The error correction function is horrible

I am new to Wavelab. The error correction function is horrible. No way to identify errors and fix them normally. We go around in circles, we do not know, even after reading the manual, if the error is corrected. We ask to go to the next error and it comes back. We think we have correct, we ask to check and it brings us the error.

But how can such expensive software work so badly?

I use it all the time because I prefer it to plugins like Restore Rig or Izotope. But since the factory preset called “Click Detection” has been added I think it will continue to garner complaints from first time users of the function. That factory preset uses the function called Digital Click Detection which is really only usable on pristine unprocessed low level material. On any normal mixes you’ll get thousands of detections and an error message. With it’s default settings it marks and “repairs” single samples only, which for the most part is pretty useless in most cases imo.

Use one of the other two factory presets. I think Steinberg should dump, modify, or rename the “Click Detection” factory preset, because this will probably come up again because of it.

“Go to the next error and it comes back” should not be happening in the latest version I don’t think.

You can tell if an error is corrected by the green corrected markers around it, if that’s set in the options.

Like bob99 I use it literally every project, but ‘manually’ … on specific clicks etc that I identify by listening. I don’t think I have ever used the auto-style ‘click detection’ for the reasons he outlined and some others. I also prefer it to third party options, but that comes down to personal preference.

Like Bob and Paul I use the error correction function on a regular basis as well. The issue that throws most new users is the error detection part, which you really have to adjust for your own needs. I did on some occasions in the past, but mostly use the error correction function after identifying an audible (and later visible) click or issue. And the options work great there, especially the repair <2 ms function, which I use in maybe 95% of the cases.

Thank you, it should be useful. But it’s disappointing. Software must do what it says it does. In any case thank you for the answer. I had to take Soundforge, which Matrix gave me with Samplitude, for the real verification, then I tested in RX. No mistake. Wavelab gives me 40!

Ok, thanks.

This automatic function is more disturbing than useful, so. I wonder what Steinberg has in mind.

The auto-function … I don’t know. Classical editing maybe?

One of the terrific things about WaveLab is that it is so easy to customize it to fit your particular preferences and workflow. So for me I have only really used error correction manually to fix specific clicks (typically the 3ms pre-set) because that’s the way I tend to work.

I’m not sure what you mean by Wavelab gives you 40 (40 clicks detected?), but I’ll have to try out Soundforge to try and see what the differences are. Wavelab has 3 different click detection methods selectable on the Error Detection & Correction tab, and each of those has adjustable parameters, so results are going to be different depending on the parameter settings and program material. I often use Click Detection 1 and Click Detection 2 with my own settings and get good results. And I often just manually declick on the Error Detect tab too, just selecting around a click and clicking “fix”.

But to me the main problem is still the other detection method, called Digital Click Detection, which is now used (unfortunately imo) in the very first Factory Preset, the Factory Preset called “Click Detection”. All 3 Click Detection methods have been in there for years and haven’t changed afaik, but there were never Factory Presets until recently. I gave up on Digital Click Detection many years ago because I would invariably get thousands of erroneous click detections and the error message, but I never really looked at what it was doing to get there until now. Now I have looked, and find that it’s detecting and correcting single samples only, which I really didn’t expect. And from what I can tell 99% of the detections are false, and 99% of the corrections are wrong.

Error Message:
Error Message.PNG
Detection:
Detection.PNG
Correction:
Correction.PNG

You can even do this on sine waves of 5 KHz and above, and the Factory Preset, run as is, detect all and correct all, will find and “fix” thousands of “clicks” in it. The Factory Preset will deform and audibly distort the sine wave waveform.

I would ask anyone to try the “Click Detection” Factory Preset, running “Detect All” and “Correct All” on a copy of a wav file, and see if you don’t get something similar.

I really think the “Click Detection” Factory Preset should be removed, because it’s possibly someone’s first introduction to the EDC and it’s really not a good start I don’t think.

I’m not suggesting that the Digital Click Detection method should be removed after all this time, because it probably has use for somebody, but it has nothing to do with detecting and correcting normal program clicks. The other two methods oftentimes do very well with that.

Exactly. I’ve never really understood what the digital click detection is meant to be doing in this regard. The factory ‘Click Detection’ preset seems to have little use in my opinion. Perhaps someone could emlighten us as to how/why this preset might be useful.

I must say that I’m also underwhelmed with WaveLab’s spectral repair. Yes, it’s far better than your average DAW but the other day I was working on a project all “in the box” in WaveLab and heard a click.

I decided to see if I could fix it in WaveLab without using RX. I couldn’t get anything to work or make sense and the detection said it didn’t detect the click. So, I just opened the file in RX, did the fix, and replaced the file in my montage.

The reason I don’t use WaveLab spectral editing much is that because I use REAPER for my analog processing (due to WaveLab’s limitations there, I alsoI do my spectral repairs in REAPER with RX as REAPER’s external editor. It’s very fast to highlight a problem section, have a copy of that section open in RX, do the fix, save/overwrite the audio file (no RX file needed), toggle back to REAPER and move on.

It allows me to quickly spot fix any issues, and in REAPER it leaves a visual trail of where all the edits were incase a new mix file is involved.

Plus, I can easily toggle back to the original audio if I do not like a fix that I did and want to start over.

I’d really love to see a way for WaveLab to send files or copies of small sections to an external editing app, and have access to the old and new sections.

As of now, if I’m working exclusively with WaveLab, I usually build the montage, get things dialed in and get a feel for any noise/clicks/pops that are present.

Then I open all the files in RX, make an RX doc and fix all the noise/clicks/pops and other issues, export new files with all the fixes, and then use the File Replace feature in my WaveLab montage to populate the montage with the fixed RX versions.

Again, I think WaveLab spectral repair is of course better than nothing (what you get with Logic, Pro Tools, Cubase, Studio One etc., but I think RX still wins because it’s a specialized tool and it would be great to have better integration between WaveLab and RX. RX Connect is not good either.

I miss the Wavelab 8.5 spectrum editor, (I still go back and use it sometimes). It’s still my favorite: Define as source, define as target, copy source to target. Move backward, move forward. And I could see things better, even after trying all the Wavelab 9.5 view options. Maybe I could get that default ugly 8.5 look in 9.5, but haven’t really gotten it yet. It was ugly but I could more easily see things to remove. Most of the time I remove clicks manually on the waveform view (not spectrum) using “common clicks 3ms fix”, which often works well, but often also need to use spectrum copy/paste, for things I can’t really see very well or fix well in the waveform view. I don’t really like Izotope, and don’t want to use external tools to do this. But that’s just me.

It’s still my favorite: Define as source, define as target, copy source to target.

You can do this in WaveLab 9.5 too: see picture:
2019-04-03_22-40-34.png
Source at left, destination at right. Press Ctrl + V to copy.

I only recently found that … it’s very cool. We all work differently and I still love the way I can switch WL to mid/side display within spectrum editor and error correction and tame pestky stuff transparently and quickly. It’s what you’re used to I guess.

However, I agree with Justin that the external tool function should be improved in Wavelab to facilitate RX because it’s so widely used. If one can make a selection in the session/“montage” (not the audiofile “workspace”) in Reaper and have it sent/returned, then you should be able to do the same thing in Wavelab, without RX Connect. I’ve never had a problem with RX Connect, but it is limited to use on audio files only in Wavelab, and does have it’s detractors. I’ve set up RX as an External Tool in Wavelab, and it doesn’t work very well. It sends ok, bur won’t return. It’s limited to audio files, and it won’t “return” the fixed file. The best I could do was “overwrite original file” in RX, which required closing the file in Wavelab, overwriting, and then reopening in Wavelab, so the function could definitely be improved to at least not have to go through that, and really should be enabled for use in the montage too.

I’m sorry PG, but I really don’t like the Source at Cursor function. I find it clunky and limiting. I really just want to draw a selection around the target, move the selection to a source area, and copy paste to the target, even if I have to move the selection back to the target. I could do that in previous versions, and I think it’s very limiting that it can’t be done anymore. Other programs that have alternate ways of doing it like Source at Cursor, whether RX or Photoshop or anything else that works this way, still allow that basic copy paste function as well. I also liked the Define as Target, Define as Source, and Copy Source to Target buttons in 8.5, and the selection Move Backward and Move Forward buttons.

Yes, and it doesn’t have to be RX exclusive. Some people might want to send audio to Cedar, SoundForge, SpectraLayers, WaveLab 8.5!, or some other audio editors.

The only time I’ve had good luck with RX Connect is in Pro Tools AudioSuite because it’s an offline process. I thought that was the best option until I tried using REAPER’s external audio editor option which skips RX Connect completely and is faster/better.

I’ve tried using RX Connect in WaveLab on an audio file and it’s too clunky. And really, this is best done right from the montage with ability to get back to the original audio if needed via a takes/playlist system.

Here’s a video of how I do it in REAPER:
Dropbox - Error - Simplify your life … n.mov?dl=0

The first time RX opens it is slow because RX had to launch, but after RX it open, it is super fast.

It’s so fast to select a small problem area in REAPER, press ENTER and a copy of that section opens in RX and is automatically focused and ready to go, I make the fix in RX, save/overwrite the file (no RX doc needed), and when I toggle back to REAPER, the fixed section is active and ready to go. I can also easily still access the original audio for that section if I need to go back or compare quickly.

I don’t see it getting any better than this. I can clean up dozens of small clicks or issues in a song VERY fast with this method.

Looks very nice indeed. But what’s with the stereo inside stereo view after RX processing? Is that a Reaper glitch?

I’m not sure what you mean. It could be that you’re seeing both the original stereo take on top, and the copy take that is RX fixed below it. In this case, you just select which take is active for that part of the song so you have easy access to the original, and the fixed version.

The dark one isn’t active, and the lighter one is active.

Ah I see. It’s two 'lanes’in Cubase speak… Cool.