The No.1 Website for Pro Audio
New Melodyne Still Degrade Audio?
Old 8th July 2013
  #91
Quote:
Originally Posted by The Famous Yard View Post
Did you not approve of the test earlier in the thread which showed no change at all to unprocessed audio? The files nulled when one was polarity inverted. They are identical. There is no degradation.
I must also point out that you have no way of knowing if I would get the same results doing a null test with my DAW - things get strange sometimes in DAWland (esp with third party plugs), and what is true on your computer may not be on mine....

You may not be hearing what I am hearing - but I sure am
Old 8th July 2013
  #92
Quote:
Originally Posted by The Famous Yard View Post
But it doesn't. Audio in Melodyne is identical until you process something (change pitch or formant etc.). When you change the pitch then of course things will change, although this is processing rather than "degradation". How good it sounds after processing is a matter of taste and opinion and depends very much on the source. Before processing it is not a matter of opinion because it can be demonstrated to be identical to the source.
Processing is degradation, you can't get that information back again afterwards, especially with something like Melodyne, it's not a mere sample based convolution but a function that reconstructs the whole waveform.

Quote:
Why will it be less noticeable at higher sample rates? Higher sample rates just add the very high frequency content. The difference between recording at 48kHz and 96kHz is you are adding the audio frequencies from about 24-48kHz. That is all. It's not "higher quality", it just records a higher set of frequencies. If there are artefacts lower than 24kHz then you will get them at both sampling frequencies just the same. If there are artefacts above 24kHz then you will hear them MORE at 96 than at 48 (which won't record them at all). However in real life you won't hear anything at 24 kHz or above anyway because it's beyond human hearing.
Because Melodyne DNA works by applying a wavelet fourier transform to bring the data into the frequency domain, adjusts the values and then transforms it back into the time domain. The higher the sample rate the greater the accuracy of the initial transform can be, and while 44.1 may be high enough to reconstruct a static oscillation at < 22.05 with short hamming windows you're looking at a substantially narrower frequency domain, higher sample rate = wider hamming window for the same length of time = more accurate frequency match. The frequencies required to reconstruct certain sharper waveforms are beyond human hearing even if the end resulting waveform is well within, essentially you hit aliasing artifacts from a whole other direction than the usual compression.
Old 9th July 2013
  #93
Quote:
Originally Posted by Hardtoe View Post
I must also point out that you have no way of knowing if I would get the same results doing a null test with my DAW - things get strange sometimes in DAWland (esp with third party plugs), and what is true on your computer may not be on mine....

You may not be hearing what I am hearing - but I sure am
I understand DAWs work a bit differently. Could you post a before and after file so I can hear what you mean?
Old 9th July 2013
  #94
I have updated the plugin a few months back - there is no "before" any longer.

But I am not trying to prove anything to you - just sharing my experiences with Melodyne for the sake of the thread....

I already settled the matter for myself when I heard the improvement - case closed, on to the next issue which needs my attention - sorry I can't help you out.
Old 9th July 2013
  #95
Well I just checked, and the famous yard is right in that it looks like melodyne retains the sample information before and after the section it adjusts, so it should remain identical although this is by no means an exhaustive test, but the section it changes is quite telling.



This image shows a quick square wave output, the top most is the midi note then under that the three audio sample outputs, the top (cream colored) is the original sample data, the next one down (light blue) is the same data put into melodyne and then exported without any adjustment,then the bottom one (brown) is the same note changed pitch in melodyne, notice the change/breakup in waveform shape and how much longer the waveform goes on for.

The square wave is pretty much a worst case scenario anyway, but it's interesting to see the effect of the algorithm, it's also worth noting that the whole note was offset by the way which may have an additional subtle impact on the feel of a performance that's adjusted.
Old 9th July 2013
  #96
Quote:
Originally Posted by Hardtoe View Post
I have updated the plugin a few months back - there is no "before" any longer.

But I am not trying to prove anything to you - just sharing my experiences with Melodyne for the sake of the thread....

I already settled the matter for myself when I heard the improvement - case closed, on to the next issue which needs my attention - sorry I can't help you out.
I didn't mean before and after the update, I meant audio before its been in Melodyne and after it's been in Melodyne but unprocessed. I'd be interested in hearing what's going on that's bothering you, as for me nothing is changing. If Melodyne has some issues when used in a certain DAW then it's worth knowing about. But if you don't want to then that's cool. I'm curious, that's all.
Old 9th July 2013
  #97
Quote:
Originally Posted by The Famous Yard View Post
I didn't mean before and after the update, I meant audio before its been in Melodyne and after it's been in Melodyne but unprocessed.

I dont hear the problem any more - I now agree with your previous test - there is nothing to learn now as the problem is gone - the only difference I hear is after I have done some editing to the captured audio in Melodyne.

that's the whole point I've been making here - bad before, good now - the update fixed whatever I was hearing

(which is what I said in the first place in post #81)

Old 9th July 2013
  #98
Quote:
Originally Posted by Hardtoe View Post
I dont hear the problem any more - I now agree with your previous test - there is nothing to learn now as the problem is gone - the only difference I hear is after I have done some editing to the captured audio in Melodyne.

that's the whole point I've been making here - bad before, good now - the update fixed whatever I was hearing

(which is what I said in the first place in post #81)

I see what you meant now. I never tested the older versions to be honest, so at least this thread prompted me to check the current version.
Old 10th July 2013
  #99
I don't mean to derail the discussion here, but does anyone know of any buzz on the next edition of Melodyne? Its been a while on the current version, which I skipped, so the next one is the one I'm waiting for.
Old 10th July 2013
  #100
Haven't heard a thing. But based on how long it took for Melodyne DNA to come to market after their first announcement then I suspect two things

a. Melodyne has a really long development cycle
b. They got burnt by announcing DNA so early and they're likely to keep real quiet in future.

If they're like most other software companies then I imagine they're got a tick/tock cycle and as the one before last was DNA, then the last was more minor then the next release might be substantial and could take longer than usual.
Old 10th July 2013
  #101
Quote:
Originally Posted by Drywsef View Post
I don't mean to derail the discussion here, but does anyone know of any buzz on the next edition of Melodyne? Its been a while on the current version, which I skipped, so the next one is the one I'm waiting for.
What are you hoping for in the next version beyond DNA?
Old 10th July 2013
  #102
Quote:
Originally Posted by The Famous Yard View Post
What are you hoping for in the next version beyond DNA?
A Vocalign equivalent, an algorithm that automatically ignores sibilance in pitch editing, tonal improvement in formant editing (which to me is unusable with the degradation it currently imparts), the ability to be rendered offline (like audiosuite), and general tonal improvement.
According to what I'm reading here it seems the the latest version did make a general tonal improvement.
Old 10th July 2013
  #103
I've also noticed that DNA removes a lot of attack from instruments such as guitar. I'll use it on the DI signal in a real pinch, but it is currently borderline unusable for my applications.
Old 24th March 2015
  #104
Lives for gear
 
jlaws's Avatar
Does anyone here use revoice pro 3 to do their pitch corrections?
Old 24th March 2015
  #105
Lives for gear
 

Quote:
Originally Posted by jlaws View Post
Does anyone here use revoice pro 3 to do their pitch corrections?
Sorry I don't have an answer for you, I have RVP3 myself, but haven't gotten around to testing the pitch correction part yet, so I'm interested in what people who do use it for pitch correction have to say ...

Thanks -

Last edited by alexis; 25th March 2015 at 09:30 PM..
πŸ“ Reply
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
πŸ–¨οΈ Show Printable Version
βœ‰οΈ Email this Page
πŸ” Search thread
♾️ Similar Threads
πŸŽ™οΈ View mentioned gear