Original Vocal / Festival Trance
Uplifting · 138 BPM · Maximum
T Mode · flagship trance engineYour vocal.
Your vocal.
Festival scale.
Turn one authorized song into a completely new trance mix. T Mode isolates the original singer, creates and scores new trance productions, protects the vocal's natural timing, locks the entrance to the beat, builds bass and sidechain movement, then masters one local WAV.
In development for the next Windows beta · local after shared-engine setup · no per-render credits1Fast neural take
2Strong scored takes
3Maximum scored takes
124–148target BPM range
THE COMPLETE T PIPELINE
More than a trance preset.
T Mode combines neural composition with deterministic audio engineering. The AI creates musical choices; BLENDLINE handles separation, selection, beat placement, bass reinforcement, sidechain movement and safe mastering.
Recover the authorized lead vocal
The local separator removes the old backing and keeps the source singer as an editable stem.
Choose the trance identity
Select Vocal, Uplifting, Progressive, Psy, Tech or Dream / Euro Trance, then set energy, BPM, bass force and production notes.
Create real neural productions
ACE-Step 1.5 composes full-length instrumental trance locally. Strong creates two takes; Maximum creates three.
Keep the strongest take
BLENDLINE measures pulse stability, useful low-end energy, clipping and clarity, then automatically selects the best candidate.
Protect the singer's natural speed
Vocal Speed Lock keeps the original pace. BLENDLINE moves the sustained entrance onto the beat instead of stretching the complete performance.
Build pump, bass and headroom
A final deterministic pass reinforces existing low end, adds quarter-note sidechain movement, ducks the music under the singer and exports a peak-safe WAV.
REAL QUALITY LEVELS
Quality means more chances.
Each level changes actual neural work. It never shortens the song into a preview. Strong and Maximum take longer because they create complete alternatives and score them locally.
1 take · 8 steps
Best for testing a style or for slower computers. One full production is cleaned, aligned and mastered.
2 takes · 12 steps each
The flagship default. Two independent productions compete on beat stability, bass, clarity and clean level.
3 takes · 20 steps each
The longest local render. Three more-refined productions are cleaned and scored before the winning master is built.
SIX TRANCE DIRECTIONS
One engine. Different club worlds.
The style changes the composition prompt and target production language, while the same vocal-protection and mastering pipeline remains active.
VOCAL TRANCE
Wide supersaws, emotional breakdowns and space for a clear lead singer.
UPLIFTING
Long builds, bright harmony, rolling bass and a high-energy release.
PROGRESSIVE
Deeper groove, patient development and cleaner, restrained peaks.
PSYTRANCE
Tighter rolling low end, sharper rhythmic detail and faster hypnotic motion.
TECH TRANCE
Harder drums, darker tension, shorter motifs and peak-time impact.
DREAM / EURO
Melodic hooks, nostalgic synth color and direct dance-floor structure.
DISK AND SPEED
Free to run. Real work to compute.
"Free and unlimited" means the standard workflow runs on the user's computer without a metered BLENDLINE cloud service. It does not promise identical speed on every machine.
About 8–12 GB
The safe 2B ACE-Step path, runtime, separator and cache. If R Mode is already installed, T Mode reuses the same engine instead of downloading it again.
Roughly 5–20 minutes
A typical full Strong render can fall in this range on a modern 8–16 GB GPU. Source length, thermal limits and model loading still matter.
30–120+ minutes
Compatibility processing can be much slower. The app calculates a local estimate and keeps Fast available for safer iteration.
Maximum normally costs about three times the neural generation work of Fast. Estimates are ranges, not promises. Generated working files may temporarily use another 1–3 GB per active project.
NATURAL VOCAL POLICY
No robotic full-song stretching by default.
Vocal Speed Lock starts on. The original vocal stem keeps its timing and pitch; BLENDLINE snaps the first sustained phrase to the nearest beat and lets the new production move around it. Turning the lock off enables optional formant-preserving tempo matching.
ControlT Mode behaviorTrade-off
Speed Lock onOriginal vocal pace is preservedBest naturalness; the production must follow the phrasing
Speed Lock offFormant-preserving tempo matchCloser grid fit; large changes can sound processed
Bass forceReinforces existing generated low endDoes not replace weak source composition
Candidate scoringSelects the strongest measured takeMusical taste still requires a final listen
NEXT DESKTOP BETA
Get the current beta →The strongest BLENDLINE mode yet.
T Mode is implemented in the development build and is now moving through full local-engine validation before the next public Windows installer. It reuses the open local ACE-Step engine and does not add per-song credits.