Vocals to hot. How?

I don’t think that processing the audio with the Sphere plugin can fix those limited/clipped areas…
There are tools though that claim to be able to repair the effects of clipping, like iZotope RX declip or Acon Acoustica DeClip, those try to reconstruct the original, non-clipped waveform. I don’t know how good the results are, but it might be worth to give those a shot (both have trial versions available).

What PreSonus interface are you using? It seems like some of them do have some sort of onboard processing (“FatChannel”), so the limiting might have happened there.

Else, it would probably a good idea to make some test recordings on a completely new project and see if you can reproduce the clipping and see where it happens, so it won’t be a problem next time.
I do think though that the original recording was too hot. If the highest sample value in the recorded file is ~-5dB, and that is already with the limiting, it is likely that without the limiting, the peaks would even be a few dB higher. That is a bit too close to 0dbFS imho, I’d go for max peak levels at around -12 to -10dbFS to be on the safe side.

2 Likes

Using plug-ins for de-clipping might salvage something, it’s worth a try.

The surprising thing is how we have a limited waveform coming into Cubase at -4 dBs or so, without an obvious source of limiting, and the preamps in the green.

For me, it’s one of the following:

  • Undetected compression on the interface.
  • Undetected compression on the input channel. (Maybe in channel strip?)
  • Capsule crapping out due to high SPL. In that case, congratulations to the singer! :partying_face: Flip the -20dB switch on the mic!

:rofl: Wouldn’t want to be in the same room with the singer then…

You write that you had the meters in view during recording. Also the waveform?
But I suppose you didn’t notice the problem during the recording. I mean, if it had already happened during recording, you would have heard and seen it at that point. Unless, of course, the recording was running unattended. (For example, because you were playing an instrument during the recording.)

Are there other recordings affected? Or just this one?

In any case, this is worth a try. :+1:

We create new surround or atmos mixes for old films and usually deal with material that has already passed through many hands. Unfortunately, clipping is not uncommon with these recordings and mixes.
@dg2468 If you want to try to “save” your recordings, I’m happy to help you with that. As long as it does not degenerate into too much work, of course. :smile:

Limiting doesn’t cause digital clipping. Conversely, the whole purpose of limiting is to prevent clipping from occurring during the recording process. As such, since you claim to hear distortion in the ‘clipped’ areas, that would never be caused by a limiter. So something else is going on here. The mic system you have appears to be too complicated for someone with little experience with microphones. In my experience, mics and devices which claim to emulate other mics are highly overrated and completely unnecessary. I would suggest simplifying your setup and purchasing a decent normal vocal microphone. There are plenty of really good ones which are quite inexpensive that do the job extremely well. In addition, a vocalist is always a monophonic source and as such should always be recorded in mono. Recording a vocal in stereo has absolutely zero benefits and makes things a lot more complicated, particularly when it comes to mixing. You can always add stereo effects at that point.

That’s a fair point.
Indeed, I have been wondering if this Mic is a little more than I can deal with.
Also, it seems to be designed to work best with the Apollo interfaces, which I have no particular interest in buying. I’m not quite ready to part with it just yet, but yeah maybe. I have to say, I think it sounds really good on my own voice though.

One big reveal from this thread has been that it’s not clipping; it’s limiting. So my first assumption was wrong. I have no idea how or where the limiter got into my signal path. Really, it’s weird. But I am relatively new at cubase, so I probably did something without realizing it. Still, it’s better to have Limited audio than clipped, I would think.

No I’m not so sure that the harshness I hear it is only in the “Limited” sections. But it does get worse as the vocalist gets louder. She has a nice strong voice but she wasn’t belting at all. So I wouldn’t have expected to need attenuation on the microphone. She was standing more than a foot away from the mic too. Some people have said that the microphone is very sensitive and can have a harsh if you’re too closer or too loud. Maybe it just requires more distance than other mics.

update: listening to the track again, yes, the distortion only occurs when the signal flattens out due to limiting.

That’s not a safe assumption. It’s not clipping in Cubase, agreed, in the sense that incoming signal is well below 0 dBFS. However, it’s possible - and I’ve done it many times unfortunately :sweat_smile: - to pass on to Cubase a clipped signal at normal levels.

For example, I can connect my preamp, crank up the gain, see that the signal peaks and clips, then back off the preamp’s output volume. In this case, if my preamp’s leds are measuring the output, they will tell me that the signal is not hot, it’s just right. But the signal is distorted because of too much gain, it’s just that the output knob has brought it down to an acceptable level. It’s just like the gain knob and the master knob on a guitar amp. You can play with a distorted sound at a low level.

But I can’ find a way to do this with a minimal setup like yours (Mic → Interface → Cubase). If the interface’s channel signal led is in the green, how can it be that the signal comes in clipped like that? It’s a puzzle!

So The clipped signal can produce that flat top I’m seeing in the audio track?
But the clipping could’ve originated somewhere prior to cubase?

So the only possibilities are the mic or the interface.

I guess it must’ve been the interface. I was watching the meter, but it’s possible it went into the red while my eyes were somewhere else. It does surprise me because we did a level check and I had the vocalist sing at maximum volume while I looked at all the levels. But maybe in the flow of the song with the piano playing along she just went louder. Has anyone had that experience?

Or,
Is it possible that the clipping was happening at the mic? They tell me it’s a sensitive mic.
This doesn’t seem likely to me, but it’s worth asking.

yes

yes, it’s possible

ok

I suggest it would be better to go back and test the setup you had during the recording. I don’t think anyone here can tell you conclusively where the clipping came from but a test will show. It will be quicker than everybody just guessing.

Will do.

Condenser mics like the Sphere will clip internally if they do not receive a full 48v phantom supply, or if the phantom supply is unable to deliver sufficient current at 48volts.
I have also seen large diaphragm condenser do this when they are damp?

S

So, a bit of an update since I last posted:

I’ve been in contact with universal audio, which has acquired Townsend Labs.
I sent them audio files, and screen capture so they could see what I was doing.
They did say the raw files were a bit too hot.

So why would the calibrator say my signal was too low even when the screen capture repeatedly showed my signal was above the minimum? Well, turns out the signal that you send into the mic has to be pretty at a consistent level for calibration to work. So a nice evenly sung “ooooo” is going to work better than “La la la” for example.

I’m pretty sure the clipping was in the audio interface as I was able to re-create the same flattening of the audio wave by pushing the interface meters slightly into the red. I just must’ve missed it during the session.

In summary, I cranked up the gain to get the calibrator to accept my signal, and in so doing, caused the signal to clip. Now that I better understand how to calibrate, I don’t need to bring my signal nearly as high and I shouldn’t have this problem.

Thank you everyone for your help. I learned a lot from reading all your comments.
Dave

4 Likes

Great that you could solve the puzzle. :+1: