Hi,
I am not to good with Audio, working mostly in MIDI. I have a track which is recorded poorly and needs to be fixed. What is the best way, using Cubase 15, Spectralayers Go or Wavelab Goot some other stock plugin?
Frankly, I am pecking around in the Dark.
thank you
Z
here is a pick of the horror track:
spectralayers have de-clip, you could try to soften the clipping with envelope shaper or try to mask the clipping with tape or saturation plugins. I guess many people go to iZotope RX for this too.
For this particular case: the event’s gain is set to +24dB. If you put this to 0dB does it still look bad? Does it actually sound bad?
Just because the waveform looks the way it does doesn’t automatically mean the audio is clipping.
Hi Johnny and thank you. What is the difference between the volume control in the Inspector and the one in the event (that you saw at 24db)? Is it even a volume control? It’s made the drack prettier to the eye.
Also why does one of my two audio tracks have this control and the other not?
Z
I think I have it. It’s a gain control not a volume control. I have set up gain for only one track, so this is why I see only one control revealed?
(Edit: I wrote this post before anyone else had replied to this thread)
Whether you can recover this file to a no-clipping version will, I think, depend on two related factors: the bit depth of the audio interface you used, and the bit depth you recorded at in software. I’m not an expert on plugins, if there’s a plugin that can remove clipping from an audio file I’d be interested to hear about it.
About bit depth: recording at 32 bit float removes the need for gain setting when you record, provided the audio interface and software are both using 32 bit float. With this setup, it doesn’t matter how badly clipped an audio recording appears to be, you can simply renormalise the file to peak at some minus value, say -0.1 dBs, and the “clipping” disappears completely. Search and read about 32 bit float for how and why.
I’ve used Wavelab Pro 11 with a Zoom UAC-232 in precisely this way, it works great for squeezing the maximum possible signal to noise ratio out of a recording, because you don’t have to adjust the gain on the interface to allow headroom in case there’s a sudden peak.
However, whilst Wavelab Pro 11 and Cubase Pro 12 both offer an option to record at 32 bit float (I mention these versions only because they are the ones I have), I don’t think the Steinberg audio interfaces offer 32 bit float. My UR24C doesn’t.
If when you recorded this file the audio interface was set to say 24 bit or 16 bit, or the recording software was set to 24 bit or 16 bit, then my understanding is that the clipping is baked into the recording, I don’t know of a way of removing it.
BTW I did not record this file. I am running Cubase 15 in 24 bit.
As Johnny_Moneto spotted, the gain is set in the event to +24 dBs. So maybe the recording isn’t clipped, it just needs the gain reducing to zero or a minus number?
Every audio event has its own gain setting. It is just a volume level modifier. If it is set to 0.0dB (= no change) then it is not displayed in the lower left corner. You can still see it in the Info Line when the event is selected.
In the Inspector we make settings for the track itself. Changing the volume level here will affect any audio event on the track evenly.
BTW - you can easily change the event gain by pressing Ctrl + Shift and the left mouse button and then moving the mouse up and down.
My humble view is:
There’s no clipping inside Cubase, as it works at 32bit float. Clipping occurs at the AD- and DA-converters, only. Too strong signal into Cubase will make the converter unable to write the signal to disc, too strong signal when sending it out of Cubase, Render/Mixdown, wil make the file contain maximum level over a periode of time and then the speakers membrane will stay max and not vibrating in giving sound.
What you say + when rendering a file. If an integer bit resolution is used there, e.g. 24bit, then clipping will occur if the level exceeds 0.0dBFS. This affects Bounce Selection, Render in Place, Audio Mixdown, and Quick Export.
Personally I create all audio files in a floating point format except for the final file, that is the one people should only listen to.
A lot of this is above my pay grade. An anyone suggest something simple? Just an app maybe?
What do you have the event gain set to now, and is it still clipping? It was still at +8.61dB in the last screen grab you shared.
Learn that stuff. You use a DAW, learn your tools like you have to learn how to play an instrument.
Hi @ZeroZero ,
why don’t you just follow the advice given here and turn the gain down to 0dB for starters. Afterwards, press F7 (Direct Offline Processing) and normalize the event to 0dB. That’s what you got and you can take it from there. If it sounds bad then it’s bad - you can only do so much to turn a poorly recorded track into a pristine one.
Rest assured, if you managed MIDI stuff than audio tracks are rather simple to understand. Nothing to be afraid of - you will get there in no time 
The only thing you must/should know is that clipping occurs only at input (AD) and output (DA) to Cubase, not inside the DAW. And the Wave-image is just an image and is not telling anything about clipping.
Take care when recording audio not to drive the input gain to much, and take care when mixing down not to get into the red at output. How to achieve this is a matter of knowing your gear and DAW.
Mixing at low level and enough headroom makes it easier to avoid clipping at mixdown.
Thank you all. I have got it going now.
Z
