Notation Software Users Forum  

Go Back   Notation Software Users Forum > Musician community sharing > Using Notation Software products with other (third party) products

Using Notation Software products with other (third party) products Find out from others, or share your experience, about integrating Notation Software products with sound libraries, audio processing software, and other hardware and software products.

Thread Tools Search this Thread Display Modes
Prev Previous Post   Next Post Next
Old 04-26-2016, 03:43 PM
Sherry C's Avatar
Sherry C Sherry C is offline
Product Manager
Join Date: Jul 2009
Location: Bad Axe, MI, USA (The Tip of the Thumb of Michigan)
Posts: 2,376
Default Re: Transcribing solo piano recordings


Speech recognition is tougher than a few notes, but that's not what we're talking about here

I used to work in hearing research (Kresge Hearing Research Institute at the University of Michigan in Ann Arbor, MI, USA), and research on speech recognition has been going on for decades, with a lot of funding from the government Given what I know of it as well as music, I would have to disagree as to a difference in complexity. Both are quite complex analyses - pretty much 4D. With both you have pitch (essential in some languages and definitely essential to music), duration, loudness, and timbre (formants in speech; timbre is essential to distinguishing between instruments and even different pitches from the same instrument).

Speech recognition is a much more universal application, and also more lucrative in terms of financial reward. Ergo a lot more time and effort have been spent there. Not so much (unfortunately) with musical analysis and transcription. You'll also notice that speech recognition technology typically requires some 'training' period for the software, as well as clearly spoken words by a single person. This is analogous to a single-note melody line for a clearly-played (eg. no vibrato, no reverb, etc.) instrument. Try singing to a speech recognition program some time and add vibrato and fermatas and other musical embellishments - the results will not be the same as with regular speech

As I mentioned previously, the Melodyne Editor actually does a phenomenal job of separating audio signals, but still does not put out a satisfactory (yet) MIDI for notation purposes unless it is a fairly uncomplicated piece (ie. a single-instrument with a "clean" recording.) It doesn't separate out the audio analysis on a per-instrument basis for multi-instrument pieces - that's a job that still requires a human to discern However, it is heartening that folks are still working on the problem.

I hope this helps explain part of the complexity and comparison of the technologies. Maybe it was more than you wanted to know (my kids tell me I do that often ).

Music is to the soul like water is to green growing things.

Last edited by Sherry C; 04-26-2016 at 03:57 PM. Reason: spelling error
Reply With Quote


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

All times are GMT. The time now is 01:38 PM.

Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
Notation Software Germany GmbH