Remix your track stem by stem — like in a studio
- Stem separation
- Mix diagnosis
- Drum replacement
- 808 and synth bass from the notes
- Remixing
- Mastering
- Instrumental and a cappella
- MIDI
Upload a track — we'll split it, find the problems and mix it again
You need a free Stream-Q account. The first 3 separations and 3 full mixes are free.
What a mix engineer does — and what the studio does
Mastering a stereo file won't fix what's broken in the mix: if the vocals sit 3 dB below the beat, they're still quieter after mastering; if the kick and bass fight over 60–120 Hz, an EQ on the master ruins both. That's why engineers ask for stems — separate tracks — and fix the problem where it lives.
The studio does this automatically: it splits the track, measures every stem and how they relate (the way an engineer listens, only in numbers) and builds a processing chain for each role — vocals, drums, bass, music.
- Vocals: low cut, dynamic cuts of “mud” and harshness exactly where they stick out in this voice, two compressors (fast and slow), a de-esser on this voice's sibilance frequency, presence and “air”, tempo-synced pre-delay reverb.
- Drums: attack, parallel “New York” compression, tape saturation; optionally your own samples under every hit or a full kit replacement.
- Bass: compression, harmonics for small speakers, mono below 150 Hz, the bass ducks under the kick (sidechain); optionally an 808, synth bass or sub following the original's notes.
- Music: low cut under the bass, clearing the 1–4 kHz band for the vocals (only while the vocals sing), stereo width.
Drum and fill replacement
Every hit is found to the millisecond: kick, snare, hi-hats, toms. Hit strength is kept, every sample gets its own micro-variations (no “machine gun” of identical hits), and the kick's polarity is aligned with the original so the low end doesn't cancel.
Fills (toms, snare rolls) stay live by default — studios do the same: triggers misfire on fast fills. Turn on fill replacement if you need it. You can also take all hits as a MIDI file and put your own samples in your DAW.
What you get
A 30-second preview is free, with an “original / new mix” switch at equal loudness (otherwise louder always seems better). The full render:
- a 24-bit WAV and MP3 320 master (streaming loudness or louder — your choice);
- an instrumental with the same mastering, and an a cappella;
- the mix without mastering — for your own mastering engineer;
- all processed stems in FLAC and MIDI of the drums and bass.
FAQ
Does a neural network mix the track?
The neural network only splits the mix into stems. The mixing is classic engineering: EQs, compressors, a de-esser, reverb, sidechain — with settings chosen from each stem's measurements. Nothing is “generated”.
Can I upload my own stems?
Yes — up to 100 files (vocals, backing vocals, kick, snare, bass, synths…). Roles are guessed from the file names and you can correct them. No separation is needed then, and the quality is the best possible.
If I change nothing, will the track get worse?
No. Everything the separation didn't assign to a stem is kept separately and added back to the sum — with no changes, the mix matches the original.
How much does it cost?
The first 3 separations and 3 full mixes are free, previews are always free. After that — Stream-Q Notes: 30 for a separation and 50 for a full mix with mastering. If something fails, the Notes come back automatically.
How long are files kept?
A project — 14 days after you last worked on it, finished mixes — 30 days, previews — 3 days.