locked
feature request - Load vocals from file. (Migrated from community.research.microsoft.com) RRS feed

  • Question

  • Fizmarble posted on 01-08-2009 7:42 PM

    It would be great to load up a .wav, or even an .mp3 of an a capella vocal line and have songsmith put music to it.  I was a little surprised to not see that option.  As a tool being marketed to professionals, or at least experienced musicians that is a feature that needs to be included.  Also, unless the singer carries a tuning fork or instrument with them, it's not always likely that the melody that comes out of your mouth is in the same tuning as the midi accompaniment.  Therefore, a simple midi pitch bend to accomodate vocals that are relative to each other but out of tune with the music.  If there is a better place to make product suggestions, let me know.

    Thursday, June 2, 2011 8:27 PM

Answers

  • nancyg43 replied on 01-11-2009 11:50 PM

    per Dan's earlier message:
    "1) Many sound cards have a "loopback" or "what  you hear" input.  If you enable that input instead of your microphone, open up Songsmith, turn the master volume down, press record, and play your audio file, it will record right into Songsmith."...

     I tried this and it worked out great!  My only problem was my vocal timing in the 'wav' file was off, but I still got some fantastic 'midi' files out of it when I exported the music only.

    I always use a Microsoft headphone with mic, so I used the "options' button in SS and enabled my soundcard rather than the headphone mic and playback, (microphone configuration) then set my 'wav' audio file up in my old standby audio program, (also changed my options there to soundcard rather than headphones) punched 'record' in SS and 'play' in my audio program, and yes, it worked!  I just had to remember to switch it all back to hear how it all worked out!

    Thursday, June 2, 2011 8:27 PM

All replies

  • dmorris replied on 01-09-2009 3:50 AM

    These are _great_ points, Fizmarble.

    Just so other readers know, I'm going to give some fairly technical answers here targeted toward musicians and/or technical folks, so if you're casualy browsing the forums, this thread might get a little dense.  You've been warned. :)

    --

    Regarding tuning: internally, Songsmith shifts its transcription of your pitch so that it best lines up with standard tuning.  So even if you sing exactly halfway between two keys, the chords will come out fine.  Then you can just press "record" again and sing on top of the chords, and you'll probably be right on key the second time.  As soon as you stat moving sliders around, Songsmith will be using your NEW vocal recording, and now it's like the first recording was just a scratch take to get you started.

    --

    Regarding importing externally-created audio files... great idea.  In fact, we're planning to release a couple of small standalone utilities as soon as things cool down after our product release to do things just like this that pro users might want access to.

    The reason this isn't supported in the app itself is that  when you record something in Songsmith, you sing along with a drum beat, so Songsmith knows your tempo and knows where measure boundaries are, which our algorithms need.  Most users don't have any other way of recording audio at a steady tempo that starts right at the beginning of the first measure.

    But, if you're a musician (and it sounds like you are, and you know about midi, etc.), you might have some other way of recording fixed-tempo audio, or you might be doing some freeform soloing where tempo isn't important.   And while there's no "official" way to stick your own audio into Songsmith, here are a couple "advanced hacks" that will let you do this right now.

    1) Many sound cards have a "loopback" or "what  you hear" input.  If you enable that input instead of your microphone, open up Songsmith, turn the master volume down, press record, and play your audio file, it will record right into Songsmith.

    2) Here's a secret (which of course isn't a secret now that I'm posting it on the forum :) )... .songsmith files are just .zip files.  If you create a nice long .songsmith file, rename it to .zip, and unzip it, you'll see a single .xml file and a single .wma file.  If you replace that wma file with another wma file (make sure the name doesn't change), then zip the files back up and rename back to .songsmith... voila, new audio. :) 

    Now there's a catch here... Songsmith chooses a key when you first sing your song, and sticking a new wma file in won't change the key of your song (something else we'd like to support via separate tools and/or in future versions).  If you're at all interested in this process (suggestion (1) might be a much faster solution), reply to this thread (or - even better - create a new thread over in the "musicians' corner" forum) and we'll give you even more details that will let you specify the key when you do this.

    --

    These are great suggestions and, in addition to thinking about this for possible future versions, we'll try to release tools that do this kind of thing ASAP, or give other people enough information to build tools that do this. We're musicians too and want to help out "pro" users as much as we can!

    Thanks for trying Songsmith!

    -Dan

     

    Thursday, June 2, 2011 8:27 PM
  • Fizmarble replied on 01-09-2009 1:01 PM

     Thanks for replying.  I messed around more last night with Songsmith and discovered that it did change my pitch.  I also discovered what you said about recording the melody over the chords that were laid down by the previous take.  I appreciated that functionality.

     Now, if I am allowed to make some more requests...  Let's say I have a cool melody that features a 3/4 bar on the 3rd measure of the verse, or a chorus all in 7/8, I would like to be able to select which bars in advance will have what timing.

    I realize the danger in trying to do too much.  People start asking for a multitrack sequencer, or someone wants reverbe or beat synced delay effects and the scope of the project is shattered.  I only brought up what I did because I think it's a fantastic product idea that could truly be realized as a professional tool with some added functionality.  

    Some of my suggestions might involve reworking how the midi engine generates new chords, which might be a limitation of PG music's software.  I don't pretend to know all of the answers.  I honestly want to see this software take off, and at this point I can't justify a 30 dollar purchase.  And did I read it wrong, or is it 70 dollars for some of the added "styles"? 

    Thanks again for your involvement with the forum and your innovation in music software.

    Thursday, June 2, 2011 8:27 PM
  • AudioCave replied on 01-10-2009 9:24 AM

     I think the danger here is trying to make this a daw, it's not.  To be more useful for more knowledgable users you shoud allow the user to just enter bars and chords right off the bat.

    Manually enter 32 bars and then manually edit the chords without having to let it make that first recording pass.

    Of course the immediate problem with auto music generation is that after a while you've heard all the songs already.

    Thursday, June 2, 2011 8:27 PM
  • nancyg43 replied on 01-11-2009 11:50 PM

    per Dan's earlier message:
    "1) Many sound cards have a "loopback" or "what  you hear" input.  If you enable that input instead of your microphone, open up Songsmith, turn the master volume down, press record, and play your audio file, it will record right into Songsmith."...

     I tried this and it worked out great!  My only problem was my vocal timing in the 'wav' file was off, but I still got some fantastic 'midi' files out of it when I exported the music only.

    I always use a Microsoft headphone with mic, so I used the "options' button in SS and enabled my soundcard rather than the headphone mic and playback, (microphone configuration) then set my 'wav' audio file up in my old standby audio program, (also changed my options there to soundcard rather than headphones) punched 'record' in SS and 'play' in my audio program, and yes, it worked!  I just had to remember to switch it all back to hear how it all worked out!

    Thursday, June 2, 2011 8:27 PM