Original Singer / New World
Cinematic jazz · warm · 110 BPM
R Mode · vocal remix studioKeep the vocal.
Keep the vocal.
Replace the world.
Upload a recording you are allowed to transform. R Mode isolates its original singer, creates completely new genre-specific music with the shared local engine, cleans the generated backing, and combines both stems on one automatic beat grid.
Local after setup · no per-render credits · Windows 10/11 x648Fast Local generation steps
12Balanced generation steps
20Maximum generation steps
WAVfull-length result
THE COMPLETE R PIPELINE
It is more than a genre filter.
R Mode does not simply place an effect on the old song. It separates the singer, builds independent music, cleans the new backing, matches both parts musically, and renders a new stem-based mix.
Keep the original singer
BLENDLINE's bundled local stem separator extracts the lead vocal from an authorized source recording. The old backing is not carried into the final remix.
Read the source rhythm
The app estimates source tempo and finds the first sustained vocal phrase so it has a musical entrance point instead of starting at an arbitrary moment.
Create completely new music
The shared local engine generates an independent instrumental from your genre, mood, production notes and target BPM.
Remove unwanted new singing
A second local separation pass keeps the generated instrumental and removes any voice or vocal texture the music engine accidentally created.
Protect the vocal while matching tempo
Rubber Band tempo processing preserves vocal formants while adapting the isolated singer to the target BPM, reducing the robotic or chipmunk effect.
Lock, duck and render
R Mode snaps the vocal entrance to the nearest beat, lightly ducks the new backing under the singer, and exports a full local WAV.
BETA 7 PERFORMANCE PROFILES
Choose time versus refinement.
All three profiles create the complete requested song. They change the number of neural music-generation steps, not the output duration.
8 steps
The practical default for quicker iteration and slower local computers.
12 steps
More refinement when you can wait longer for the new instrumental.
20 steps
The slowest R profile for final renders where refinement matters most.
FOUR SIMPLE CHOICES
You direct. R Mode handles the pipeline.
The app exposes the creative decisions that matter while keeping the complicated separation and timing work automatic.
Choose a source you may transform
Use an owned or licensed recording. Common audio and video formats are converted on the computer.
Describe the new production
Select genre, mood and BPM, then add notes such as live brass, tight drums, darker harmony or a larger chorus.
Create the vocal remix
Keep BLENDLINE open while the local separator and shared engine complete the six-stage pipeline.
Listen, correct and export
Use the responsive player to seek through the result. If the entrance needs a major correction, use the compact -5, Auto or +5 alignment controls, then download the WAV.
WHAT THE APP ACTUALLY USES
Clear inputs. Clear output.
R Mode uses the large optional shared engine, so it does not need a second music-generation installation. Its bundled stem model handles vocal isolation and cleanup.
PartR Mode behaviorImportant limit
Original vocalSeparated, tempo-matched and retainedSeparation quality depends on the source mix
Original musicRemoved from the final workflowHeavy vocal reverb may leave traces
New musicGenerated from your genre directionGenerative results vary between renders
TimingAutomatic beat lock with compact correctionsIt does not rewrite the singer's phrasing
R MODE FACT SHEET
Requirements and realistic expectations.
“Local and unmetered” means your computer performs the work without a metered BLENDLINE cloud service. It does not mean every computer renders at the same speed.
Source
One authorized audio or video recording
Direction
Genre, mood, target BPM and production notes
Engines
Local stem separation + optional ACE-Step 1.5
Output
Original vocal over a new instrumental WAV
Privacy
Standard processing remains on the computer
Cost
No per-render cloud credits from BLENDLINE
GET A CLEANER RESULT
The source still matters.
Use the cleanest recording available. A centered lead vocal with moderate reverb is easier to isolate than a live recording, a choir, or a singer buried under loud instruments.
Prefer lossless or high-bitrate audio
WAV, FLAC or a clean high-bitrate file gives the separator more useful detail.
Start with a close target BPM
Large tempo changes can stretch consonants and make phrasing feel less natural, even with formant protection.
Describe instruments, not artists
Use production language such as smoky jazz quartet or bright synth-pop rather than requesting a living artist imitation.
Review before publishing
Listen for vocal bleed, timing errors and generated similarity. A BLENDLINE export is not automatic copyright clearance.
LOCAL VOCAL REMIXING
Download BLENDLINE →Keep the performance. Change everything behind it.
Install the shared engine once, choose R Mode, and transform only recordings you have the right to use. A generated remix is not automatic copyright clearance.