How do I fix echo and bad sound in a remote interview?

This page may contain affiliate links from which Blur Studio earns commissions.

Echo is a harder problem than plain background noise, and it's worth being honest about that before you spend credits trying to fully fix it. Background noise, a hum, traffic, chatter, is a separate sound layered on top of the speech, which is exactly what noise-removal tools are built to strip out. Echo is different: it's the voice itself bouncing around a hard-surfaced room and arriving at the microphone slightly delayed and distorted, which makes it part of the voice signal rather than something clearly separate from it.

What actually helps

1. Run the recording through ElevenLabs' Voice Isolator anyway. It won't eliminate echo caused by room shape, but it does genuinely help with any separate background noise layered on top, hum, fan noise, distant sounds, which often compounds a bad-room recording and makes it worse than the echo alone.
2. Listen to the result at a realistic volume, on real speakers or headphones, not just through a laptop's built-in speaker, before deciding whether the improvement is enough to use the recording as-is.
3. Consider trimming or re-recording just the worst sections if certain parts of the interview are dramatically worse than others, rather than processing the entire file for a problem that's concentrated in one part. Fixing one wrong sentence in a recorded video shows how to patch a single section.

The file limits are the same as for any other use: up to 500MB and 1 hour of duration per file, with processing costing roughly 1,000 credits per minute of audio. Noise-only problems are covered in removing background noise from a job site video.

A worked example: two guests, one echoey, one clean

Say you recorded a remote interview where you sound fine (a decent headset, a quiet room) but your guest is in an empty home office with a laptop mic and real echo. This mixed-quality situation is common and worth handling per-speaker rather than as one blanket fix:

1. Split the two tracks if your recording software gives you separate files per participant, which most video call platforms do. Processing your clean track and your guest's echoey track together wastes effort and credits on the half that's already fine.
2. Run only the guest's track through Voice Isolator, since that's where the actual problem is. Your own clean audio doesn't need the step at all.
3. Listen specifically to whether the guest's echo decreased or just the background hum did. Separating these two effects matters for deciding whether this result is good enough to publish or whether the guest's segments need re-recording.
4. Remix the two tracks back together once you've processed the one that needed it, rather than reprocessing your already-clean audio unnecessarily.

This per-track approach is specifically useful for interviews, since it's rare that both sides of a remote conversation have equally bad audio, and treating them separately gets you a better result for less processing cost than running the combined file through as one. Making recorded customer calls easier to hear applies the same tool to phone recordings.

Be honest about what this won't do

If the echo is severe, a genuinely hard, empty room with a lot of reflective surfaces, no cleanup tool fully separates the "real" voice from its own reflections, because they're the same sound arriving at slightly different times, not two different sounds mixed together. Expect improvement at the margins, not a transformation. If the interview is important enough and the echo is bad enough, the more reliable fix is re-recording in a better space, not pushing harder on cleanup tools after the fact.

It's worth setting the right expectation with whoever's waiting on this content before you even start cleaning it up. Telling a client or a producer "I can improve this, but it won't sound like a studio recording" up front avoids a worse conversation later, after you've spent credits and time, where the honest answer is still "this is about as good as it's going to get."

Fixing this before the next recording

For future remote interviews, a few cheap, low-effort changes prevent this entirely rather than needing to be fixed afterward: recording in a smaller room with soft furnishings (a studio isn't needed, a room with curtains, carpet and furniture absorbs reflections far better than a bare space), sitting closer to the microphone so less room sound gets picked up relative to direct voice, and using a basic directional microphone instead of a laptop's built-in mic, which tends to pick up ambient room sound indiscriminately. None of this costs much, and all of it does more for your next recording than any amount of post-processing on this one. If you're coordinating interviews with outside guests you don't control, a customer testimonial, an expert quote, a podcast guest, send these three tips to them directly before the call rather than hoping they figure it out themselves. A one-line message, "please join from a smaller, softer room if you can, and sit close to your mic," costs you nothing and prevents a problem you'd otherwise be stuck cleaning up afterward with limited results.

Worth mentioning separately: most video call platforms (Zoom, Teams, Meet) already run their own real-time echo cancellation during the call itself, which is a different thing from Voice Isolator and happens before the recording is even saved. If a guest's live audio sounded fine during the call but the recording still has echo, check whether you're recording the platform's processed audio output or a separate raw capture, since some recording setups bypass the platform's own echo cancellation entirely and capture a rawer feed than what people actually heard live.

How to tell echo from plain background noise before you process anything

These two problems sound similar when you first notice "this recording sounds bad," but they need different fixes, so it's worth a quick ear-check before deciding what to do. Background noise sounds like a separate, identifiable sound underneath the voice, a hum, a hiss, traffic, chatter, something you could point to and say "that's a different sound." Echo sounds like the voice itself trailing off or doubling slightly, a hollow or "in a box" quality to the speech itself, not a separate sound you can point to. If you're not sure which you're dealing with, listen specifically for whether turning down the volume makes the problem feel smaller (more likely genuine background noise) or whether the hollow, boxy quality persists at any volume (more likely room echo baked into the voice itself). Getting this diagnosis right before you spend credits processing the file saves you from running an echoey recording through a tool that was never going to fully fix that specific problem.

Common mistakes worth avoiding

• Processing the whole interview when only one speaker's side has a problem. If you can split tracks per participant, only run the problematic one through cleanup; it's cheaper and gives you a clearer sense of whether the fix actually worked.
• Expecting a full fix and being disappointed by a partial one. Set the expectation (with yourself or whoever's waiting on the content) that echo improves at the margins, it doesn't disappear, before you spend the time and credits.
• Not checking whether your recording setup captures raw or platform-processed audio. Some setups bypass the video platform's own real-time echo cancellation, capturing a rawer, worse feed than what was actually heard live on the call. Check this once and fix your recording setup rather than relying on after-the-fact cleanup to compensate every time.
• Giving up on a salvageable recording too early. A recording with real content and moderate echo is usually worth the modest cleanup cost before deciding it's unusable; reserve the "just re-record it" decision for genuinely severe cases. If you only need a transcript, cleaning up audio before sending it to be transcribed may be enough.

How I'd approach this

For a one-off bad recording, running it through the isolator and deciding from there is the right amount of effort, there's no point over-investing in cleaning up audio you're not going to reuse often. If your business does remote interviews or recorded calls regularly and keeps running into the same room-noise or echo problems, the better fix is usually on the recording side, not the cleanup side: a cheap mic upgrade or a better-chosen recording space pays off every single time afterward, rather than fighting the same echo problem on every file. Where I'd actually get involved is a business recording interviews or calls at real volume, enough that per-track splitting, processing and remixing by hand becomes its own time cost; at that point, a small automated pipeline that handles the split, the cleanup decision and the remix is worth more than the manual version, especially once you're doing it every week rather than occasionally.