Good day! I’m not sure I’ve gotten an answer to my question yet. Probably because I have not asked it correctly.
I’m not sure whether clipping is bad or not. Someone said that it can damage equipment. What if I have some audio that clips 3dB off the tops of the waves. Is it neccessary to fix this, or might I just reduce the volume by 3dB and leave the wave clipped? It seems that, with that treatment, no damage could occur to equipment.
Also, can you recommend a good forum to learn more about processing sound? I want to ask about using different kinds of equipment, like acoustic guitars, and that might be out of the range of this discussion forum.
Not digital equipment. The digital system just doesn’t care. You might run into analog problems. If you play highly clipped music loudly into a good speaker system, the high-pitched, crisp, gritty clipped sound could damage the tweeter.
I’m not sure whether clipping is bad or not.
It’s bad. Clipping is a point where the music system stopped following the song. In digital land, the digital system normally assigns numbers to all parts of your song. Clipping is when the song is so loud, the system runs out of numbers. Worse, yet, the system starts making up its own trash.
It’s not good. It’s not. It’s also permanent. You can’t start with clipping and get back to the original song.
might I just reduce the volume by 3dB and leave the wave clipped?
It’s up to you. The quality of the song will not change and you will have a 3dB quieter sound which is almost inaudible.
Clipping is one of the four horseman of reliable ways to kill your show (#2).
can you recommend a good forum to learn more about processing sound?
Damn. I think I’m riding all them horses! Well, no, I don’t ever use mp3. I always use wav. I’m recording in hotels so I can’t do much about street noise and other noise in the hotels. The only thing I have control over is the clipping. For that I always record the base track in Dual mode with a backup at about -6 or -8dB. For overdubs I can’t do that, but I do tilt my double mics so that one catches the sound straight on and the other elliptically. I can edit out clipping with the wave from the angled mic.
I can only travel lightweight, so I carry a Tascam DR40. That’s better than nothing. I’m actually doing pretty well. And no, this is not spose to be studio recording. This is what it is and it’s just different. It’s alot easier to experiment outside of a studio. But I see that if you pile tracks on top of eachother, the noise does build up. Studio recording does sound a lot better, though, so I have mixed feelings about it. I could treat it as demos to be used for professional treatment later in a studio. But like Springsteen, and his home recording for Nebraska, I am pretty sure that I’ll find out that - that’s the album. I don’t know. McCartney’s first album recorded at home in his basement is much better than any of his studio work afterwards. But he had pro equipment, was located on a farm and prbably had the basement soundproofed. Even so some reviewers complained about the production. I think it’s a classic album from him. That’s the definitive McCartney album for me.
The Tascam recording might be improved by using an external mic, but I can’t travel with guitars, and mics and hard drives. Plastic and steel is heavy stuff to be lugging around.
I carry a Tascam DR40. That’s better than nothing.
Unless you have some special requirements it should be fine and it’s probably better than lots of things! There’s no fan making noise and compared to a computer there are not as many things to mess-up. It could be GREAT in a soundproof studio!
For that I always record the base track in Dual mode with a backup at about -6 or -8dB.
Your main track can be at the same level. With digital recording there’s no need to get close to clipping so you can leave plenty of headroom. You can boost later.
If you are old enough to remember analog tape, you needed a hot signal to overcome tape hiss, plus analog tape is more forgiving and it starts to soft-clip when your go 'into the red". But with digital… No tape noise but it hard-clips at exactly 0dB.
Thanks for the replies. Why I am confused is not only that my writing is unclear and I go off on tangents but that the responses differ significantly. Koz says NO clipping.
It’s not good. It’s not. It’s also permanent. You can’t start with clipping and get back to the original song.
OK, so why have clip fix?
Someone (Doug?) in another conversation suggested that light clipping is OK. I’ll try to find that quote. But even in this discussion he talks of “soft clipping.”
Yes, I do remember. I always went for the hottest signal I could get. Maybe I am creating some problem by trying that now? But I am trying to mask environmental noises. I spose I’m not accomplishing much. Bringing the volume up or down on the mic is going to affect background noise either way.
ClipFix is an attempt to rescue clipped audio that would otherwise be trash. It works by looking for flat topped peaks and guessing what the missing waveform should be. The “guess” is a mathematical algorithm called “cubic interpolation”, which in layman’s terms means “replacing the flat top with a smooth curve”.
As you might imagine, it can produce a reasonable guess if there is only very mild clipping, but if there’s a lot of clipping then it really doesn’t stand any chance of producing good results because it has no way of knowing what the “missing” (clipped) audio should be like.
At best, ClipFix is just a guess at what the clipped peaks may have been like to make the damage less noticeable, but it’s more of a sticking plaster than a cure. For sever clipping it’s like putting a sticking plaster on a broken leg
Thanks Koz, Doug and Steve. I remember first recording on a TEAC (?) reel to reel. I’ve also recosrded with a Tascam 4 track cassette. I always had the Vu meters just touching the red. I read this instruction in a book on recording. I had no isea that it was to overcome tape hiss. So, I can use much lower input levels? I was recording at 70 to 90 depending on the signal strength. I’m now trying some lower setting like 50 or 60. Is there a limit to that? Can I set my level to 1 or 2? Why or why not?
Steve - does cubic whatever have something to do with analytical geometry? I know that sound wave is three dimensional. It has to be represented on the two dimensional window of Audacity. A cube is 3D. I would spose that you need at least two cubes to represent a sound and a third to connect them together. What dimensions are being represented by a cube? I imagine the height for amplitude, and another dimension for the longitude of the wave cycle or pitch and another for duration? But there are also other elements to a sound such as timbre from the overtones, and attack and decay.
In other words, I don’t know what you are talking about?
Your sound card probably captures audio as 16-bit numbers. That means there are 65536 discrete amplitude levels, which allows the digital audio to match the (continuously varying) analog sound very accurately. If you “only” use half of the available range, then there are still 32768 distinct levels, which is still enough for very high quality. However, if the recording level is extremely low, then there will not be enough discrete levels to be able to accurately represent the analog sound, and the sound quality will suffer.
This file is close to the best quality possible if you recorded at -70 dB with a 16-bit sound card:
Not really. It’s just a way of smoothly filling in a gap between two points on a curve.
Yes. -6dB (50%) is fine, and if your levels are unpredictable and you occasionally get clipping you can go lower. Your recorder works at 24-bits which give you enormous dynamic range and you could record at -20dB (10%) with no loss of sound quality!
That’s assuming the levels are low because you’ve turned-down the recording level… If your levels are low because you’re across the room from the mic (or something like that on the acoustic/analog side) that will affect quality.
I was recording at 70 to 90 depending on the signal strength. I’m now trying some lower setting like 50 or 60. Is there a limit to that? Can I set my level to 1 or 2? Why or why not?
You’re “settings” aren’t important, it’s the actual level you’re getting.
Steve - does cubic whatever have something to do with analytical geometry? I know that sound wave is three dimensional.
Steve will have to explain the algorithm but audio can be completely defined in 2-dimensions (time & amplitude).* [u]digial audio[/u] is simply a series of samples (amplitude values or the “height” of the wave) with each sample representing one point in time. (i.e. 44,100 samples pre second). When you play the audio, the digital-to-analog converter “connects the dots” and re-creates a smooth analog “waveform” (a varying voltage). The digital values could represent the position of the loudspeaker as it vibrates to reproduce sound.
The “normal” waveform is time domain representation. It can also be represented as frequency over time (the frequency domain) and that’s what you see in the spectrogram view. Some effects/processing is done in the frequency domain. But “basic” digital audio (analog-to-digital-converters, digital-to-analog converters, WAV files, etc., are all time-domain.
This wiki page refers to a “Fourier transform.” Isn’t this analytical geometry?
“In mathematics, a Fourier transform (FT) is a mathematical transform that decomposes functions depending on space or time into functions depending on spatial or temporal frequency, such as the expression of a musical chord in terms of the volumes and frequencies of its constituent notes. The term Fourier transform refers to both the frequency domain representation and the mathematical operation that associates the frequency domain representation to a function of space or time.”
Then light clipping is OK but it must be corrected with ClipFix?
No clipping is OK. Clip Fix produces a mathematical algorithm, not the original sound. It reduces the gritty, annoying, crunchy sound without actually fixing the music or voice.
Unlike tape, the digital system has straight-line transfers. No gentle magnetic saturation and ultra-sonic bias. You get louder until the system runs out of numbers and whacks off your peaks with a sharp knife.
The general recommended voice recording level is peaks about -6dB or half-way on the blue waves.
The bouncing sound meter starts changing color right about there, going from green toward yellow. If you hit red, you’ve got clipping damage.
Also remember the digital system isn’t insanely noisy like tape. 16-bit sound range is 0dB (clipping) down to -96dB! You can slide sound quieter and louder anywhere in that range without damage. Don’t run out on either end, and don’t overload your analog system. Watch those red PEAK lights.
You can certainly produce a robust, dense sound for a show, but it’s strongly recommended you produce a clean recording, save it for protection, and then get there with effects. We can go from a clean recording to anything else, once you produce distortion right at the top, you’re stuck.
Yes, but. Pay attention to your production values and associated productions.
A podcast I like has started cranking out messy work and making mistakes I don’t remember happening before. Be sure to Like and Subscribe and remember to hit that bell icon. and then snaps right back to normal volume just when you’re leading forward to hear they they said. Yes, it’s attention-getting, but if you grab too much attention, I’ll lean forward and change the channel.
Two flying shows settled on different announce volumes. One settled down here and the other is a good hand-full of dB louder. I’m wearing out the Volume/Balance knob—and yes I do have one. I don’t have to click on anything or fish the phone out of my pocket.
I know of three production sound standards. The audiobook one where they tell you how loud to get in actual electrical measurements, the LUFS one which is supposed to be an up and coming universal standard, and the cheat where you Effect > Amplify to -1dB, just short of clipping and ship it.
Audacity Effect > Loudness Normalization supports both RMS (Audiobook) and LUFS.
Honorable mention goes out to The New York Times for their Mini Crossword Puzzle “success” fanfare. It’s loud enough to crack plates and notify the neighbors. You learn to turn the volume down before you put that last letter in the box. I can’t tell if they did that on purpose or not.