On 9/16/2013 1:36 PM, Eric Dannewitz wrote: > Honestly though, what sort of "breakthroughs" are you wanting? I'd > rather see refinements of stuff that is already in there but doesn't > work well. Like audio to pitch, the whole parts thing in Finale, > making play back for scores better, maybe even being able to create > smartmusic files with recorded audio, etc... I can't see how they are > going to come up with "breakthroughs". That is like asking Apple to > have a new product category every 3 to 4 years. It's insane.
I would consider audio-to-notation to be a breakthrough. Perhaps we are not talking about the same thing. There will be a time when a program will be able to analyze a recording and make a really good stab at building a notation score from that audio. After all, humans do this all the time. It is called transcription. There is no fundamental reason why a computer cannot be programmed to do a better job than humans do. The Melodyne product already gives us a look at that part of our future. Melodyne can take an audio file of a polyphonic instrument and accurately recognize all the notes and their associated overtones. And it can already save that information to MIDI. What it cannot do today is take a mix of multiple instruments and break that apart the same way, but it clearly is a solvable problem. I don't really understand a preference for the pedestrian and an antipathy for the big ideas. I certainly want to see bugs fixed, but really, our expectations should be higher than that. It should not be an either-or proposition. When word processors first came out, they were mostly a digital model of the process used with a typewriter. But then we got generational improvements that went way beyond that. When digital recording software first came out, it mostly modeled the reel-to-reel recording process and still embodies a lot of that terminology, but today's DAWs go far beyond what was possible with reel-to-reel recording decks. It is not acceptable to me to say that notation software should just do a cleaner, faster job of what was done with quill and parchment for 500 years. But that's essentially where we are -- still stuck in that first generation vision for the most part. I'm not denying there have been some improvements, but really the products just don't exemplify the vision we have seen in most other technology fields. S0 here are some other examples of things that should be part of the vision: 1) In 2013 I shouldn't still have to fiddle with layouts on my parts. How many Finale releases have we seen that bragged about great new algorithms that avoid collision of printed elements? Yet, I still have to manually edit every &^%%$#$#% part by hand just to achieve the most minimally readable parts. That would be a breakthrough -- one that might have expected in 2000, not 2014. 2) How long have we had spell checkers and grammar checkers in word processors? 15-20 years anyway. Why don't we have the same things in music notation by now? Why isn't there a function that says "In measure 14, 2 instruments have B naturals that are in conflict with the B-flats in 5 other parts." Why isn't there a function that evaluates my voice leading and suggests better options? 3) Much of harmonization is more-or-less formulaic. Finale offers a little help with the BIAB Harmonization plug-in, but by now, the state of the art should be much farther down the road. I'd like to point the notation program to an audio snippet from an Elgar symphony and tell the notation program "Suggest a harmonization of my melody that is like that." OK, I don't expect that in the next year, but people developing notation programs should have goals like this -- a much bigger vision than just how big to make the margin on the page. 4) There is a little support for drum grooves. That's good for 2003, but in 2013, the DAW world is now very advanced in looping. I ought to be able to draw upon a library of 1000 drum patterns and loop them in my score just as I would in a DAW. Likewise for the other instruments. What sense could it ever possibly make for me to have to write drum parts from scratch each time? 5) The whole concept of chords has always been a complete mess in Finale. The playback is so bad that nobody will use it, and you can't use any common system of chord spelling without spending loads of time building your own library. What I want to do is type in a chord and have the notation do a sensible comping from that. I don't expect it to be as powerful as BIAB, but I note that Sibelius has a plug-in that does some of this. This doesn't seem like a really heavy lift. 6) Likewise, the notation program should be able to read the notes in all the staves and determine the most likely chords in use. In many scores, there would be enough information for the program to guess 80% of them correctly, and then I could spend my time on the other 20%. And the same should apply to any harmonization wizards. It should not be necessary for me to type chord names. The notation program should be able to read what is already in the bass, piano, and guitar to determine how to harmonize the sax soli. All of these things would, of course, be optional. A person who only wants to model quill and parchment would never be forced to use the advanced capabilities, just as nobody is forced to use a spell checker. _______________________________________________ Finale mailing list [email protected] http://lists.shsu.edu/mailman/listinfo/finale
