Guides / Research
The DAW Conversion Benchmark
Round 1 · measured August 15, 2026
Every converter — including ours — claims fidelity. This page measures it: the same fixed input projects through Doseedo, logic2ableton (v2, open source) and DawVert (open source), outputs parsed field by field. Zeros were verified against the raw output files. Where Doseedo drops something, that's printed too.
Method
Fixed public fixtures; each tool run on 2026-08-15 (logic2ableton and DawVert at that day's latest GitHub main, Doseedo production API). Outputs parsed with format readers; counts are what arrives editable in the target project. An import kit of loose MIDI/stem files is honest output, but it is not a project — it's labeled as what it is. Fixture set: a known-content session (6 tracks, 25 MIDI notes, 2 placed audio clips, tempo 100, 3 markers) authored as a real .als/.logicx pair; two real FL Studio files from PyFLP's public test assets; a real 27-track, 3,977-note REAPER project from a public GitHub repo.
Logic Pro → Ableton Live
Input: the known-content .logicx (6 tracks · 25 notes · 2 audio clips · tempo 100 · 3 markers).
| Output | Tracks | MIDI notes | Audio clips | Tempo | Markers | |
|---|---|---|---|---|---|---|
| Doseedo | native .als project | 6/6 | 25/25 | 2/2 | 100 ✓ | 3/3a |
| logic2ableton | native .als project | 2/6 | 0/25 | 2/2 | 120 ✗ (default) | 0/3 |
a Round 1's first pass measured 0/3 here — a real Doseedo bug (a marker-reader asymmetry on writer-authored projects). It was fixed and redeployed the same day; the 3/3 above is the re-measured production result. Finding our own bugs is what this page is for.
Ableton Live → Logic Pro
Input: the same content as a real .als project with Samples/.
| Output | Tracks | Tempo | Markers | Content | |
|---|---|---|---|---|---|
| Doseedo | native .logicx project | 6/6 | 100 ✓ | 3/3 | 25 notes + 2 audio clips embedded (converter report) |
| logic2ableton | import kit — 4 .mid files, 4 stem WAVs, CSV manifest, instructions | — | — | — | you rebuild the project in Logic by hand |
FL Studio → Ableton Live
Input A: PyFLP's multi-channel.flp (2 channels · 2 notes · tempo 140). Input B: PyFLP's FL 20.8.4.flp (19 channels · 44 musical notes + one automation-clip channel · tempo 69.42 · pattern-mode, empty playlist; the file re-declares a pattern index, so naive note counts read double).
| A: tracks | A: notes | A: tempo | B: result | |
|---|---|---|---|---|
| Doseedo | 2 | 2/2 | 140 ✓ | 11 tracks · 44/44 notesb · tempo 69.42 ✓ · 8 markers |
| DawVert | 2 | 0/2 | 140 ✓ | crash (KeyError, FLP input plugin) |
b Pattern-mode project (nothing placed on the playlist): Doseedo lays the patterns out from bar 1. The automation-clip channel's events are intentionally not rendered as MIDI notes. Round 1's first pass also printed a 44/96 "gap" here — that one was a bug in the benchmark's own ground-truth counter (the duplicate pattern declaration), kept in this footnote for the record.
FL Studio → REAPER
Input: multi-channel.flp.
| Output | Tracks | Notes | Tempo | |
|---|---|---|---|---|
| Doseedo | native .rpp project | 2 | 2/2 | 140 ✓ |
| DawVert | native .rpp project | 6 | 0/2 | 140 ✓ |
REAPER → Ableton Live
Input: a real 27-track worship-band project from a public GitHub repo (3,977 source MIDI notes across 14 sounding tracks).
| Output | Tracks | Notes | Tempo | |
|---|---|---|---|---|
| Doseedo | native .als project | 14 (all sounding tracks) | 3,890/3,977c | 120 ✓ |
| logic2ableton | does not read .rpp | |||
| DawVert | does not read .rpp | |||
c The gap is REAPER's own item semantics: notes past an item's right edge don't sound in REAPER either, and the converter honors the item boundary. The 3,890 are the notes that actually play.
What this doesn't measure (yet)
Automation lanes, plugin state, buses/sends and warp are only exercised by the Logic↔Ableton pair here (see the field-by-field spec for those claims and their limits). Rounds with deeper fixtures — automation-heavy sessions, third-party plugin chains, tempo-mapped projects — are the next iteration, and every new tool that appears gets added. Spot an error or want a tool included? Tell us — corrections get published.
Run it yourself
The fixtures are public: PyFLP's test assets, the REAPER project, and the known-content Ableton/Logic pair is reproduced by any DAW following the counts above. logic2ableton and DawVert are open source; Doseedo converts 5 projects a month free without an account.
Test it on your own project
The benchmark uses fixed fixtures — your session is the benchmark that matters. 5 conversions a month free, no account.
Open the converter →Related
- Every Logic ↔ Ableton converter, compared
- What carries over — the field-by-field spec
- Every way to move a project between DAWs
FAQ
What is the best DAW project converter?
Measured on fixed fixtures (August 2026): Doseedo was the only tool that produced a native project with the MIDI notes, audio clips and tempo intact in every direction tested (Logic↔Ableton, FL Studio→Ableton/REAPER, REAPER→Ableton). logic2ableton moved audio stems but no MIDI notes and defaulted the tempo; its Ableton→Logic lane outputs an import kit, not a project. DawVert created tracks with correct tempo but no notes survived on the FL test files, and it crashed on one of them. Full counts and caveats — including the fields Doseedo itself dropped — are in the tables.
How is the benchmark measured?
Each tool converts the same fixed input files. Outputs are parsed with format readers and counted field by field: tracks, MIDI notes, audio clips, tempo, time signature, markers. Surprising results (like a zero) are verified against the raw output file before being reported. Fixtures are public; the exact tool versions and dates are listed.
Is the benchmark fair to other tools?
The page reports Doseedo's own failures with the same prominence — including a marker round-trip bug found during this run and note-count gaps on pattern-mode FL projects. Every zero was double-checked in the raw output file, and each tool's own scope claims are noted alongside its results.