: a single sharp clap can be much louder than normal speech. Repeated distortion on voices is the concern. Turning down playback after recording does not restore clipped speech.

Joe Gilder’s demonstration in How to Set Levels shows both what clipping sounds like and why leaving room for louder speech matters. His equipment differs from yours; concentrate on the recorded result, not his particular mixer controls.

Finally, replay the announced 15-second quiet interval at 6:45–7:15, then a nearby stretch of speech. Can you hear ventilation, traffic, computer fans, handling sounds, or a steady hum? A noise that seems minor during silence may mask quieter words. Also listen to the character of the voices. A distant, roomy sound that follows each person’s own words suggests sound reflecting around the room. Try moving the shared mic or speakers appropriately and reducing hard, reflective surfaces before relying on processing.

A different fault is delayed echo: for example, you hear the remote guest’s words return faintly a moment later, or a room speaker’s words are repeated. That points toward playback entering a live microphone or another active audio source. It is not the same as the natural reverberation of an echoey room.

Transom’s Remote Recording: Riverside and Talk Sync provides useful context for headphones, device selection, and the difference between a clean remote mic and a mic that captures more room sound.

If you hear a delayed repeat with a remote guest, identify whose words repeat and on which recording. Check headphones and playback volume at the relevant end, and check that the phone used as the group camera is not also supplying a second live room microphone. Make a short retest after changing one likely cause; do not simply assume Riverside’s echo controls will make the original raw track clean.

Second pass: does the picture remain usable?

Watch the recorded phone-camera video, not only its live preview. Start with the introductions, sample the middle of the conversation, and check the ending. This catches changes that a single opening frame cannot: someone leans out of shot, the phone shifts, or the camera changes exposure.

Evaluate three things separately:

• Framing: Are all three in-room speakers visible throughout? Are faces large enough to read, with reasonable space above heads? Check the overlap section, when people may turn or lean toward one another.
• Stability and focus: Does the group shot remain steady and reasonably sharp? Note any bump or focus change that lasts into the discussion.
• Lighting: Can you see facial expressions on every speaker? Look for a bright window turning people into silhouettes, deep shadows on one face, or brightness visibly fluctuating during the take.

A higher video resolution does not repair a poorly lit or cut-off subject. For an editor, a stable shot that consistently includes the speakers is more useful than an impressive resolution setting paired with unusable framing.

If a remote guest participated, inspect their recorded image as well. Note a problem on that guest’s track separately from a problem on the group camera; they may require different corrections.

Check that sound and picture stay together

Find the clap near 0:00–0:45. In recorded playback, the sharp sound should coincide with the moment the hands meet on screen. Do not mistake an echo or a second noise for the original clap. If you can view the phone video and room audio together in a simple timeline, compare their timing there; otherwise, check the combined recorded playback and tell the editor which original files need confirmation. When the phone video and microphone audio were recorded as separate assets, a good-looking live call alone is not a synchronization check.

Then inspect clear speech near the end of the take. Mouth movement should still agree with the words you hear. A clap that lines up at the start but speech that is visibly early or late near the end may indicate drift. Record that observation rather than moving the audio to fit only the first clap: shifting one track once may not solve timing that changes over the recording.

Do not demand frame-perfect certainty from a small browser preview. Your practical question is whether the timing looks natural at the beginning and near the end, or whether the editor needs to investigate the original tracks.

Turn observations into a recording-day decision

Keep a short review log under the take identifier. For each finding, record when it occurs, what you heard or saw, its likely cause, and what you will retest. For example: “R01, 2:35 and 3:10: third room voice hard to understand when turned away; try a different seat and repeat a short conversation test.” This is more actionable than “audio needs improvement.”

Use these decision levels:

Do not try to make the raw files meet a finished podcast’s loudness target or apply heavy noise reduction as part of this evaluation. The editor needs to know what was captured, including limitations. If the remote guest was absent, mark remote audio, echo, picture, and synchronization as not tested, rather than passing them by default.

Conclusion

A useful rehearsal review answers seven concrete questions: Are the voices balanced enough to understand? Is spoken audio free of clipping? Are background noise and delayed echo controlled? Does the recorded shot keep everyone framed, stable, and visible? Do audio and video agree at the clap and near the end?

Keep your findings under the rehearsal’s take identifier, make one targeted change at a time, and verify important fixes with another short recording. In the next lesson, you will use these results to build a recording-day run sheet, including checks, responsibilities, interruption procedures, and a backup plan.