Controlling the rate of a speech sample
I’m trying to construct a patch in MacMSP which alters the rate of a recorded speech sentence. What I want to do is to alter the rate of certain elements in the speech signal. For instance, I want to slow down / speed up certain words within a phrase or sentence. In the spoken phrase, "The window is wide open", for example, I want to slow the rate of the speech at the moment "wide" is spoken to give the word greater emphasis. Can anyone suggest any methods of how to do this?
I have been experimenting with sending messages to line~ objects, controlling the tempo by telling the line~ object to change the speed of the sample at certain points, however the changes are to obvious and often change the pitch of the speech (which I want to avoid).
Any advice would be fantastic.
Check out this awesome granular time-stretching patch and see if it does what you need.
This is great thanks. Is there any way you can automate things in Max? For example, to automate the tempo over the course of a spoken sentence.
Have a look at [mtr] which is designed for automation. Using [function] in combination with [line] would be another way to go.
I will look into the mtr thanks.
I’ve tried implementing a function/line before to automate the amplitude of my speech sample. When using it with the groove object, is there anyway of getting the automation to loop along with the speech sound? At the moment I have to click the button to get it to automate within the loop.
Also, on another note… Is there away of making the lines between objects/messages to go at right angles (like in the tutorials) instead of straight lines?
Turn on Options > "Segmented Patch Cords" and you’ll be able to route them around and make your patches a bit easier on the eye.
You can use the sync output of [groove~] to send a bang when the loop reaches the end. The method I usually implement is below.
----------begin_max5_patcher---------- 634.3ocwW0sjaBBF8ZySACWayHD7uNsWrOGc1oCQItzQgLJtMc2YyydUH+G2 Dhw3dCF9Tfy24b3Cx6SbfykqXUPv2A+B3379DGGcn1.Na56.KnqRxoU5OCNu VojBnq4UKopjW3hreWxRTloA6EO0yEfh7Zef20BddyXD0EbQNSomOz9fxZ0o QMgT+aIyL2v4TQFb2LwS0PRN+OeCgfsw9Xxj1FWKSF1aozj0WHYhM4BR+f3u ss6bA2Yt3cNXi6CVKXUUzL1VvpXqzXDlKkKAnKjCnYQs3NvjCXi3bsbXgTnD zBMuCepjSyg2lRsWkZmpJ9a5WfPS8i8QQgmSJ3n9vJB1eaF7YjREO6RxJxSa KCBaamMyNK5cSIMfRzLntIlVc4LNoWThRlkkytd1aryC7FTtP049yvATbmWu XAqbMnRVvZ98kpEQBz5LosMNZX0Y7mwAKxkTEz8zpUWWt8Gx5Bkrk4zDFHsr tnsFwTJ2BphnoJDl70VlnS5gLfdnrRo7UlMdHyQ.wl5ldjK6glMTdnM0Jb6S UCbe3oqdtNJh7EbtN1a.EcVZFasEonez9sBiP8Bcx2ixEng7Dye7y0.uoVPN yBN9viGbAha29iFxyZRY4JpMdFD4Q3YFRZI3PZQOPXNWb5s+0oXa7i4pJYcY x1Eei4qIq2s9orJEWPU7lZHG7QM2KErmJdgmlxDGdo3Bd5RYycF1.Bvycpb1 hIbzQKW2XBeaPBQBm56BhL+Q.unsctWnZARiOhguNRMfCg7M6SI65MRXc7D5 GfNemHJzBDEMpHhXAh7G2sm1TwHXTgDxJmj23hIa7Rnw0Lchr7IXJ7NvTSmO l7e.+0OQXB -----------end_max5_patcher-----------
In my "User Page" (see above user pages link), you will find a spectral time stretcher that stretches automatically more the "vowels" than the "consonants". Look at the "play-with-transients" patch. You might be interested.