Part of our guide to transcribing a full band.
There are two ways to turn a recording with several instruments into notation. You can transcribe everything in one pass, with every instrument written out from a single upload, whether that is a rock rhythm section, a jazz combo with sax and trumpet, or a string quartet. Or you can transcribe one instrument at a time, focusing the model on the part you care about. Songscription now does both: Multi-Instrument handles the one-pass route, and single-instrument mode handles the focused one. Neither is always better. Each has a clear job, and knowing the tradeoffs saves you time and cleanup. This guide explains how the two approaches work, what each does well, and how to pick between them, plus a third option for when you just want one playable version of the song.
The two approaches
One-pass multi-instrument transcription takes a full mix and writes out every part in a single step. In Songscription you pick "Multi-Instrument" in the instrument picker, upload a recording (a file, or a YouTube, TikTok, or Instagram link), and get a separate track for each instrument it hears: keys, guitars, bass, strings, brass, woodwinds, saxophones, synths, voice, and drums. The result is a multi-staff score plus multitrack MIDI you can open in Ableton, Logic, FL Studio, or GarageBand. Every keyboard track is split into left and right hands.
The other approach is to transcribe one instrument or one part at a time. You pick the instrument you want and the model isolates that part from the full mix on its own, so there is no need to split the song into stems first. It is one part per run, but the model has a single, narrow target, which is why it tends to produce the cleanest version of that part. For the bigger picture of what AI can and cannot do here, see whether AI can transcribe multiple instruments.
What one-pass does well
The obvious win is speed. One upload gives you every part at once, already laid out as a full score, with no need to run the song five times and line the results up afterwards. That makes it ideal for getting a whole arrangement on paper quickly, handing each player their chart, studying how the parts fit together, or pulling a multitrack MIDI file into a DAW to remix, reharmonize, or rebuild the song with your own sounds.
The honest tradeoff shows up in dense mixes. The model is untangling every instrument at once, and a bass note, a guitar chord, and a kick drum can share the same moment and the same frequency range. Most parts come out usable, but some will need cleanup in the built-in editor, and accuracy is highest on clean, well-balanced recordings. The same goes for bigger ensembles: it recognizes strings, brass, and woodwinds, but the more instruments overlap in the same range, the more cleanup to expect. A choir comes back as one voice track rather than separate soprano, alto, tenor, and bass parts.
What per-instrument does well
The reason per-instrument transcription is cleaner comes down to focus. When the model only has to write down the notes of one part, it does not have to decide which note belongs to which instrument across the whole mix. The narrower the target, the more accurate the output. Songscription models are generally robust enough to isolate the instrument you want directly from a full mix, and you choose the difficulty for that part instead of accepting a default.
Because the model pulls the part out of the mix on its own, you do not need stems for either approach. On an especially dense or crowded recording, optional stem separation can make a part cleaner still by pulling the guitar, the bass, or the vocal line out before transcription, but it is an aid rather than a requirement. For the step-by-step version of the focused workflow, the instrument transcription guide walks through transcribing individual parts.
Combining the two
You do not have to choose one approach for a whole project. A practical pattern is to run Multi-Instrument first to get every part, then re-transcribe the one part that matters most in single-instrument mode. If you are a bassist learning a song, the Multi-Instrument score shows you how your line sits against the drums and guitar, and a single-instrument bass transcription gives you the cleanest version of the line itself. Every transcription is editable and downloadable as PDF, MIDI, or MusicXML, so you can swap a cleaner part into the arrangement in a notation program.
If you are starting from separate tracks rather than a finished mix, transcribing each stem on its own is the cleanest route of all, and transcribing multi-track audio to sheet music covers that workflow. If you only have the finished mix of a full band, transcribing a full band recording walks through Multi-Instrument end to end, with per-instrument and piano cover as alternatives.
When you want one playable version
Sometimes you do not want separate parts at all. You want to sit down and play the song. Piano cover mode, piano plus vocals, and lead sheet modes condense a whole song into a single playable part instead of writing out every instrument. That is the right choice when the goal is a piece you can perform yourself rather than a record of what each player did.
Which approach to choose
Use Multi-Instrument when you want every part fast, in one upload, and are happy to tidy a few spots in the editor. Use single-instrument mode when you need one part as clean as possible. Use piano cover or lead sheet mode when you want one playable version of the whole song. Stems are optional in every case. If you are not sure, start with Multi-Instrument to see every part, then switch to single-instrument mode for any part that needs more precision.
Frequently Asked Questions
Is one-pass multi-instrument transcription better?
It depends on what you need. One-pass transcription, which Songscription offers as Multi-Instrument, is the fastest way to get every part of a recording, with a separate track for each instrument it hears, from keys, guitars, and bass to strings, brass, saxophones, and drums, from a single upload. Per-instrument transcription is the better choice when you need one part as clean as possible, because the model focuses on that instrument alone. Many people use both, Multi-Instrument for the whole song, then single-instrument mode on the one part that matters most.
When should I transcribe one instrument at a time?
Transcribe one instrument at a time when a single part matters more than the rest, such as the bass line you are learning or the piano part you are performing. In single-instrument mode the model isolates that instrument directly from the full mix, so you do not need to split the song into stems first, and you choose the difficulty for that part.
How do I get a full score in one pass?
Choose Multi-Instrument in Songscription's instrument picker and upload the recording as a file or a YouTube, TikTok, or Instagram link. It transcribes every instrument in one pass and gives you a multi-staff score with a separate part for each one it hears, whether that is a rhythm section, a horn section, or strings, plus multitrack MIDI you can open in a DAW. Some parts may need cleanup in the built-in editor, and accuracy is highest on clean recordings.
Does per-instrument transcription take longer?
Yes, for a full band it takes more steps, because you transcribe each part separately and then assemble them. Multi-Instrument does every part in one upload, so it is the faster choice when you want every part. For a single part, per-instrument is just as fast and usually cleaner, since you only run it once and the model concentrates on that one instrument.
The quickest way to see the difference is on a recording you know well. Upload it with Songscription, try Multi-Instrument for every part or pick one instrument, and compare.
