The Glossary: Audio terminology explained

[otek]

Following requests from posters who might find a lot of the engineering discussions and associated terminology hard to understand, I've decided to start a thread about this.

This is a place to ask about any unfamiliar word, term or phrase used in the discussions on any of the fora. Anything that may seem unclear. As such, there won't initially be any material here, but more and more will be added as new questions are asked and new terminology discussed.

May I also ask that questions or words regarding larger topics, such as gain staging or mic placement, be kept as specific as possible. In other words, you may ask about the term gain staging, but the explanation will be kept reasonably brief - you may expect something like:

"The concept of organizing and controlling all the various amplification stages in a given signal chain, to avoid distortion and optimize signal-to-noise ratio."

If you need more than this, make the question more specific, or better yet, post it in a separate thread.



Thanks,

otek

[TSTW]

cascading busses?

[Skwaidu]

cascading busses?

Not a very common term IMHO... I'll try anyway.

= To run audio through several busses in series to accomplish something, usually combined processing... Either to process tracks together for a sound(like group compression) or for convenience. (EQing a stack of vocals with one eq)

Example: Take 8 tracks of drums, combine 2 snare mikes to a buss, gate and eq, send said buss to another buss with the rest of the drum tracks, compress, send to mix buss... Compress!!! :D

[lebouche]

Two and a half questions

1.)WTF is normalised about..cable wise like on my patchbay tis

1.5)it the same as balanced an unbalanced which I don't understand either.

2.)What is the difference between normalizing a wav and changing the gain.

Tanks,

[Brendo]

1. Normalling... "Full Normalled" wires top row to bottom row unless you have a plug in either row. "Half Normalled" wires top to bottom, unless you have a plug in the BOTTOM row, otherwise you can take a copy of the signal from the top, while the top to bottom is still hard-wired inside. Or you can have no normalling, which normally wires NOTHING together.

1.5.... no. Balanced signals have a signal component, a polarity inverse signal component, and ground. Noise induced in the signal path while travelling over a 3 conductor cable, such as an XLR mic cable or a TRS cable, is induced equally in the hot and cold conductors... the signal on the cold, the polarity inverse signal, is flipped back to correct polarity in the balanced input of your gear, and any noise induced in the cable is then cancelled out and you get a bit of a signal boost from summing the two signals.

2. Normalizing generally brings the signal to 0dB or to a specified number, e.g. -3dB while changing the gain allows you to change the gain in decibels, e.g. a 3dB boost or cut.

[otek]

It should be added to item #2 that normalizing is a destructive process that writes changes to the audio file, whereas changing the gain can be either non-destructive (via fader or plugin) or destructive (changing gain in the sample editor).

Also, changing the gain or normalizing are both processes that compromise the resolution of the audio, and should ideally be avoided, unless you run into very specific problems that require the change be made.

[st robert]

all right, how about these pro audio terms:

clownfucker

red shirt mix

i just don't wanna use them on the wrong project....

rob

[st robert]

also, i wonder what makes something "class a" and why would a transformerless design be of benefit?

[Spock]

OK, your last question first. Because to build a good quality transformer means the device will cost more, and the device will be heavy. To this end, a lot of old TVs did all kinds of tricks to dump the power transformers for the heaters.

The other reason in some cases, like an amp output, is that you don't need to do the impedance convertion that a transformer gives you. A tube is not going to be able to directly drive a 4 to 16 ohm load, but a transistor can do it just fine.

Class A is simple, it just means that the one active device is conducting the whole cycle and never gets to saturation or cutoff.

[otek]

clownfucker

red shirt mix




clown·fuck·er (klounfŭk'er)
n.

1. Audio term coined by Volthause (http://womb.mixerman.net/member.php?u=207).

2. Person engaged in the process of recording guitar or bass using modeling amplification.

See also related forms:

clown·fuck - verb
clown·fuck·ing - noun
clown·fuck·ish - adjective
clown·fuck·ish·ly - adverb



red shirt mix - Philosophical concept pertaining to audio engineering, established by Loudist, the basic premise of which is the existence of, adherence to, or ability to discern, the essential motif or element in the structure and combination of recorded tracks within a song at any given moment, and to promote this motif or element at all times, by way of all available engineering methods and techniques.

[HOOK]

going back to "Normalizing vs gain" and splitting some hair in the process.....

Normalizing: your computer analyze your soundfile first and find the "loudest" part, and then "amplifie" the file so the "loudest" part is just below zero (or whatever value you choose...)

Gain: Let you apply a certain amout of "amplification" to your file (destructive or nondestructive) regardless if you hit the ceiling or not....


NB!! The " " is there because we are inside a computer and as we all know this is lalaland and does not really exist! :Wink: :Roll eyes:



HOOK

[TSTW]

Mook.

I know what they are. I want to know where does the name derive from?

[Immanuel]

also, i wonder what makes something "class a"

Class A uses full power load all the time. Therefor it always has the power awailable, and it does not need to "request" it from the PSU. This also means, that unless you use it to the max all the time, you are waisting a considerable amount of electricity, which is transformed to heat.

[zombie69]

2 bus - is that the same as a stereo master fader?

[otek]

2 bus - is that the same as a stereo master fader?

Yes.

The main stereo pair of output channels for the whole mix.

Depending on context, "stereo bus" may be synonymous with this too, but it can also be taken to mean any stereo bus within your signal flow.

[otek]

This thread is intended as a quick and easy reference for people who want to find out the meaning of a specific word, or even a recurring phrase or term on the fora.

Therefore, I have elected to delete a few posts in this thread - including one of my own - that don't specifically address this purpose, so as to have a minimum of clutter in here. I hope those of you who have gotten their posts removed won't have any objections to this.


Thanks,

otek

[otek]

Some common signal flow and editing terms:

Comping: The process of combining selected parts from several different takes, to make up one cohesive whole. Typically used for crucial mix elements, that often contain some measure of improvisation - e.g. vocals or solos.

Summing/Bouncing: the combination of several recorded tracks onto one or two tracks. Was common in analog recording, where the track count was strictly limited, but may still be used in ITB recording for convenience.

Summing/Bussing: the combination of several input sources onto one channel of a console or track on a recorder.

Doubling: The process of recording the same part twice (or more) to create a fuller sound. Most commonly used on vocals and guitars. Maintaining the consistency of the performance, pitch and timing from take to take, is essential with this technique.

Note that in certain cases, such as with labels on specific pieces of equipment, the terminology may vary somewhat. The above descriptions are very commonly used, however.


otek

[HOOK]

What would be you gentlemen´s definition of "Headroom"?



HOOK

[otek]

What would be you gentlemen´s definition of "Headroom"?

Headroom is the amount (in dB) with which a signal can exceed a given reference level before clipping.

In analog, this reference is usually equal to the +4 standard (1.23 VRMS). This corresponds to a standard level of -18 dBFS in the digital realm.


otek

[HOOK]

So you could say that, as the given reference level in the digital realm is 0 dBFS, there are no head room in the digital realm?



HOOK

[Immanuel]

It is my impression that in digital, headroom is often refered to as the difference between the maximum expected level and 0dBFS.

[otek]

So you could say that, as the given reference level in the digital realm is 0 dBFS, there are no head room in the digital realm?


Are you being facetious? :Razz:

I just said that the reference level in digital (with respect to the +4 standard) is -18 dBFS.

There is something in EBU lingo referred to as the PML, or Permitted Maximum Level, which is specified to -9 dBFS. So according to that reference, the headroom would be 9 dB.


otek



PS. Also, guys, try to keep guesswork and "Impressions" out of this thread. Stay as close to facts as you can, and if you post something that is subject to debate, please give a source reference.

Thanks!

[Cary Chilton]

Great thread Otek, should have been here a long time ago.

[otek]

Great thread Otek, should have been here a long time ago.

It was, relatively speaking. It was started over a year ago, little over a month after this forum opened its doors.

Sorry I took so long. :D


otek

[Paul Weitheiml]

Much appreciated!:Thumbsup:

[Immanuel]

I just said that the reference level in digital (with respect to the +4 standard) is -18 dBFS.

Various standards have been published for digital operating levels. The early pseudo-video digital recorders mentioned last month adopted -15dBFS (15dB below full-scale) to equate with +4dBu (or 0VU). The European Broadcasting Union have specified a very similar standard of -18dBFS to equate to 0dBu (ie. -14dBFS aligns with +4dBu). In America they tend to use -20dBFS.
Above was found in this article in Sound on Sound http://www.soundonsound.com/sos/jun98/articles/digital2.html

[otek]

Thanks for the info, Immanuel!

I will do a cleanup in this thread soon and among other things, make yours the de facto reply post.


otek

[Bob Olhsson]

Actually it's -18 with an RMS meter which reads -20 on a VU meter.

[Immanuel]

One of my converters is a Creamware A16 from 1997 (obviously, this is not the same as the current A16 ultra). It uses 18bit converters, and I just double checked the specs.

A/D: 0dBFS = +10dBu
D/A: 0dBFS = +8dBV

While this is anecdotal, it adds a bit of historical touch.:)

[waterboy]

Here is one that I haven't bothered to ask until now:

*NoiseFloor (is that right?)

I know what Wikipedia says: but I would like the layman definition, please.

Thank you.

[MacGregor]

Here is one that I haven't bothered to ask until now:

*NoiseFloor (is that right?)

I know what Wikipedia says: but I would like the layman definition, please.

Thank you.

Generally spoken noise floor is every unwanted signal, but normally people mean low level hissing created by mics, preamps, tapes, thus 'floor'.

DrummersMakeNoise,too!Mac

.

[waterboy]

So when somebody makes a comment like; raising or lowering the noise floor, they are talking about the level at which this unwanted hiss is heard?

[otek]

So when somebody makes a comment like; raising or lowering the noise floor, they are talking about the level at which this unwanted hiss is heard?


Yes.

In any audio circuit, there is a constant static discharge due to thermal electronic activity, airborne signals and electromagnetic fields, which gets amplified by the signal chain.

When people are talking about lowering the noise floor, they are seeking to attenuate noise or make sure it stays out of the audio path. There are multiple ways of doing this, such as balancing the signal path, shielding and grounding.

Also, observing proper gain staging can be a very effective way of managing the noise floor.


otek

[TSTW]

Slew rate, slew distortion? Reading the Ulysses thread.

thanks in advance

[Ethan Winer]

Slew rate, slew distortion?

Slew rate is how fast an output voltage changes. This is related to frequency response, but not exactly the same because it also depends on the absolute volume level. So if you measure the frequency response of a device at a low signal level it might be flat to past 20 KHz, and with very low distortion. But at a higher signal level the output stage can't keep up with the input and the result is distortion. In this case a sine wave turns into a triangle wave (the distortion), and the overall level is usually reduced too.

--Ethan

[Tim Halligan]

"What's happening with Rihanna?"

Standard replacement for "Who gives a fuck?"

Suggested by BlackieC.

Seconded by me.

:lol:


Cheers,
Tim

[MGMc]

Gravity: One of those mysteries best left unsolved.

[Darth_Fader]

Pardon me but I want to add a couple of definitions:

Intensity: Aka: Sound Pressure Level. The actual power of the wave propagating in the atmosphere. Measured in dB SPL or like units. Purely a physical measurement.

Loudness: the SENSATION (i.e. percieved) level of a signal. Measured in sones, phons, or as dBEQ, but dB is really not the appropriate measurement. Purely perceptual, but can be measured via complex calculation.

[otek]

Thanks, JJ, for adding that - and for posting in a thread that gets much to few visits these days, myself included.

Welcome to the Womb!


otek

[westicle]

Transients

'splain?

[otek]

Transients

'splain?

The attack part of a sound.

Alternatively, the American word for guys sleeping under a pile of newspapers at the Victoria Street railway overpass.


otek

[Darth_Fader]

The attack part of a sound.

Alternatively, the American word for guys sleeping under a pile of newspapers at the Victoria Street railway overpass.


otek

More specifically, a transient signal is a signal for which the signal statistics vary significantly with time, over a wide variety of time intervals.

Many things vary with time, but over a periodic interval. These are not transient, when you consider them at a time interval of many periods, their statistics settle down to something stable.

Most onsets are in fact transient in nature, of course.

However, look at a vocal pitch. Taken as a single pitch spike, it looks transient. Taken over a second's duration, however, it looks very, very regular. This is why a transient has to be nonstationary (that's the technical term for changing statistical parameters) over a wide variety of time intervals.

If we're talking about the ear, 200 milliseconds would be a good maximum time interval to consider as far as a maximum "transient" window. Things can kinda-sorta become transient at under 60 milliseconds in a perceptual sense, which is not the same as the definition given above, of course. A very low-pitched voice can in fact be transient as far as the ear is concerned, hence some interesting issues with stuff like Thurl Ravenscroft singing "Asleep in the Deep" or the SQAM "German Male Speech'. When this kind of thing happens in the ear, the psychoacoustic effects are often called co-articulation, which is a bit of a misnonmer, perhaps, but that's what it's called.

[westicle]

Awesome, thanks guys.

[cat]

Some abbreviations I come across I don´t get
HPF & LPF
LFO
L.C.R. panning
DDL

[MKZ]

Some abbreviations I come across I don´t get
HPF & LPF
LFO
L.C.R. panning
DDL

They're probably already here in this thread but,

High pass filter

Low pass filter

low frequency oscilator

Left-Center-right

"direct download" ?

[zoff]

They're probably already here in this thread but,

High pass filter

Low pass filter

low frequency oscilator

Left-Center-right

"direct download" ?

DDL = Digital Delay Line

[MKZ]

DDL = Digital Delay Line

aha! no shit! haha!

[MacGregor]

"Recording studio"

A recording studio is a facility for sound recording and mixing. Ideally, the space is specially designed to achieve the desired acoustic properties.

Sometimes misleadingly referred to as "shop" by people grown up in Banjo Centers.

Mac
Hehehe...

[DrumnBum]

Can someone explain aliasing?

[otek]

I believe that would be one for Darth Fader to explain properly.


otek

[Darth_Fader]

Can someone explain aliasing?

Aliasing...

This isn't a definition, it's more of the description of a process. Aliasing is the process of creating aliases for a frequency outside the passband of a sampled system, for a definition. Yes, I realize that's (*&(*& useless.

So what is an alias, you ask?

An alias is what you get when you put a signal outside of the passband (for audio DC to FS/2) where FS is the sampling rate into the A to D.

A digital system has a bandwidth of FS/2. Period. No more, no less, nohow. Mathematically, you can not represent anything outside of that bandwidth (although for purposes not used in most audio capture, but heavily used in audio processing, that bandwidth does not have to be DC to FS/2, but let's not worry about that right now) in a sampled signal. (This doesn't have to be a digital signal, analog sampling provides us with exactly the same effects.) If you do put in something outside that bandwidth, you get an alias. An alias is a NEW FREQUENCY that did not exist in the original signal, and as such, is generally a "real bad thing".

Why is this? Well, to use an example, if you have a signal at 7/8 the cutoff frequency of a standard PCM signal, the DIGITAL REPRESENTATION includes 7/16, 9/16, and N*7/16 and N*9/16 (for integer N) signals, all of which ARE THE SAME FREQUENCY DIGITALLY. (note, the 7/16 is relative to SAMPLING FREQUENCY, which is twice the cutoff frequency)

This is discussed in great detail in a deck titled something like "Conversion ..." at www.aes.org/sections/pnw/ppt.htm and note slides 25 to 29.

Another thread somewhere is probably where we should discuss this if you want. Ping me with a pm if you start such a thread.

[meLoCo_go]

Can someone explain aliasing?
Too add what j_j said:
Remember that movie effect when the wheel on the car starts to rotate and it first it is as normal but then it sorta slows down, then stops, then starts to rotate backwards?
That's aliasing too.

If you do put in something outside that bandwidth, you get an alias.
I think that I'd add a little example on how something outside the bandwidth may appear.
Say you have a 10kHz sinetone@44100 and you run it through some process that generates harmonic distortion (like tapesim or ampsim) -- your 2nd harmonic would be 20kHz which is fine as it is below 22050Hz which 44100 may correctly represent. BUT your 3rd harmonic would be 30kHz which is higher than 22050! What would happen? It would not dissapear, but sort of "bounce back" into the 22050Hz bandwidth at a frequency of 44100-30000=14100Hz (if I didn't screw up my math). So, you have a new tone which is harmonically unrelated to fundamental.
So, any process that generates signals outside of digital bandwidth would result in signals "bounced back" into the bandwidth, and that would simply add more noise. The oversampling techniques increase the available bandwidth to reduce aliasing.

[DrumnBum]

Thanks guys, that really helps a lot. I realize that it's a subject that is pretty in depth, but it really helps give me a basic understanding so that at least when I hear the word used, I can follow the discussion much better.

:vuvu: