Is clipping bad?

I don’t know much. I keep Show Clipping checked, and whenever I see clipping, I correct it. I thought I read somewhere that if it sounds good, it is good. I know about overdriving guitars on stage and in the studio. Is that the same as a clipped wave? I loaded an old 70’s punk recording to audacity to view it and it was massively clipped. Then I watched some videos of engineers talking about mixing and they said “Don’t bring us a clipped recording to mix.” So what’s the lowdown? Should I try to keep my recording stage unclipped and then experiment with clipping in the mix? that sounds reasonable because I’ll always have the unclipped original if I screw everything up.

Maybe the style matters? I’m trying to produce a hip hop style track now. What’s the norm in that style? But I can see where some kinds of music should be constructed of more purely recorded and mixed sounds, such as the Beatles or Beach Boys and a lot of the more modern commercial groups.

Don’t just “try”, ensure that you “succeed”. :wink:

Clipping is when the peaks of the waveform are literally clipped off. Clipping during recording is permanent and cannot be totally fixed (There are effects that can attempt to restore clipped peaks, but they are rather limited in what they can do because they have no way of knowing what the peaks should actually look like, because they are not there.)

So called “hard clipping” can also cause damage to a speaker’s tweeters due to excessive amounts of very high frequencies caused by the clipping.

Easier said than done. I still was not able to EQ that last project correctly. Anyway, I hold my dual mics in a position that captures the waves from different angles and I also use a -06dB setting because my DR40 has dual recording. So, I can correct clipping with correct “clips” of a non-clipped recording. HA! I use a clip of a non-clipped to replace a clipped.

My recordings are made to be heard through headphones. I want to spare the rest of the population.

It can also depend on where. “Fuzz Guitar” can be achieved by overloading the guitar amplifier and in the case of older tube guitar amps has a distinctive sound because of the way vacuum tubes work. The amp is now a part of the instrument. You can’t get that sound with the guitar by itself and you can’t get it with transistor amps unless they have a “tube sound” setting.

“Clipping” while inside Audacity doesn’t matter. Audacity uses an internal format that doesn’t overload, so if you find yourself with red bars on the timeline and you caused them with a filter or an effect, it should be possible to just reduce the volume and bring everything back to normal. The audiobook people can run into this. It’s not unusual to get red bars in the middle of mastering. Doesn’t matter.

You can have almost clipping. This one drives developers nuts. We can’t actually measure clipping since for most sound files there is nothing over 0dB, and so there’s nothing to measure. The best we can do is measure intent. There is sound at 0dB and if it could have gone over, it would have. There is a balancing act of one sample at 0dB is not a big deal but three or more is. There are no natural sounds with 3 or more samples at 0dB, so it must be clipping damage. It’s this last one that gives you a song in a blizzard of red bars that has no overload or clipping damage. The music producer intentionally did that.

Clipping almost anywhere else means the digital sound file didn’t follow the show. If you have a raw sound file with red bars in it, then you have sound damage. That’s the studio note to never show up for production with a clipped sound file. Those are not recoverable. Further, a studio that produces a bad show makes everybody think they screwed up. Everybody loses.

if it sounds good, it is good.

To whom and on what? Clipping tends to make very high pitched crackling, clicking, ticking, or crashing sounds. Those will drive a 16 year old girl nuts, but a 50 year old man may not even hear them. Home performers can have sucky speaker or earphone systems and are shocked the first time they hear their works on a good quality sound system. There is a “Hollywood Standard” headphone model that is not particularly good for casual listening, but is revered on the set for showing you damage before anyone else can hear it.

The show has to sound good to the person writing the checks.

Koz

I did hear about dual recording systems. Fascinating idea, and field engineers will love it, but that gives you a production recording with a clean show in there somewhere. It’s your job, through hours of editing to coax it out. That might be useful for an occasional accident, but not a regular production process. Far more efficient to pay attention during the recording and turn the volume down if needed.

That’s also the caution of going into a recording or shoot intending to fix it later. Planning on disaster recovery isn’t a plan, and it’s possible there is no recoverable show.

Koz

If you want clipping for “creative reasons” it should be added after recording so you can adjust the amount of clipping, or you can change your mind.

DR40 has dual recording. So, I can correct clipping with correct “clips” of a non-clipped recording. HA! I use a clip of a non-clipped to replace a clipped.

Or you could just throw-away the clipped version or leave yourself plenty of headroom and skip the double-recording. :wink: With digital you’ve got loads of dynamic range so you can leave plenty of headroom. Pros often record at -12 to -18dB. (In the analog days you wanted a hot signal to overcome the tape noise. With digital, recording levels are not critical as long as you avoid clipping.)

Or, if you are recording “live” where levels are unpredictable and there’s no chance for “take-2” you can use a hardware limiter in front of the analog-to-digital converter.

I loaded an old 70’s punk recording to audacity to view it and it was massively clipped.

Assuming this is a professional recording It probably wasn’t clipped in the 1970s. :wink: It may have been clipped when it was digitized. If it was digitized commercially it wouldn’t have been clipped during digitization but it may have been clipped during re-mastering (some [u]Loudness War[/u] CDs/MP3s are intentionally clipped). Most “serious” music lovers consider clipping to be a bad thing, even when done intentionally.

Or, it may not actually be clipped, especially if it’s an MP3… Audacity shows potential clipping where it goes over 0dB or where there are multiple 0dB samples in a row. Audacity isn’t analyzing the wave shape. If you have a 0dB normalized file and you make an MP3, the MP3 often peaks over 0dB. MP3 slightly changes the wave shape making some peaks higher and others lower and MP3 can go over 0dB without clipping. It’s not (necessarily) clipped* but Audacity will “show red”. Or, if you do something in Audacity that boosts the levels (like boosting the bass), Audacity will show red even though the waveform is not clipped (yet).

On the other hand, if you have a clipped file and you reduce the levels in Audacity it will no longer show red even though the waveform is still clipped.

\

  • If you have an MP3 (or other format) that goes over 0dB you’ll clip your digital-to-analog converter if you play-back at full digital volume. But as far as I know the slight clipping caused by MP3 compression is not audible.

I’m slow to understand this stuff. I’ll give you a bit more detail on how I record. I’m embarrassed because it’s obviously amateur, but I spend much time on the writing and good writing can outshine poor production. However, I would like to improve the sound if it makes the lyrics come through more intelligibly.

I recorded all of the backing clips previously scrap by scrap and bit by bit in a hotel room. So, now I’m editing all of that together to get complete backing tracks. It’s a little sloppy because I’m doing it on the fly, and my environment leaks into the recording, which ain’t bad, because I hate the sterility of modern recordings. They sound like they live in a vacuum. I don’t. If a truck rolls by or a bird lands on my window sill and starts chirping, that’s what you get. Of course, that accumulates in multitrack editing and can get too dirty. I edit in a heaping helping of the vacuum packed stuff that you can get from exporting Musescore clips into the mix, to balance the grime.

To do the vocal dubs, I use a Tascam linear PCM DR40 handheld. The mics are adjustable. So, I put them in the outward or open position as opposed to the closed position. Then I hold the unit in front of my face with the narrow edge of the devise facing me. In this position, I have one mic facing my voice and the other is facing directly opposite. So the waves are entering directly into the first mic and passing over the second mic, relatively speaking. I also have the dual setting at -06.

I load these two stereo recordings to audacity and split the tracks to mono, and now I’ve got four signal options to choose from. that leaves me with a hot track at full volume, a softer one (which clips less if at all but sounds tinny,) and then duplicates of those at -06. I have to have the hot track because I can’t monitor as I record, due to having to monitor the backing track with audacity running. My other option is to load the backing track to the DR40 and monitor while in Overdub mode, but then I lose my four option because I’m not in Dual Mode when in Overdub. I should probably buy a small mixer and an external mic, but I’m unemployed due to the current worldwide situation.

Show Clipping is the first thing I look at. Then I look for 100 hz rumble and correct that. I go through the vocal phrase by phase and pick the best option and copy and paste the best options from the four options. One track might have rumble but the other three don’t. Two tracks are clipped but the other two are less so or not at all. So I just work through copying and pasting together the best of them. I fix clips this way if I can, and sometimes with clip fix in very short sections. I end up with a reconstructed stereo track, with the best options in each channel.

So, as you can see, most clipping is happening at the mic. And I gather that this is no good. I might play with that dual setting and go -08, then -10 and see if that works better. Other clipping is happening in the mix. Every three of four days I get ready to make a premix export in wav. I first test how the mix is going to look when exported by mixing down to a new track. If I have red bars, I look in the all of the tracks above the mix and note where frequencies are clashing. Sometimes I bring the gain down of a whole track, and other times I reduce the volume of a short section that is clashing with another section by using the Amplify in effects.

I need to learn more about how to read a waveform. I can spot rumble by sight, but there are other wave defects that I don’t understand. Some are common and I feel that I’m missing something because there are ten factory presets in the Graphic EQ, but I don’t know where or when to apply them. I assume these must be there for correcting waveform defects. I would appreciate a list of waveform pictures and the matching effect that would correct it. If I learned that, I think it would push my production forward another notch. Does such a chart already exist? Or would someone care to make this chart for waveform deformities and corresponding effects for correction?

After the pandemic, I might go out and splurge on an external mic and mini-mixer. That will improve the sound. I also have an USB mic in storage in another city that I’ll pick up later. I’m pretty much holed up in a hotel until the pandemic subsides and am immobile and making do with what I’ve got, which isn’t much. But that’s the adventure. I’m quite satisfied with some of the results so far, but am interested in exploring more options.

Thanks for your consideration.

Here’s a screenshot of an edit I’m working on. You can see the four channels have been split into four tracks. The top two are the wave passing over and entering into those two mics in opposite directions. The bottom two are the same at -06db.
Screenshot from 2020-09-11 21-31-12.png
Below, I have highlighted a wave deformation that I do not know how to correct. It’s a common one and I would think one of the presets would work for this. But, as you can see, I have been able to rely on the mic that caught the wave at a different angle and it’s -06db counterpart. I can copy, paste and amplify that and get by.
Screenshot from 2020-09-11 21-23-49.png

Rumble filter ? …

rumble filter.gif