Please correct the errors below.
The player Learn and teach using sheet music synced with video Notation and tab editor Easily create interactive sheet music, for free Sheet music scanner Beta Turn PDFs and photos into interactive sheet music About Soundslice Here’s our story
Help Transcribe Embed Plans Community
Practice Teach Embed Transcribe Community Plans Help Store
Log in Sign up

Lyrics support in PDF scanner

Our PDF/image scanning feature, launched in beta in September, now supports lyrics!

Screenshot of editor showing original image and detected notation, including lyrics

When you upload a PDF or image, you’ll now see the lyrics in your slice. Our lyrics editor makes it fast and easy to tweak the text in case of misdetections, though in our experience the detection is good as long as the photo is reasonably clear.

We’ve trained our machine-learning system specifically to detect lyric syllables and run OCR (optical character recognition) on the individual letters. We took extra care to detect and preserve hyphens between syllables, and we should do the right thing for “extenders” (the horizontal lines that specify a syllable should be held over multiple notes).

For now, it only works for languages that use Latin characters, such as English.

We still consider our PDF/image scanning to be in beta, but this was one of our highest-priority features post-launch. This makes the tool extra useful for choral directors and people teaching/practicing vocal repertoire. Enjoy!

Comments

You really need to increase the number of pages per scan. I know it's still beta but Newzik which I also use has upped their limit to 500 pages a month...
Thanks for the feedback! Yes, we'll be increasing the limits in the future.
Hi Adrian and other in Soundslice,

I tried to write this before. The import of the song's texts under the sheet music is incorrect for Czech users, due to the fact that the character set, e.g. UTF-16, is apparently not supported.

In this case, it would be easier to write the entire text manually, and before the coding of languages such as Czech (that is, with diacritics) can be solved, it would not be bad to have the option to set before loading:
[.] Import lyrics ?
[.] Import chords ?
[.] import the H chord as a B
[.] Prefer TAB over notation when PDF has both

Karel Jacko - CZECH
I haven't dealt with this issue for years. But I googled this:

The modern standard is a universal character set containing (almost) all characters of (almost) all world alphabets called Unicode, especially in UTF-8
Yes, the optical character recognition doesn't yet support diacritics, but we'd like to add support in the future! Please note that we do indeed encode lyrics as UTF-8 (which is why things work properly if you add lyrics with diacritics in our editor).
Thanks

Post a comment

Log in to post a comment.