Pitch Shift to Achieve Key Change

[Scott Kendrick]

... we messed up... we didn't recognize that the tune we were covering was not the ideal key for our singer. Is Melodyne the best there is for pitch shifting instruments - specifically acoustic guitar, bass, electric guitar?  I tried it, but the acoustic doesn't sound great after doing so. (tried most/all the algorithms)... was hoping not to have to re-record.

thanks everyone

[John Bradley]

In my experience, you can pitch shift +/- about 2 half steps. Beyond that vocals and acoustic instruments start getting noticably wonky.

If you record your electric guitars DI and do the amp modelling in-the-box, you can go a little further with pitch shifting, as feeding the shifted (DI) through an amp sim will mask quite a bit the added artifacts. Obviously you can get away with more shifting if the amp is heavily distorting the signal. Not so much if you're feeding it into a squeaky clean Fender Twin or Roland JC-120 model.

[Scott Kendrick]

yea - we were planning on just 2 half steps, and I think all the electrics are DI - the  acoustic is the problem as it's pretty prominent. When I tried it with Melodyne it sounded like a bad recording... it's possible I"m doing something wrong.

[John Bradley]

I’ve only ever tried pitch shifting monophonic signals. Shifting polyphonic sounds (I assume the acoustic is strumming some chords, as they so often do), I’m not sure if you can do that at all and not have it sound weird, even if you have Studio with its polyphonic support. Maybe on a nice clean piano, but guitars are jangly creatures full of weird overtones and tuning compromises. 
 

Good luck!

[bitflipper]

Don't try pitch-shifting stereo tracks. Don't try it on guitars that are heavily effected. It won't work well on anything that has a lot of unpitched content, such as tuned percussion. Clean mono tracks only.

Make sure you're not over-thinking it. If the vocalist struggles to hit a few high notes, for instance, it would be a lot easier to pitch-shift just those problematic notes on the vocal.

[John Vere]

It’s been a while but I lowered the pitch of a song by a whole step ( E down to D). and I’m pretty sure I just used the same transpose I use for Midi short cut N. it was 2 Acoustic guitars and an electric lead. Worked just fine. But as Bitflipper has said the tracks need to be clean. 
It one reason I try and start with as much midi as possible until I do the vocals. I always record the final guitar parts after the vocals so I don’t walk all over them. 

[rfssongs]

Pitch change is really iffy, best done in a very limited way. Sometime I copy a chorus part and create melodyne harmonies. The thing is though, if you use midi to test your song you can make any key change you want. Then if you are a live sounds kind of guy you have something solid you can start adding tracks to in the key & with the tempo you want.

[Scott Kendrick]

This is all super helpful.. I did go back and try doing a full track import into melodyne standalone (vs. as a plugin shifting in real time in Cakewalk). It worked pretty well on the acoustic using the unversal algo - the polyphonic had more weird artifacts in the result and I don't think it's necessary as I"m shifting the whole thing. I'm doing it on the original raw files - there is one electric track that has a bit of delay, but it's a pretty simple track of chords, so the transition was pretty clean.

Re: the overthinking (which is good advice)... in this case, our vocalist had to sing most of the song a whole octave up from the original to be in her range (female vocalist vs. male)... so a good part of the song was super high, and then she dropped down when the original got more into her comfortable/power range. She pulled it off, but the shifting made for some unexpected transitions. Raising the key hopefully allows her to sing in the original arrangement... Feeling more positive about this now. Frankly I'd rather re-record it, but it's a lot of work and I'm the drummer/engineer, ha ha, so I totally understand the reluctance :).