How to make "distorted vocals"?

[conroekid]

Hello and thanks to everyone, for making such a great program and forum.
I apologize, if this has already been covered or is on the wrong board.
My question is about making distorted vocals.
I accidentally recorded a vocal, with the level too high, and it sounds great. (I am working on a lo-fi pop record.)
I want to do this again, but am afraid that I will hurt the program, my sound card, or the computer.
How loud is too loud?
Also, is using the "amplify" effect and distorting it that way safe?
P.S. I am using version 1.2.3.

[kozikowski]

It's not like the older analog amplifiers where you could hurt something by overloading the sound channels. In a computer, overload or clipping is effectively where all the digital sound numbers go to "1" and no further information is possible.

There ways to cause damage, but they involve analog before and after the computer. You can cause physical damage by plugging a very high level sound signal into the microphone input of your computer and you can cause speaker damage (also analog) by putting a very highly distorted performance in at high volume. Distortion tends to produce crisp, sharp, screechy signals and that tends to pile up in the tweeter. No, tweeters are not supposed to smoke like that.

It's not likely you'll get into either of those conditions. Long before the loudspeaker fails, you'll be peeling the dog from the ceiling.

Koz

[steve]

As Koz said, overloading in the digital realm is not a problem, but overloading the input to your sound card can make it smoke. Similarly, playing back recordings with a lot of high frequency content (such as is caused by clipping) at high volume can blow the tweeters in your loudspeakers (they do not smoke, they just go "click", and then all the sound from the speakers sounds muffled due to the lack of tweeters).

There are several plug-ins available for producing various kinds of distortion.

Using the "Amplify" effect with "allow clipping" will not hurt Audacity.
If you apply a bit of Equalization to drop down the highest frequencies you may find that you can retain the desired sound while reducing the risk to your speakers (and you local Canine and Chiroptera communities will thank you).
You can always check the frequency content to ensure that you are not overcooking it too much by looking at the spectrum (Analyze menu) - this is much improved in the latest 1.3 version.

Another simple way to deliberately clip audio is using the Nyquist prompt.
(the examples below work with Audacity 1.3.x - they may work with Audacity 1.2 but I've not tested them)

For a mono track, if you enter the following text into the Nyquist prompt dialogue box, it will clip the selected audio at a level of 0.7 (as indicated on the vertical scale of the track).

Code: Select all

(snd-clip s level)

Note that you can change to "0.7" to any value you like between 1.0 and 0.0. A level of 1.0 will not clip anything, while a value of 0.0 will clip all the audio and produce silence

For stereo tracks it is a little more complicated, but the following code will work for mono or stereo:

Code: Select all

(setq level 0.5)
(if (arrayp s)
(vector (snd-clip (aref s 0) level)(snd-clip (aref s 1) level))
(snd-clip s level))

Note that in his version, the clipping level is set in the first line ("0.5")

Another simple distortion type (mono only) is given below.
The interesting thing about this distortion is that it distorts low level signals as well as high level signals, so even a modest amount of distortion will be clearly evident throughout the entire processed selection and not just the peaks.
The strength of this effect is controlled by changing the "wet" value on the first line ("0.2" in this example)

Code: Select all

(setq wet 0.2)
(setq hps (hp s 50))
(setq pos (sum 1.0 (clip (diff hps 1.0) 1.0)))
(setq neg (diff 1.0 (clip (sum hps 1.0) 1.0)))
(setq dist (sim (cue (snd-sqrt pos))(cue (mult (snd-sqrt neg) -1.0))))
(sim (cue (scale (diff 1.0 wet) s))(cue (scale wet dist)))

[conroekid]

Wow. Thanks, so much, to both of you, for your great help!

[conroekid]

Thanks again, for the great help. Y’all’s insight was great and this was my first time to use plot spectrum analysis and Nyquist, both of which are really fun! So, thanks, for introducing me to them.
Anyway, I think that the sound that I like most is: recording it at an overly loud level and then checking the plot spectrum, to see if it will damage the loudspeakers. But my problem is that on the plot spectrum, I don’t know what to look for. I have version 1.3.7 and have looked at many places and see a lot of peaks, less than 1000 Hz, that are over 0 dB. And I’ve seen some that are over 1000 Hz that exceed 0 dB. Am I right to look for excesses of 0 dB? And does it matter at what Hz these excesses happen?
Y’all have been a great help already, so thanks, for that.

[steve]

As an example, this is likely to be unhealthy for tweeters and drive dogs crazy:

screenshot.png (52.68 KiB) Viewed 3489 times

[conroekid]

Thanks, so much, for the great help. How do these look?

[conroekid]

I think that my images didn't show up. I'm sorry.
If you want to look at them, I think that they are at these urls.

http://viewmorepics.myspace.com/index.c ... D=47500676

http://viewmorepics.myspace.com/index.c ... D=47500721

[steve]

If you want to look at them, I think that they are at these urls.

I just get the error:
"This photo cannot be viewed because it has either been deleted or marked as private by the user."

[conroekid]

I'm very sorry that those links didn't work. Thanks, for all the great help.

http://www.flickr.com/photos/[email protected]/3332565042/

http://www.flickr.com/photos/[email protected]/3332571860/

http://www.flickr.com/photos/[email protected]/3332565038/