The No.1 Website for Pro Audio
New Melodyne Still Degrade Audio?
Old 9th February 2013
  #31
Gear Head
 
preachers's Avatar
Quote:
Originally Posted by djanthonyw View Post
I just use the DAW for the timing.
but sometimes manual timing can cause some artifacts by split -> time stretch -> corssfade workflow.
crossfade artifacts when time stretch after middle split within a word can't be avoid completely even when you ensure you are split at zero cross spot. but in melodyne, it won't.
Old 9th February 2013
  #32
Lives for gear
 
Rufuss Sewell's Avatar
If you throw a Melodyned track back through a nice analog EQ and compressor for a bit of harmonic grit, the digitally stuff goes away.

At least on tracks that were record dry, cleanly and edited properly.

A roomy vocal can suffer with Melodyne.
Old 9th February 2013
  #33
Lives for gear
 
djanthonyw's Avatar
 

Quote:
Originally Posted by preachers View Post
but sometimes manual timing can cause some artifacts by split -> time stretch -> corssfade workflow.
crossfade artifacts when time stretch after middle split within a word can't be avoid completely even when you ensure you are split at zero cross spot. but in melodyne, it won't.
Using elastic or flex time is the same exact thing as using time stretching in Melodyne, except Melodyne has more artifacts. Sure, it's more convenient to adjust pitch and timing in one app or plugin such as Melodyne, but you pay for it by loss of quality.
Old 10th February 2013
  #34
Gear Maniac
 
ol drippy's Avatar

Last edited by ol drippy; 10th February 2013 at 08:15 AM.. Reason: double post
Old 10th February 2013
  #35
Gear Maniac
 
ol drippy's Avatar
I still definitely hear it affecting the highs, even just exporting out of pro tools into melodyne editor with no fixes and back into pro tools. It doesn't null but then I also couldn't get Autotune EVO to null either when just tracking the pitch in graph mode with no edits.
Old 10th February 2013
  #36
Quote:
Originally Posted by preachers View Post
you can't compensate those degradations by using a eq
I would argue that, in practice (while perhaps not in theory), you can absolutely compensate with EQ if indeed that works for you. The judgment wether or not you can comes from using your ears.

Talking here from my own perspective, if I can make up for the (very slight) HF loss by using EQ, to the point where I cannot distinguish between the two files in a blind A/B test, then that is as far as I need to go. I trust my ears enough. However, I dont trust my psychology as much, which is why I subject myself to blind tests, to avoid being biased by what I think must be true (this is not saying anyone else does, but just pointing out that its a huge factor when talking about these small differences in sound)

For me, there's no point to worry about a changed texture of the sound, if you cant perceive that change in texture.

Once you actually start tuning in Melodyne, its a different story of course. All tuners have artifacts when tuning. But here is where the workflow of it comes into play (again, for me).

Tuning vocals manually is about musicality, and thoroughly understanding vocals, timing, and pitch.
You are consistently making judgments on:
1. How precisely a note should be hit (perfect pitch is not the goal)
2. How much you should tone down a vibrato thats too strong
3. How much you should correct a note that is drifting in pitch (one of the strongest points in Melodyne in my opinion)
4. Timing, by what exact amount should you change things around (and yes, this could be done in flex time or in the DAW, but since the interplay between pitch and timing is so integral for the feeling, working on them simultaneously is crucial, at least for me)
etc etc

All these are musical considerations, and in order to achieve a good performance, one must have a tool that works for you.

For ME, it happens to be Melodyne Studio, as I need key commands to really have a good workflow. Furthermore, Mel. Studio allows multichannel, if you're working on many vocal tracks.

If a tuned vocal comes out sounding unnatural, then you've done a bad tuning job. You cant really blame the tool, if you've used any of the ones mentioned in this thread.
If we're talking about which one gives you the better result with AUTO-tuning, thats a different story. I personally cant compare AT and MD here, because I never use MD in any auto-fashion. I do however sometimes use AT for AUTO-tuning. Never on voice, but sometimes on instruments. I find it to be better than Logic's pitch correction, but I dont know if thats telling much
Old 26th February 2013
  #37
Lives for gear
 
chet.d's Avatar
 

Considering that I'm one that agrees that once a file is in melodyne it acquires an undesirable deterioration for lack if a better word...

Anyone think there's any "sonic" benefit at all in upgrading to the latest editor 2 ? I now have the previous melo editor ver.
thoughts?
Old 26th February 2013
  #38
Lives for gear
 

This is why I don´t use Melodyne. Bought it many years ago and noticed this degradation. The audio degradation is a deal breaker.

Does anyone know why the Melodyne degrades the audio and why they can´t improve this after all these years?

I mean you don´t have to do anything to the notes. Once the audio is "recorded" into Melodyne, the audio is degraded.
Old 27th February 2013
  #39
Lives for gear
 
djanthonyw's Avatar
 

Quote:
Originally Posted by GearNerd View Post
This is why I don´t use Melodyne. Bought it many years ago and noticed this degradation. The audio degradation is a deal breaker.

Does anyone know why the Melodyne degrades the audio and why they can´t improve this after all these years?

I mean you don´t have to do anything to the notes. Once the audio is "recorded" into Melodyne, the audio is degraded.
I have no idea, but I'm definitely not touching Melodyne again. Sticking with WavesTune. I also love in WavesTune how I don't have to record extra audio that takes up more space on my hard drive. This also makes for easy recalling of sessions if I need to.
Old 28th February 2013
  #40
Lives for gear
 

Can anyone make a comparison between any artifacts in Melodyne vs. Cubase's VariAudio?
Old 28th February 2013
  #41
Lives for gear
 
jrhager84's Avatar
 

I've never had anybody notice a touched vocal part with Melodyne. Ever.

Sent from my SAMSUNG-SGH-I777
Old 28th February 2013
  #42
Gear Addict
 
ontariomaximus's Avatar
 

THAT was interesting.
Old 28th February 2013
  #43
Lives for gear
 
djanthonyw's Avatar
 

Quote:
Originally Posted by jrhager84 View Post
I've never had anybody notice a touched vocal part with Melodyne. Ever.

Sent from my SAMSUNG-SGH-I777
Post a vocal that hasn't been run through melodyne, and a vocal that has, and I will be able to pick out what is what.
Old 28th February 2013
  #44
Quote:
Originally Posted by ontariomaximus View Post
THAT was interesting.
Do you mean the YouTube video? THAT was nonsense. That was probably why comments were disabled for his video. If he doesn't know why distortion happens in software and how not to distort his recordings then he is seriously incompetent. It is not a "flaw" in PT.
Old 28th February 2013
  #45
Quote:
Originally Posted by djanthonyw View Post
Post a vocal that hasn't been run through melodyne, and a vocal that has, and I will be able to pick out what is what.
I will put a file up later.
Old 28th February 2013
  #46
Lives for gear
 
djanthonyw's Avatar
 

Quote:
Originally Posted by The Famous Yard View Post
I will put a file up later.
No post processing on the Melodyne file, please.
Old 28th February 2013
  #47
How are you guys using Melodyne? I may correct like 2-4 notes in a song and after a day or two I forget where they were. Can't hear any difference. Are some people running entire vocal tracks through it front to back?

I did an experiment a while back - copied a track, ran one through Melodyne and kept the original clean - and when played together they sounded a little phase-y, but I chalked that up to melodyne correcting notes and pulling them a few cents here and there off the original - so of course they'd be phase-y. From there I just decided to only use it when I had to and for only a note or line at a time (punching it in and out during breaths or pauses). I never hear any parts now and think, "THAT's where I used it."

Tried Autotune out and felt like it sounded the same. No real discernible difference to me. Now I don't have perfectly golden ears, but I've been recording long enough that I trust em.
Old 28th February 2013
  #48
Quote:
Originally Posted by djanthonyw View Post
Using elastic or flex time is the same exact thing as using time stretching in Melodyne, except Melodyne has more artifacts. Sure, it's more convenient to adjust pitch and timing in one app or plugin such as Melodyne, but you pay for it by loss of quality.
I can tell you regarding flex time, that melodyne is sometimes better and sometimes flex time is better. It depends on the original audio, and I have found that to be the case with ANY digital monkeying around.
Old 28th February 2013
  #49
Quote:
Originally Posted by djanthonyw View Post
No post processing on the Melodyne file, please.
well you should be able to tell because one won't be in tune.
Old 28th February 2013
  #50
Lives for gear
 
djanthonyw's Avatar
 

Quote:
Originally Posted by Musiclab View Post
well you should be able to tell because one won't be in tune.
No, I wasn't even talking about tuning, I was talking about just having audio recorded into Melodyne, without even tuning, introduces artifacts.
Old 28th February 2013
  #51
Lives for gear
 
djanthonyw's Avatar
 

Quote:
Originally Posted by Musiclab View Post
I can tell you regarding flex time, that melodyne is sometimes better and sometimes flex time is better. It depends on the original audio, and I have found that to be the case with ANY digital monkeying around.
From my experience, Melodyne sounds worst than Flex Time. Melodyne's artifacts remind me of badly encoded MP3s.
Old 28th February 2013
  #52
Ok. Time for the challenge. Attached are 6 files which are pairs of 3 small vocal snippets. Each time one of the pair has gone through Melodyne but with no pitch or formant adjustment of any kind. The other is just raw vocal. There are no plugins on either track (other than Melodyne) although a compressor was used to record the original vocal. I am using Melodyne Editor within Pro Tools 9 on a Mac (not sure if any of that matters but just being complete). They are 24/44.1 mono files so no sample rate or bit rate conversion has taken place.

Anyone who wants to have a guess, please elaborate a little about what you feel is different about the two and how close they are. "Degrade" is a bit non-specific, so please put whether you feel the high or low or mids are affected or whether one is less "human" than the other.

I did three short parts to lessen the chances of anyone getting the answers randomly.

If you want any more info then let me know. I think I have covered everything important. I will reveal answers once there enough guesses to make it interesting.

One more thing....Obviously I can't stop you doing this, but PLEASE TRY TO GUESS BY LISTENING ONLY rather than doing a null test to hear the difference as that rather spoils the fun. I will post a difference file of each snippet with the results.

Let's see what happens....
Attached Files

Vox 1a.wav (2.02 MB, 1610 views)

Vox 1b.wav (2.02 MB, 1604 views)

Vox 2a.wav (2.70 MB, 1711 views)

Vox 2b.wav (2.70 MB, 1712 views)

Vox 3a.wav (2.05 MB, 1696 views)

Vox 3b.wav (2.05 MB, 1791 views)

Old 28th February 2013
  #53
Cool idea. Thanks for taking the time to post this.
Old 28th February 2013
  #54
Lives for gear
 
jrhager84's Avatar
 

Quote:
Originally Posted by djanthonyw View Post
No, I wasn't even talking about tuning, I was talking about just having audio recorded into Melodyne, without even tuning, introduces artifacts.
You *do* realize you have to 'split' the S sounds, right?
Old 28th February 2013
  #55
Lives for gear
 
djanthonyw's Avatar
 

Quote:
Originally Posted by jrhager84 View Post
You *do* realize you have to 'split' the S sounds, right?
I never said anything about S sounds.
Old 28th February 2013
  #56
Gear Maniac
 
EricBradley's Avatar
 

Melodyne does degrade audio.
It's really easy to hear.

The biggest problem I have with it is that it destroys high end on untouched notes. If you just import a file into Melodyne everything nulls but when you make a small adjustment everything breaks.

I've provided two spectrograms. The first one is untouched. The second one has a correction on the first note (between 0-1.5 s). You can clearly see how the high end is degrading somewhere over 7 kHz over the whole time period.
Attached Thumbnails
New Melodyne Still Degrade Audio?-not.png   New Melodyne Still Degrade Audio?-melodyned.png  
Attached Files

Bailey NOT_bip.wav (2.02 MB, 2529 views)

Bailey Melodyned_bip.wav (2.02 MB, 2432 views)

Old 28th February 2013
  #57
Gear Maniac
 
EricBradley's Avatar
 

Quote:
Originally Posted by The Famous Yard View Post
Ok. Time for the challenge. Attached are 6 files which are pairs of 3 small vocal snippets. Each time one of the pair has gone through Melodyne but with no pitch or formant adjustment of any kind. The other is just raw vocal. There are no plugins on either track (other than Melodyne) although a compressor was used to record the original vocal. I am using Melodyne Editor within Pro Tools 9 on a Mac (not sure if any of that matters but just being complete). They are 24/44.1 mono files so no sample rate or bit rate conversion has taken place.

Anyone who wants to have a guess, please elaborate a little about what you feel is different about the two and how close they are. "Degrade" is a bit non-specific, so please put whether you feel the high or low or mids are affected or whether one is less "human" than the other.

I did three short parts to lessen the chances of anyone getting the answers randomly.

If you want any more info then let me know. I think I have covered everything important. I will reveal answers once there enough guesses to make it interesting.

One more thing....Obviously I can't stop you doing this, but PLEASE TRY TO GUESS BY LISTENING ONLY rather than doing a null test to hear the difference as that rather spoils the fun. I will post a difference file of each snippet with the results.

Let's see what happens....
Why did you invert phase on every second file?
These files null perfectly just as they are. (down to -132 dBFS)

Now try a pitch-shift on the first note and you will hear the whole take change character.
Old 28th February 2013
  #58
Didn't the dude ask us to just use our ears or something?
Old 28th February 2013
  #59
Quote:
Originally Posted by EricBradley View Post
These files null perfectly just as they are. (down to -132 dBFS)
Well that was a short shootout.

Quote:
Originally Posted by brianellefson View Post
Didn't the dude ask us to just use our ears or something?
The dude did indeed. The reason every other file was polarity reversed is because I was doing my own test to see the difference. When I bounced I just muted the other track.

Yes they do null. So it seems that just running audio into Melodyne does nothing at all to degrade it. I did it the simplest way I could and the two null.

@djanthonyw. I see you took a guess. I appreciate that you had a go as you were the one who asked for a test and took part rather than just sniping from the sidelines, so kudos to you. We have all suffered from expectation bias or placebo effect so I don't blame you for that. Testing at least sorts out the truth.

@EricBradley I wish you'd followed the test but there you go. In the spirit of research I will tune the first word and see how it affects the rest of the file. I'll do it in the morning as it is late here. Re. the spectrograms you posted if you flip between them you can see that the frequency balance has changed on the first section- the tuned bit - but to me it looks the same on the rest of it, but the picture is a little darker. I don't know what significance that has in your software.

Like many of you, I have never had a single complaint about Melodyne from artists. In fact quite the opposite. They are usually really impressed by what it can do when used with discrimination and skillheh. However I have never tested the degradation scientifically, although, as with all pitch correction software it has its limits and artifacts appear depending on material and how much you're changing it, so I'm glad that I now have and also that I have discovered it does nothing to the audio by simply tracking it in.

EDITED for factual incorrectness re another poster.
Old 28th February 2013
  #60
Lives for gear
 
djanthonyw's Avatar
 

My guess for audio ran through Melodyne:

1a
2b
3b
📝 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…
🖨️ Show Printable Version
✉️ Email this Page
🔍 Search thread
♾️ Similar Threads
🎙️ View mentioned gear