You have edited a short home video and the music sitting under it is fighting the spoken parts. Before you publish, you want to separate vocals and accompaniment from audio file sources so you can hear each layer on its own, then decide which one should sit lower. This walkthrough follows that exact situation from import to export inside ShengXi Vocal Remover, and it stays honest about what separation can and cannot fix.
Why a Sound Check Comes Before Any Editing Decision
A sound check is not about making the mix perfect. It is about answering one question: does the balance between voice and music support what the video is showing? When the vocals and accompaniment arrive as one combined track, that question is hard to answer by ear alone, because turning the whole file down also turns down the part you wanted to keep.
Separation gives you two controllable pieces. From there you can preview them together, adjust each side, and export a version that fits the video. Treat this as a listening exercise first and an export task second.
What this method assumes
- The source is a video from Photos or a local audio file you already have on the device.
- You have permission to process the material and to publish the finished video.
- You accept that results depend on source quality, mixing, noise, and compression, so two files will not behave identically.
Import the Source and Choose a Processing Mode
Open the app and bring in your file. ShengXi Vocal Remover accepts videos from Photos as well as local audio files, and the extraction, separation, and export all happen on the device. Nothing needs to be uploaded to a server, which matters when the footage includes family members or a location you would rather not send elsewhere.
The app offers two processing modes: fast and fine. A sensible first pass is to run the fast mode, listen, and decide whether the result is good enough for your video. If the combined track is dense, heavily compressed, or recorded in a noisy room, the fine mode is the better second attempt. Neither mode promises an identical or perfect split, and you should not expect one.
Read the source before you judge the result
Ask yourself three quick questions while the file is importing:
- Is the music loud relative to the voice, or are they already close in level?
- Does the recording sound clean, or is there background hiss, traffic, or room echo?
- Was the audio compressed somewhere along the way, for example by a messaging app or a video editor?
These answers shape your expectations. A clean, well-mixed file usually separates more cleanly than a noisy phone recording, and knowing that in advance saves you from chasing a result the source cannot support.
Use the Mixed Preview to Judge Balance
After separation, the app lets you preview the vocal and accompaniment parts mixed together, with separate volume controls for each. Start with both at a moderate level and listen to the section of the video where the problem is most obvious — often the opening seconds or a moment when someone speaks over the music.
Try a short sequence of small adjustments:
- Lower the accompaniment a few steps and listen for whether speech becomes clearer.
- Raise the vocal slightly and check whether it starts to sound thin or artificial.
- Return to the original balance and compare. Sometimes the separated version confirms the original was fine.
Keep notes on which setting felt right. Because the app stores recent history, you can return to an earlier result and compare instead of starting over.
What Separation Can and Cannot Solve
This is where expectations matter. Separation is a processing step, not a repair service. If the video's audio has a persistent hum, a passing siren, or clipping from a too-loud recording, those problems will likely remain in both the vocal and accompaniment parts. ShengXi Vocal Remover does not offer noise reduction, transcription, a song library, or multi-instrument stem splitting, so plan around those limits rather than working against them.
If the separated vocal still sounds muffled, the cause is usually the source, not the volume slider. Try the fine mode once, then accept what the file allows.
You can also use the separation as a diagnostic. If the accompaniment sounds cleaner than expected but the vocal feels brittle, that tells you the original mix was probably bright and compressed. That information is useful even if you decide not to change anything.
Export a Version That Matches the Video
Once the balance feels right, export the result and place it with your video in whatever editor you are using. Export the mixed result rather than guessing at settings later, and keep the source untouched so you can return to it. The app supports sharing and cleanup, so you can remove intermediate files that you no longer need once the final version is saved.
A short permission reminder
Before you publish, confirm you have the right to use the music and the footage. The app processes what you give it, but it does not grant rights to any song or video. Permission is your responsibility, not a feature of the tool.
Common Questions About This Workflow
Should I always use fine mode?
No. Fast mode is a reasonable first attempt and often enough for a home video. Reach for fine mode when the combined track is dense or when the first result leaves the vocal hard to follow.
Can I adjust only one side?
The app provides separate volume controls for the vocal and accompaniment parts, so you can lower one without touching the other. Use the mixed preview to hear how the change affects the whole.
Will the result sound exactly like the original recording?
It will resemble it, but separation is an estimate. Results vary with source quality, mixing, noise, and compression, and some material separates more cleanly than others.
Where do my files go?
Processing happens on the device, and nothing needs to be uploaded to a server. Results are kept in recent history so you can revisit them, and you can clean up files you no longer need.
The sound check is finished when you can watch the video once, without stopping, and hear both the voice and the music doing their jobs. If that happens, the separation did what it needed to do.