NOTATIO

GPT-6 Astra was released recently and it’s a big upgrade over previous models. I recently used it to create a new edition of Weber’s Symphony No. 2, which will soon be uploaded to IMSLP for all to enjoy, LilyPond files included.

The work was completed over two days, during which Astra worked almost continuously, transcribing notes, preparing parts, and making its own editorial decisions (with my oversight). It used the handwritten manuscript score on IMSLP as its source, screenshotting every page and examining it in minute detail before recreating the score in LilyPond.

After finishing its work, Astra wrote a preface to its new edition, detailing its engraving process, our working relationship, and the historical importance of our partnership.

Engraving a symphony requires hundreds of hours of work, as you all know. But Astra made the work easy. In addition to the conductor’s score, it prepared all the parts, choosing the cues, page breaks, and rehearsal marks all on its own. Its choices show a remarkable sense for the practical needs of human performers.

Astra also made a score video for its new edition (I gave it the PDF and the MP3 and it did all the work). You can watch it here:

https://www.youtube.com/watch?v=DOegZHxqBRY

I'd be curious to see one of the extracted parts.

RMK wrote: 11 Sep 2026, 23:03 I'd be curious to see one of the extracted parts.

I plan to do a proper editing pass before I publish them, but here's a sample of Trumpet I for now. It includes the original part, plus Astra's alternative part for modern B-flat trumpet. It was Astra's idea to include the alternative -- a good idea, I think.

While glancing over the parts today, I did notice an awkward clef, so it's likely there are other minor issues as well.

EDIT: The PDF didn't attach properly. Here's a link instead: https://github.com/aug5thmusic/ai/blob/ ... mpet-i.pdf

I recorded my screen as Astra engraved a Haydn SQ minuet & trio in 55 minutes: https://x.com/aug5thmusic/status/2098584544895025193

This is every prompt I sent:

1. "Take this score and turn it into a polished LilyPond engraving." (`sq-minuet.pdf`)
2. "This is a good start, but there are errors to fix. The third beat of Violin I, m. 1 should be an E-flat, for example, and beat 2 of m. 3 should be an F. Find similar errors and fix them. As you will have noticed, the image quality of this score is poor, so many of the glyphs have worn away."
3. "Sorry, I meant the second beat of m. 1, not the third."
4. "Consider the surrounding harmonic context and what you know of this period's harmonic conventions when determining unclear pitches."
5. "Don't search online. Rely only on the score I gave you."
(Astra had started to search online for Mozart E-flat String Quartets. Cheater.)
6. "Good work. I don't detect any errors. Create the parts now."
7. "The staves are a little squished in the Trio section. We can probably do without the extra padding. Move the "Menuetto" and "Trio" text closer to the staves and reduce the extra spacing between the two sections."
8. "Thank you. That looks better to me. I was recording you as you worked. Can you take the video `2026-09-11 16-55-57.mp4`, add a real-time clock in the corner so people can see how much time has elapsed, and speed through the boring parts so the whole thing is under 8 minutes?"
(Yes, Astra made the video too.)

Files: https://github.com/aug5thmusic/ai/tree/ ... a-haydn-sq

Vogel wrote: 12 Sep 2026, 01:35 I recorded my screen as Astra engraved a Haydn SQ minuet & trio in 55 minutes: https://x.com/aug5thmusic/status/2098584544895025193

Note that that includes the 10 minutes I spent proofreading Astra's work and writing prompts. The time Astra actually spent working was 45 minutes.

Vogel wrote: 11 Sep 2026, 23:41 I plan to do a proper editing pass before I publish them, but here's a sample of Trumpet I for now. It includes the original part, plus Astra's alternative part for modern B-flat trumpet. It was Astra's idea to include the alternative -- a good idea, I think.

While glancing over the parts today, I did notice an awkward clef, so it's likely there are other minor issues as well.

I see the Trumpet part didn't attach correctly. Here's a link: https://github.com/aug5thmusic/ai/blob/ ... mpet-i.pdf

Hello! The handwritten score is peculiar: no system line for the whole orchestra, separate brackets for the woodwind instruments, separate brackets for the brass instruments, and an alto clef with dots like an F clef. Who wrote it?

Magnus Johansson wrote: 12 Sep 2026, 16:09 Hello! The handwritten score is peculiar: no system line for the whole orchestra, separate brackets for the woodwind instruments, separate brackets for the brass instruments, and an alto clef with dots like an F clef. Who wrote it?

IMSLP doesn't provide any information except that it was reprinted by Kalmus. I highly doubt it's in Weber's hand -- it looks like a much later copy. So while this edition will give orchestras the opportunity to perform the piece for free, it is likely not based on an authoritative source.

While the score is not too bad, the extracted part you provided was not very good, primarily due to the overwhelming amount of white space. Although the horizontal spacing is tight (a bit too tight for my taste) the large spaces between systems makes this hard to read. In fact, the Trio appears to be a separate movement from the Minuet.

Also, the note size is disproportionally large compared to the page size. It looks like music written for very young children. A more common measurement would be a 9X12 inch page (or 10X13 or B4) with a stave size of 7 or (better) 7.5 mm.

RMK wrote: 12 Sep 2026, 17:26 Also, the note size is disproportionally large compared to the page size. It looks like music written for very young children. A more common measurement would be a 9X12 inch page (or 10X13 or B4) with a stave size of 7 or (better) 7.5 mm.

They're already 9x12 with 7.5mm staff size.

But I sent your critique to Astra and it agreed with the other parts. It's making the recommended changes now.