You can also browse Wikipedia:Featured articles and Wikipedia:Good articles to find examples of Wikipedia's best writing on topics similar to your proposed arti
Where to get help
How to improve a draft
You can also browse Wikipedia:Featured articles and Wikipedia:Good articles to find examples of Wikipedia's best writing on topics similar to your proposed article. Improving your odds of a speedy review To improve your odds of a faster review, tag your draft with relevant WikiProject tags using the button below. This will let reviewers know a new draft has been submitted in their area of interest. For instance, if you wrote about a female astronomer, you would want to add the Biography, Astronomy, and Women scientists tags. Editor resources
|
| This is a draft article. It is a work in progress open to editing by anyone. Please ensure core content policies are met before publishing it as a live Wikipedia article. Find sources: Google (books · news · scholar · free images · WP refs) · FENS · JSTOR · TWL Last edited by Colorado001 (talk | contribs) 11 days ago. (Update)
Finished drafting? or |
Songscription is an automatic music transcription platform that uses artificial intelligence to convert audio recordings into sheet music, MIDI files, piano roll and guitar tablature. The company was founded in late 2024 by Andrew Carlins, Alex Alvarado-Barahona, Katie Baker, and Tim Beyer and launched publicly in June 2025. It is based in the San Francisco Bay Area and has described its product as "the Shazam for sheet music."[1]
Automatic music transcription (AMT) — converting audio into musical notation — has been studied in music information retrieval for several decades. The task involves multiple sub-problems: pitch detection, onset and offset timing, rhythmic quantization, and instrument-specific notation conventions.[2] Earlier machine learning approaches typically addressed these sub-problems in separate stages, requiring complex post-processing pipelines to produce readable scores.
A 2024 paper by Songscription co-founder Tim Beyer and researcher Angela Dai, presented at ISMIR 2024, proposed treating the full transcription pipeline as a single sequence-to-sequence problem. This new model used a transformer architecture with a custom compound-token representation for symbolic music that converted piano performance MIDI directly into musical scores. It was the first method to predict notational details such as trill marks and stem direction directly from performance data, without a separate classification stage. Compared to prior deep learning approaches and hidden Markov model-based pipelines, the method outperformed existing approaches on the MUSTER transcription benchmark while reducing sequence lengths by a factor of 3.5.[3] Songscription's initial production models were built on this architecture.
Three of the four founders met at Stanford University, where CEO Andrew Carlins was enrolled in a joint MBA/MA in Education program. Carlins has said the company's mission is personal to him: he grew up with a stutter and found that singing allowed him to speak fluently, which shaped his interest in music accessibility.[4]
The company was accepted into the StartX accelerator at Stanford and received pre-seed funding from Reach Capital before its public launch.[5]
Songscription accepts audio files (MP3, WAV), MIDI files, and live recordings. Its transcriptions are exportable in four formats: PDF sheet music, MIDI for use in digital audio workstations, MusicXML for notation software such as MuseScore, Sibelius, and Finale, and Guitar Pro format for guitar and bass.
Transcriptions can be edited and played back at variable speeds in a browser-based score editor. A piano roll view is also available, displaying notes as color-coded bars against a virtual keyboard, which is intended for users who do not read standard notation.[6]
At launch, piano was the most reliable instrument. By late 2025, the platform supported piano, violin, flute, guitar, bass guitar, and trumpet, with additional instruments in development.[4]
The AI models were trained on a combination of real performances and matching scores contributed by professional musicians, public domain sheet music, and synthetic data. Synthetic examples were generated by rendering scores to audio and adding simulated acoustic conditions (background noise and reverberation) to improve robustness on real-world recordings.[1] As of late 2025, the company was in negotiations with major music publishers for licensed training material.[4]
Songscription uses a freemium model. A free tier allows limited monthly transcriptions; paid tiers — aimed at hobbyists, educators, and professionals — increase limits and add priority support.[7]
In November 2025, about five months after its public launch, the company reported over 150,000 users across 150 countries and announced a $5 million funding round led by Reach Capital, with participation from Emerge Capital, 10x Founders, and Dent Capital.[4] Guns N' Roses guitarist Ron "Bumblefoot" Thal joined as an advisor at the time of the announcement.[6]
Coverage in TechCrunch, MusicRadar, and Music Business Worldwide focused on the platform's practical value for musicians who cannot find published notation for the songs they want to learn or teach — a gap that Songscription's own survey put at roughly 60% of the repertoire its users were seeking.[5] Reviewers noted that the piano model performed well across a range of repertoire, though complex or fast-paced passages still benefited from manual correction.
{{cite conference}}: CS1 maint: multiple names: authors list (link)
| Part of a series on |
| Artificial intelligence (AI) |
|---|
| Glossary |
Category:Music technology companies Category:Artificial intelligence companies Category:Music notation software Category:Stanford University spinoffs Category:Companies founded in 2024 Category:Software as a service
Informasi ini disarikan dari Wikipedia dan disajikan kembali untuk tujuan edukasi. Konten tersedia di bawah lisensi CC BY-SA 3.0. Kami tidak bertanggung jawab atas ketidakakuratan data yang bersumber dari kontribusi publik tersebut.
- provide significant coverage: discuss the subject in detail, excluding routine coverage like product launches, staff appointments, or financial reports and listings in databases or listicles;
- are reliable: from reputable outlets with editorial oversight;
- are independent: not connected to the subject, such as press releases, the subject's own website, or sponsored content.
Please add references that meet all three of these criteria. If none exist, the subject is not yet suitable for Wikipedia.