Melodyne mess with your sound? My preliminary conclusions
-
- KVRist
- 382 posts since 6 Apr, 2005 from Fair NJ, the Garden State, US
Ok, so now that I've bought melodyne, I can actually save files and reload them into SX3.
The demo, I thought, sounded horrible. A/B'ing against identical files in cubase -- both apps running at the same time -- I couldn't believe how nasty melodyne sounded.
The thing is, the playback engine in melodyne must be terrible. Some folks on this forum confirmed this idea, and now I've had a chance to prove it to myself.
Took a 24-bit .wav -- about 6 minutes long -- and loaded a copy of it into melodyne. Did a couple manipulations in the first measure, just to make sure that melodyne was fully engaging the audio, and then saved the result.
Loaded the melodyne-treated track and the original track side-by-side in a new SX3 project (excluding the first measure, which I'd messed with). Thought I heard a BARELY perceptible difference, tho I couldn't say one was better than the other. Certainly nothing like a/b comparison when using the melodyne playback engine.
Ran both tracks thru a spectrum analysis thingy, and found that the only meaningful deviation was above about 13K and below about 20hz. The original .wav had a bit more high end than the melodyne -- looks like maybe as much as half a db at 20K. The melodyne track has a little more sub-bass, maybe 1/4 db at 20hz.
So, well, the conclusion is that melodyne does, in itself, alter the audio loaded into it, but not nearly as radically or as detrimentally as the playback engine would have you think.
Without a direct A/B comparison, there'd be no way of telling the track was altered. And, in fact, after significant manipulation on another track in a full mix, the singer didn't notice that pitch correction had been applied.
Last note: I can confirm that melodyne messes up the timing just a little bit. When I played both files together, you could hear the phase.
The demo, I thought, sounded horrible. A/B'ing against identical files in cubase -- both apps running at the same time -- I couldn't believe how nasty melodyne sounded.
The thing is, the playback engine in melodyne must be terrible. Some folks on this forum confirmed this idea, and now I've had a chance to prove it to myself.
Took a 24-bit .wav -- about 6 minutes long -- and loaded a copy of it into melodyne. Did a couple manipulations in the first measure, just to make sure that melodyne was fully engaging the audio, and then saved the result.
Loaded the melodyne-treated track and the original track side-by-side in a new SX3 project (excluding the first measure, which I'd messed with). Thought I heard a BARELY perceptible difference, tho I couldn't say one was better than the other. Certainly nothing like a/b comparison when using the melodyne playback engine.
Ran both tracks thru a spectrum analysis thingy, and found that the only meaningful deviation was above about 13K and below about 20hz. The original .wav had a bit more high end than the melodyne -- looks like maybe as much as half a db at 20K. The melodyne track has a little more sub-bass, maybe 1/4 db at 20hz.
So, well, the conclusion is that melodyne does, in itself, alter the audio loaded into it, but not nearly as radically or as detrimentally as the playback engine would have you think.
Without a direct A/B comparison, there'd be no way of telling the track was altered. And, in fact, after significant manipulation on another track in a full mix, the singer didn't notice that pitch correction had been applied.
Last note: I can confirm that melodyne messes up the timing just a little bit. When I played both files together, you could hear the phase.
Grist for the glamour mill.
-
- KVRist
- Topic Starter
- 382 posts since 6 Apr, 2005 from Fair NJ, the Garden State, US
I have to bump this once. Feel bad that I kind of trashed melodyne based on the demo. And you know they NEED my blessing to have any success in this business . . .
Today I did the A/B routine for an engineer I work with sometimes. He thought it sounded like melodyne's playback engine reduced the bit depth, which might be right. He also agreed that, once you bring the melodyne-processed file back into cubase or whatever, it sounds nearly identical to the original.
He'd never heard of melodyne -- one of those weirdos who spends all his time working and not looking at gear online. He's going to buy the 8-track version. I probably should have as well. It's a first-class bit of software, I think.
Today I did the A/B routine for an engineer I work with sometimes. He thought it sounded like melodyne's playback engine reduced the bit depth, which might be right. He also agreed that, once you bring the melodyne-processed file back into cubase or whatever, it sounds nearly identical to the original.
He'd never heard of melodyne -- one of those weirdos who spends all his time working and not looking at gear online. He's going to buy the 8-track version. I probably should have as well. It's a first-class bit of software, I think.
Grist for the glamour mill.
-
- KVRist
- 378 posts since 11 Mar, 2002 from Victoria BC
That cracked me up!bobby yarrow wrote:
He'd never heard of melodyne -- one of those weirdos who spends all his time working and not looking at gear online.
-
- KVRian
- 1238 posts since 10 May, 2002 from Sweden
I agree. I'm sure it's very useful as a pitch correction tool but I tend to see it as an instrument. Feed it with a good set of sampled (monophonic) phrases and you'll be able to create pretty much any melody line you can think of - and it's going to sound more believable than if you'd played it with a regular multisampled instrument.bobby yarrow wrote: It's a first-class bit of software, I think.
Example:
Original sax phrase
Reconstructed phrase using Melodyne
I really need to get some new material, I think that's the 3rd time I post those.
/Yoss
-
- KVRian
- 1238 posts since 10 May, 2002 from Sweden
No Melodyne examples, but I've got plenty of material (Vocal Planet, Symphony of Voices, DIVA) so I should be able to do quite a lot. Let me see what I can come up with. Like I said, time for some new material.A3ntar wrote:LOL, that was pretty good Yossarian. DO you have any vocal samples?
/Yoss
-
- DC TC
- 2518 posts since 30 May, 2004
with its audio to midi capabilities, would melodyne sound good if i sang the melody of an instrument and converted it to midi to use it on a vsti? i suck at playing instruments but i'm thinking if i can sing it then i'll be all set. in the demo, i wasn't able to try this as there was no midi conversion. oh yeah, and it crashed my computer.
-
- KVRian
- 1238 posts since 10 May, 2002 from Sweden
I haven't really used this feature in Melodyne. My playing also sucks but I'm still not sure that it would be easier to sing a tune than to enter it by playing the keys. Even if the pitch-to-midi part works I have a feeling I'd still be in for a lot of MIDI editing before I got anything useful. Still, I shouldn't knock it until I've tried it.
/Yoss
/Yoss
-
- KVRist
- Topic Starter
- 382 posts since 6 Apr, 2005 from Fair NJ, the Garden State, US
I think the pitch-to-midi thing doesn't come with Uno. There are plenty of pitch-to-midi plugins tho, right?
Anyhow, my yamaha ez-ag compensates nicely for my uselessness as a keyboard player . . .
Anyhow, my yamaha ez-ag compensates nicely for my uselessness as a keyboard player . . .
Grist for the glamour mill.
