You sing alto in a church choir. Last Tuesday, someone propped a phone on a music stand and recorded the whole rehearsal. Listening back at home, you can hear the problem immediately: the piano dominates the first verse, and your section is buried under everyone else. You want a practice file where the accompaniment is clear enough to sing along with, and a second file where the voices are loud enough to check your entrances.
This walkthrough follows a single evening of working through that recording with ShengXi Vocal Remover. There are no perfect outcomes promised here — choral recordings are messy, rooms are noisy, and a phone microphone does not capture a balanced mix. What you get is a repeatable way to make decisions.
Step one: bring the file in and confirm what you have
The rehearsal video is in Photos. Open the app and import it from there. Everything — extraction, separation, export — happens on the device, so there is no upload step and no waiting on a server connection. That also means the recording stays where it started: on your phone.
Before choosing any mode, spend twenty seconds assessing the source:
- How loud is the room noise between phrases? If the air conditioning hums through the whole take, expect it to survive separation.
- Is the piano acoustic or amplified? An electric keyboard recorded close to a speaker behaves differently from a grand piano in a hall.
- How compressed does the file sound? Phone videos compress audio, and heavy compression limits how cleanly voices can be pulled away from accompaniment.
Write a short note to yourself: “Goal: piano-forward version for practice, voice-forward version for checking entrances.” Two goals, two exports. Deciding this now prevents aimless tweaking later.
Step two: pick a processing mode and listen critically
The app offers fast and fine processing modes. Start with fast — not because it is better, but because it tells you quickly whether the recording is workable at all. If the fast pass leaves the piano recognisable and the choir recognisable, you have something to refine.
Checkpoint: after the first pass, play thirty seconds from the middle of the piece. Can you still tell where the sopranos enter? If the accompaniment has swallowed all sense of the voices, try the fine mode before adjusting anything else.
If the first pass sounds thin or hollow, that is normal for difficult sources. Ask yourself one question before moving on: is the problem the separation, or is it the original recording? If the phone was sitting next to the tenors, the balance issue existed before you opened the app. Separation cannot fully undo a recording where one section was physically closer to the microphone.
Step three: build the practice mix
Switch to the mixed preview and use the separate vocal and accompaniment volume controls. This is where the session becomes useful rather than technical.
- Pull the accompaniment up until the piano line is easy to follow.
- Bring the vocals down but do not mute them. Hearing a faint guide vocal helps you stay in tempo during long rests.
- Sing through the first verse against this mix. If you keep drifting ahead of the piano, lower the vocal guide a little more so you rely on the accompaniment.
Then build the second version. Push the vocals forward and reduce the accompaniment. Play the passage where your entrance is uncertain — the one after the key change, for example — and listen for whether your starting note is audible before the rest of the section joins.
Adjust in small moves. Large jumps in either slider make it hard to remember what you liked two steps ago.
Step four: export, then keep one honest record
Export both versions before you close the app. Give them names you will recognise a week from now — something like “Tuesday rehearsal – piano lead” and “Tuesday rehearsal – voices forward.” The recent-history list keeps earlier passes available, so if you want to revisit the fast-mode result and compare it against the fine-mode one, you can.
Do one last check with headphones rather than the phone speaker. Speaker playback hides a lot of separation artefacts, and you will make better decisions about which version to keep.
What to remember before the next rehearsal
A few practical notes that came out of this session:
- Two goals, two exports. Mixing levels that work for singing along rarely work for checking entrances.
- Judge the source first. Results vary with source quality, mixing, noise, and compression. If the phone was too close to one section, no setting will rebalance the room.
- Fast mode is a scouting pass. Use it to decide whether fine processing is worth your time.
- Headphones for final decisions. Especially when the recording is noisy.
One boundary worth stating plainly: you need permission to process and publish any recording. If the choir recorded the rehearsal, ask before sharing an export more widely. The app separates audio — it does not grant rights to the music, the arrangement, or the performance.
If the next rehearsal is recorded from the same spot on the music stand, the whole routine takes about fifteen minutes, and most of that is listening rather than adjusting. That is the point: a local audio processing app is most useful when the settings serve a goal you already had in mind.