Pitch Edit v2

[vncnt]

This morning I tested Melodyne 5 Studio, just before I was going to buy it as a companion to Reason 12.
It has been disappointing experience, and I think I'm going to wait for a version 6 before I reconsider buying.

What caught my eye the most in my first hours as a Melodyne user.
Despite all beautiful features, their interface looks and feels like a clunky translation from Apple to Windows.
From 20 years ago!

Let me explain what we're used to in Reason, and what Reason still can learn from Melodyne.
Not as an attack on Melodyne but for both sides to learn from.
Some issues may be a matter of choice or maybe Apple platform related. I don't know because I've never used Apple products.

I tested with a 3,5 minute song, and also with a worst case scenario.
An untrained horrible voice: mine. To make it less painful for the neighbours it needs: pitch correction, pitch drift correction, pitch modulation correction, formant correction (more on that later), start time correction, note length correction,

My system:
Intel Core i7-3770K CPU @ 3.50GHz, 32 GB ram, 64bit Windows 10, SSD system disk, SSD project disk, GTX690 gpu, external USB audio Focusrite 18i20 (model 1) with asio drivers.

Melodyne:
- Opened the File dialog to use the Search field to jump to my project drive: as soon as I moved the mouse to select from the list it closed the list. All the time.
---> Not a deal breaker but I didn't expect this from a € 699 program that is recommended by professionals.

- Once the 3,5 minute 44,1 kHz 16bit/ch stereo song was loaded and analyzed, each play/pause command had a 0,5 s delay, and the interface became too slow to handle. Apart from a spike on a single core, the cpu was almost running idle. No severe disk activity. Nothing strange.Tried with the onboard audio card: no difference. It seems that something in this program needs optimizing (for Windows maybe?).
---> This is a deal breaker. Info: Reason has no trouble running whatsoever on this configuration, for years.

- Tool Buttons sometimes need to be left alone for at least a second before I can select alternate options underneath that button.

- Zoom: the vertical zoom feels inverted (if you're used to Vegas Pro and many other NLE's).

- The time cursor: many times it doesn't even respond well after clicking. Also, you can only change the current time in a narrow part of your screen.
---> At least Reason responds always to this basic type of command.
---> Both programs could learn from Vegas Pro (and other NLE's): when clicking in the background (not clicking on a note or sound blob): change the current time.

- Tool switching needs to be done with a button, a shortcut key or with the pulldown menu. I couldn't assign tools to a combination of Shift/Ctrl/Alt + LMB drag, or MMB drag, or RMB drag. Even worse: each tool button is reset to the default tool when switching to another tool to the left or right side of that tool button.
---> Reason uses handles to change the 'Drift', 'Transition', and can use the note representation to move the start/end points with or without stretching. I didn't realize how fast and easy you can switch between those tools in Reason until I tried Melodyne 5.
---> Reason also handles Shift/Ctrl/Alt combinations but completely ignores MMB and RMB drag mouse activity. In many programs those mouse buttons are used to pan and zoom. A very nice improvement.
---> Reason makes it easier to edit the biggest problem. In Melodyne I feel forced to work in a more structured way - like a robot. I don't like that. I prefer the fact that Reason is open to my chaotic approach.

- The light-gray interface. Every serious program has a dark mode these days. Why? Because a workday, running multiple screens, can be long. Sometimes I feel sunburned at the end of day.
---> Reason also lacks a dark mode but the light gray of Melodyne is even worse.

- Melodyne can reduce sibilants in amplitude. In Reason you need to unselect those sounds if you don't want to change them together with multiple other notes that do need pitch correcting and/or a transition.
---> Reason could use a practical solution for this. Like a filter, or a lock. Something.

- Tool tips are partly hidden behind the mouse cursor.
---> Reason does that too. Slightly better maybe.

- Shotcut keys are not visible in tool tips.
---> Reason does.

- The Melodyne algorithms seem to detect every weirdness in my voice. I haven't found a way to combine multiple sound blobs successfully. Not visually. Not audible. Sometimes sounds as an echo, sometimes as separated sounds.
---> Reason treats this kind of weirdness as a single note. Is the Melodyne algorithm overly complicated?
---> The Reason algorithm sometimes seems to add a kind of rumble or noise to the sound when transposing. This makes the sound feel more processed. When I enhance the high frequencies with an EQ that side effect seems to be masked.
---> Melodyne seems to transpose pro voices cleaner but struggles a lot more than Reason when dealing with horrible voices (or takes).

- I needed to clone sound blobs in Melodyne by Copy/Paste. Ctrl+drag would have been logical (and much faster) in my opinion.
---> I understand that Reason doesn't do polyphonic notes (at all) but once a note has been pitch detected it could open the door to experimental use of these sounds when Reason allowed them to be combined to polyphonic results, as if you've used MIDI input to synthesize voices, like the Neptune.

- I liked the Pitch editing in Melodyne: it can even invert the pitch envelope of the detected note.
Its Pitch Drift tool changes the balance between start pitch and end pitch (like rotating the pitch envelope): nice!
Melodyne can edit the fade-in and fade-out per blob: very nice, especially when you need to dial down (or shorten) Dutch/Flemish 'g' sounds to a less frightening level.
---> Reason could learn from this!

- With my untrained voice, sometimes there are differences in overtones, the high frequencies are reduced with higher notes.
Interestingly enough, when using the (fantastic 'in-blob' edit method, opposed to Reasons annoying separate dial) Melodynes Formant Tool compensates this kind of unprofessional sound blob sequences while Reason changes the formant in a more traditional way (as in: not removing the problem). I still can't explain the difference very well. Seems like Melodyne restores the high overtones while Reason feels like it is changing a filter with a high Q-factor.

I haven't touched DNA chord editing or Harmonics editing. I will in the next 29 days of my trial.
This trial has made me doubt whether I want to buy version 5 in the short term.
I think the slow performance, when using Melodyne on a pc with a cpu and memory that is barely touched, is suspicious.

Also, my opinion about Pitch Edit in Reason is now a lot more positive, even though it does not handle everything.
Garbage in, garbage out defenitely applies here. But it's not just about singing, it's also about voice acting and special voice effects.

This is it for now.
I hope this can slightly help to improve both programs.

[mcatalao]

Hi. Interesting take on some things.

I use Reason's pitch editor A LOT for singers, and I also have Melodyne Editor 5 in my set of tools.

To my ear, reason studio has done a great job and for most of my work I end up using reason's pitch editor. It is really a great tool, so well integrated within reason's sequencer, and the integrated workflow is the main reason melodyne collects dust on most of my projects.

I agree with a lot of the stuff you say, but let me complement with some things i feel about reason's pitch editor.

First, the way reason "senses" tone is not so good as Melodyne's pitch detection. Melodyne can detect any tone at any octave. I've seen reason fail to understand the tone of very low vocals, bass instruments, and odd instruments like bagpipes. Also reason struggles with ambient noise, that I often have to take care before opening pitch editor. So there's a point where Reason pitch editor still has to evolve.

Second, Reason's pitch edit algorithm is a lot grainier than melodyne's specially with vocals with higher frequency content. Female voices seem to lose a little bit of definition, not only in the tone (sometimes it sounds a tad muffled) as also there's the loss of some transients in the way. This is not negligible, but it's workable, specially if you use a more surgical approach instead of the general "tweak everything to the medium tone and remove drift". But then again... When I'm working on client's stuff that need A LOT of pitch editing I'll probably create a melodyne project, take care of pitch and then import to reason. This is another point reason would have to evolve a bit too, because melodyne can be more transparent.

Third, I really like the scale tools in melodyne. For composing, sending voicing ideas to singers, and even for some choirs it's great, and SO FAST to create vocal harmonies. So I really wished reason to implement this, because the integration with the sequencer is already so good, that having the scale/mode glue would be a great workflow improvement.

I understand melodyne is ugly as heck, at least in windows. But the functional feats of that software outperform the lack lust of the GUI. DNA has saved my butt more than once or twice, specially on stuff that I can't have the artist come in and redo some stuff! I've successfully tuned guitars, pianos, accordions and 3 voice a-capellas. I even used it for separating stuff and remixing after losing a project. It was not perfect, but the client didn't complain.

Together with Izotope RX these two pieces of software have saved me hours of cutting and pasting and redoing stuff and bringing musicians and vocalists into the studio, which either would bring the project out of budget, impractical or taking the money/time out of my pocket!