current state of pitch-shifting/transposing technology in DAWs

[Bob_Ross]

Well, I think I've finally come up against the limitations of the two decade-old technology I'm using in my recording studio!

I got a simple editing gig: Client wants to splice two versions of the same song together. Tempo is close enough that it won't be a problem...and in fact the client really loves the slightly faster tempo of the version that's coming second in this mash-up since the whole point is to go from a quiet acoustic version to a hyped-up punk rock version of the same tune.

Problem is, the quiet acoustic version is in the key of Bb, and the hyped-up punk rock version is in D.

My DAW is a ~20 year old version of MOTU Digital Performer, and the native pitch-shifting/spectral manipulation technology is simpy not up to the task of transposing an entire full band + vocals mix up or down multiple half-steps.

So, it's 2024 now. Is anybody's pitch-shifting/spectral manipulation technology up to the task of transposing an entire full band + vocals mix up or down multiple half-steps? Worst case, one tune needs to go down or up a major third; best case, one tune goes up a major second and the other goes down a major second.

(Note, I'm probably not going to buy any new software or hardware to do this myself, even if it is a simple task nowadays; I'll just hire an engineer with contemporary tools, send the raw files out and have them transposed before I do the editing.)

 

[ugly_bassplayer]

I whole band major 3rd up,
Ouch, good luck splicing that one together.

 

[scott sinner]

I don't think you’re going to get decent results with that. That is just too big a shift to sound natural. Timbres change significantly at that stretch.

But doesnt hurt to try. Reaper’s plugins are very good for this, for free. REAPER | ReaPlugs
Use ReaPitch and try both format settings and the different algorithms.

Do you have Melodyne? You said you werent going to buy any, but melodyne can do a pretty good job too.

 

[silky smoove]

Melodyne is what springs to mind for me. It won’t sound great doing that kind of transposition, but it’ll probably sound better than just about anything else. Another option would be to use Izotope RX to extract the individual elements of the mix and transpose them all separately but leave non-pitch elements like drums out from transposition.

 

[Aloe]

in what quality do you have the recording to be transposed?
if 24/96 or greater, give it a try. if what you have is 16/44.1, you might be disappointed with the result.
(in fact, it's the second number that matters)

 

[silky smoove]

if what you have is 16/44.1, you might be disappointed with the result.

In what way would sample rate matter for this kind of operation? There’s no harmonic generation happening that would attempt to create frequencies above Nyquist and cause aliasing. The transposition is not so severe that it’s likely to exceed Nyquist, and any content that did get shifted that high could be addressed by a simple low pass filter. Maybe I’m missing something, but I see no relevance to sample rates for this.

 

[Carl Hillman]

Well, I think I've finally come up against the limitations of the two decade-old technology I'm using in my recording studio!

I got a simple editing gig: Client wants to splice two versions of the same song together. Tempo is close enough that it won't be a problem...and in fact the client really loves the slightly faster tempo of the version that's coming second in this mash-up since the whole point is to go from a quiet acoustic version to a hyped-up punk rock version of the same tune.

Problem is, the quiet acoustic version is in the key of Bb, and the hyped-up punk rock version is in D.

My DAW is a ~20 year old version of MOTU Digital Performer, and the native pitch-shifting/spectral manipulation technology is simpy not up to the task of transposing an entire full band + vocals mix up or down multiple half-steps.

So, it's 2024 now. Is anybody's pitch-shifting/spectral manipulation technology up to the task of transposing an entire full band + vocals mix up or down multiple half-steps? Worst case, one tune needs to go down or up a major third; best case, one tune goes up a major second and the other goes down a major second.

(Note, I'm probably not going to buy any new software or hardware to do this myself, even if it is a simple task nowadays; I'll just hire an engineer with contemporary tools, send the raw files out and have them transposed before I do the editing.)

Shift both halves of the song to concert C. One up a step, other down a step…

 

[Engle]

Shift both halves of the song to concert C. One up a step, other down a step…

This is a good idea. I pitch shift things in Reaper often for practice tracks. It works but more than 2 or 3 steps starts sounding unnatural.

 

[Bob_Ross]

Shift both halves of the song to concert C. One up a step, other down a step…

You mean what I suggested in OP?

Worst case, one tune needs to go down or up a major third; best case, one tune goes up a major second and the other goes down a major second.

Yeah, I've already tried that. My 20-year-old technology is still too primative to make that listenable.

The good news is, I sent a rough example of no pitch shift, just a weird cross-fade from the Bb version to the D version, and the client just got back to me and said "That's fine, no one will notice." (!!!)

So, hooray for tin-eared clients, I guess.

Still, I would feel better if I could make the edit less awkward sounding, so still gonna pursue pitch shifting if possible/practical. Thanks for all the suggestions so far.

 

[Plain Old Barry]

Audacity is surprising good for free. I use it all the time to good result to pitch shift cover tunes I need to learn.

Try it. It's free.

 

[pfschim]

Just got out of several days in a local pro studio. Got to talking to the engineer about this very thing and he said that 1/2 step or 1 whole step is do-able, but after that you end up with some pretty noticeable timbre issues. And this is in a pro studio setting with a very solid engineer who is well known for being able to "fix" a lot of performance issues. These latest sessions had a lot of non-fixed pitch stuff happening (flute, piccolo, violin and a few somewhat pitchy voices) and he did some pretty clever stuff throughout the tracks. I believe he's using ProTools FWIW.

that one whole step down, one whole step up approach that Carl Hillman suggested sounds promising

good luck OP!

 

[logdrum]

Antelope Audio interfaces have a plugin that works really well using hardware based (DSP + FPGA ) processing. It is called auto-tune synergy but you have to have their interface for it to work. It's uncanny. You can play an out of tune guitar and it will correct it, during tracking. That's how good it is. I do not know if it will pitch shift and entire recording. It might.

 

[themickster]

Well, I think I've finally come up against the limitations of the two decade-old technology I'm using in my recording studio!

I got a simple editing gig: Client wants to splice two versions of the same song together. Tempo is close enough that it won't be a problem...and in fact the client really loves the slightly faster tempo of the version that's coming second in this mash-up since the whole point is to go from a quiet acoustic version to a hyped-up punk rock version of the same tune.

Problem is, the quiet acoustic version is in the key of Bb, and the hyped-up punk rock version is in D.

My DAW is a ~20 year old version of MOTU Digital Performer, and the native pitch-shifting/spectral manipulation technology is simpy not up to the task of transposing an entire full band + vocals mix up or down multiple half-steps.

So, it's 2024 now. Is anybody's pitch-shifting/spectral manipulation technology up to the task of transposing an entire full band + vocals mix up or down multiple half-steps? Worst case, one tune needs to go down or up a major third; best case, one tune goes up a major second and the other goes down a major second.

(Note, I'm probably not going to buy any new software or hardware to do this myself, even if it is a simple task nowadays; I'll just hire an engineer with contemporary tools, send the raw files out and have them transposed before I do the editing.)

Pitchshifting works well, but not on overdriven guitar nor singing. I always record guitar dry and give it software overdrive.

 

[lark_z]

+1 on Audacity. The furthest I've pitch shifted was dropping "Ramblin' Man" from G# down to E (is that right? first cup of coffee here) for my guitarist/singer buddy. It worked fine. And as @Plain Old Barry indicates, you can't beat the price.

 

[JRA]

...one tune goes up a major second and the other goes down a major second.

that's the best approach, IMO, and doable with most modern DAWs.

 

[dbsfgyd1]

Well, I think I've finally come up against the limitations of the two decade-old technology I'm using in my recording studio!

I got a simple editing gig: Client wants to splice two versions of the same song together. Tempo is close enough that it won't be a problem...and in fact the client really loves the slightly faster tempo of the version that's coming second in this mash-up since the whole point is to go from a quiet acoustic version to a hyped-up punk rock version of the same tune.

Problem is, the quiet acoustic version is in the key of Bb, and the hyped-up punk rock version is in D.

My DAW is a ~20 year old version of MOTU Digital Performer, and the native pitch-shifting/spectral manipulation technology is simpy not up to the task of transposing an entire full band + vocals mix up or down multiple half-steps.

So, it's 2024 now. Is anybody's pitch-shifting/spectral manipulation technology up to the task of transposing an entire full band + vocals mix up or down multiple half-steps? Worst case, one tune needs to go down or up a major third; best case, one tune goes up a major second and the other goes down a major second.

(Note, I'm probably not going to buy any new software or hardware to do this myself, even if it is a simple task nowadays; I'll just hire an engineer with contemporary tools, send the raw files out and have them transposed before I do the editing.)

Best wishes. I have a feeling that in the time you’re going spend to put those two versions together, you could have the band back in the studio to recut the project. JMO, but hope you find a solution to your dilemma.

 

[brianrost]

Celemony Melodyne is the only software that can handle polyphonic pitch correction, even mixes. It works really well, but like all pitch shifting it starts sounding weird after transposing about a third up or down. You might be able to shift Bb to D or better yet, shift them both to C.

To get the best results takes time to learn the software and then time to execute (imagine trying to tweak every note in a mix). It works wonders on things like out of tune guitars and pianos or shaky intonation on fretless instruments.

The two lowest priced versions do not handle polyphony. You need either Editor ($400) or Studio ($700). So if the client is willing to pay you a solid chunk of that and give you a month to work on it…maybe

Sub the work out.

 

[JazznFunk]

At some point it's probably easier and more cost-effective to re-record one of the versions instead of spending hours/days/weeks editing it so that it kinda sorta maybe sounds sorta not bad. My general rule of thumb in post-pro with pitch shifting is that unless it's a single instrument, where I'm fixing very minor variances in order to tighten things up, I'll re-record a part before I get wild with pitch shifting. At some point, it begins to sound unnatural, even with the best software.

 

[MuseChaser]

in what quality do you have the recording to be transposed?
if 24/96 or greater, give it a try. if what you have is 16/44.1, you might be disappointed with the result.
(in fact, it's the second number that matters)

After many, MANY years of working with DAWs since the early days, the second number matters MUCH less than the first number. 24 bit yields a much lower noise floor and greater dynamic range (especially imoortant on classical and acoustic jazz recordings). I use 48kHz/24 these days for almost everything. There really isn't anything gained by using higher sampling rates unless you're planning doing a LOT of processing prior to the final mastering.

 

[Avigdor]

Maybe consider adding a transition with just some rising chords (can be newly recorded) just to go to the new pitch. Power chords in Bb, B, C, C# and D with increasingly heavy guitar and some drum hits or a synth pad with a smoothly rising pitch might just give you a transition between the parts. The band may well come up with a creatively much better idea on how to do it. And some gradual speedup to match the tempo can also be included. Hard to tell without hearing the material if this is even remotely possible at the desired transition point, but maybe an idea.