Sound card lingo...?

  • Thread starter Thread starter PRiZ
  • Start date Start date
P

PRiZ

New member
Sup!...I'm new at this board, I've been looking for a sound card for a while. I had the C-port or DSP24 with ADC & DAC in mind, but am not sure anymore what card I should buy.
Although I've generally learned alot about sound cards I can't seem to understand the Spec language...Please, if you could help me out with a few and eventually have them all explained so I can understand what each card is somewhat capable of...

what do they do and what are they needed for etc...

- channels (like multi-channels etc)
- lightpipe (often after something)
- ADAT
- AD/DA converters or A to D converters
- converters
- midi I/O (I/O = in and out, i know this)
- Digital I/O
- 24 bit internal DATA path
- 24-bit resolution
- Up to 96kHz sample rate conversion
- 120dB SNR
- 120dB Dynamic Range
- 1/4" TRS analog 4 In/8 Out I/O
- 4 16 channel MIDI jacks
- S/PDIF, Optical stereo digital I/O
- Real time (when it says real time effects or something)

The list goes on...
I read the post about converters and found it fascinating, but couldn't understand most of what was being said...
like noise floor db108 ect, is this how low the sound can go or what? Thnaks for reading, look forward to any replys.
 
Outboard converters are typically a box that converts an analog input to digital (A/D) and/or digital to an analog output (D/A). You can buy A to D converter units that just convert audio to digital, or D to A units that just convert digital to audio, or combination AD/DA units that do both. You can buy units that handle just a single chanel or audio, or two chanels for stereo, or even more chanels for multitrack recording. Analog inputs and outputs can be either balanced or unbalanced. Unbalanced inputs/outputs are usualy on 1/4" jacks. Balanced inputs/outputs are usualy either XLR connectors or 1/4" TRS, which is just a 1/4" sterio jack with audio signal on the tip and ring and a seperate ground on the slieve.

In an A/D, audio inputs are converted to digital which must be sent somewhere to be processed or recorded. Similarly a D/A converted must get its digital output from somewhere. There are several standards that have developed over time for sending digital audio between various pieces of equipment. The terms lightpipe, ADAT, S/PDIF are used to refer to one of the standards for digital I/O.

A/D converters work by sampling the audio signal at regular intervals and converting it into numbers. D/A converters do the reverse converting a stream of numbers into an audio signal. There are several comon rates at which this conversion is done 44.1 Khz (44,100 time a second - CD rate), 48 Khz (ADAT rate), 88.2 Khz and 96 Khz. Because of the need to maintain compatibility with older equipment, some never devices include sample rate conversion which allow a digital stream at one rate to be converter to another. For example if your using an effects unit that can only handle a 48Khz digital stream, but all your samples are at 96 Khz, it might be handy to have a unit that can perform 96 Khz sample rate conversion to 48Khz.

Another variable is the size of the number that is represented by each sample. the larger the number the more accurate the representation of the original analog signal. CDs represent each sample at a 16 bit number, which means it can have 2 to the 16th power, or 65,536 different values. 24 bit resolution can represent 2 to the 24th power values, which is 256 times as much as 16 bit audio. While this may sound great, the reality is that the audio side of things usualy places more of a limit on accuracy than the digital side does.

The terms noise floor, signal to noise ratio (or S/N ratio), and dynamic range all refer to the size of the noise (hum, hiss, digital noise or other artifacts) to size the audio signal. The maximum dynamic range of 16 bit CD quality sound is limited to 96db (6db * 16bits). The maximum dynamic range of 24 bit audio is 6db * 24 bits or 144db. The problem is that with todays technology it is pretty difficult to manage an audio circuit with better than 120db S/N ratio. In fact even the best converter chips usualy can't manage much more than about 120db S/N ratio. So what happens is that 4 bits of that number wind up being noise and your really only getting about 20 bits worth of S/N ratio. To keep this in perspective, the threshold of hearing is 0 db and the pain threshold is around 130db.

Some do-it-all units also have MIDI for controlling keyboards, effects units, or anything that that has a MIDI interface. A single MIDI connector can handl 16 channels of MIDI data and some units can have more than one of these.
 
Thanks....

Thanks that all made pretty clear sense...when you said:
"In fact even the best converter chips usualy can't manage much more than about 120db S/N ratio. So what happens is that 4 bits of that number wind up being noise and your really only getting about 20 bits worth of S/N ratio."
If that's so, why are basically all cards these days 24 bit and not 20 bit. And that would mean most of these cards out today probably don't even have good enough converters but use 24 bit anyways, Is this true?

When you siad"
"There are several standards that have developed over time for sending digital audio between various pieces of equipment. The terms lightpipe, ADAT, S/PDIF are used to refer to one of the standards for digital I/O." Do you know which standard is the best or most recommended?

Lastly you were talking about midi connectors...I have a Digital reverb ( effects processor ), and I want to also get a keyboard. Does everything these days basically work in didgital, or in other words would I not need any midi unless I had any old midi external machines?

I'm still not sure abouit Channels or when they say something is real time, isn't everything real time nowadays...(confused)
 
I was afraid you'd ask that. I was careful to say you only get about 20 bits worth of S/N ratio, which is not necessarily the same as only having a 20 bit converter. You see, you may get 4 bits of noise, but it's also possible to get 4 bits of extra signal mixed with that noise. If that were the case and you could recorded a signal that was small enough to fit in those 4 bits, when you played it back you'd hear both noise and the signal you recorded. So you are actually getting something, although not much. What determines if you actually get all 24 bits is whether the converter "maintains linearity" down to all those 24 bits. Which is basically a way of saying it accurately converts audio to numbers through all 24 bits.

However, there is another answer to your question. Computers tend to represent numbers in 8 bit chunks so even if you had a 20 bit converter the computer would wind up representing that number as 24 bits anyway because because it would want to process and store it as three 8-bit bytes. Essentially, once you make the leap beyond 16 bits you might was well go to 24 bits anyway since it free.

Digital standards is one of my weak spots. Someone else more familiar with the ins and outs might better answer your question. However, for what its worth, I get the impression most people choose a standard based on what they want to connect to, not what is best. In fact I've never heard of a any of them refered to as "the best" format.

MIDI is a way of sending messages such as note-on, note-off, change-program, change volume, etc to a device. It's used on keyboards to turn them into a very sophisticated player piano. No audio information is used in MIDI. My effects box has a MIDI in which allows me to change effects programs. wet/dry mix, and other parameters in real time while the sequencer plays keyboards or audio. If you work totaly in real time, recording live instruments you may have no need for MIDI, however it's a nice feature to have for effects and triggering keyboard stuff even if you do most of your work with live instruments.

The simplest explanation of channels is that each channel of A/D is one more track you can record at a time. Two channels can record stereo, eight can record eight tracks at once. On playback having two channels allows stereo playback, six would allow a six speaker surround playback. This only refers to recording and playback. Inside your computer the number of tracks is limited by your software. If your software and system allowed you could have 100 tracks, mix them down to stereo and output it on two channels.

As to your last question the answer is no, everythiing is not always real time. In fact nothing is. Computers only really do one thing at a time. They just do it fast enough that it appears to be doing lots of things at once. But we probably don't want to go there. The way you originally refered to real time was probably in the context of audio playback. You see there's a delay from the time the audio driver gets audio data to when it plays back. This delay can be long enough to be audible. To compensate the computer can send the data early making is seem like the delay doesn't really exist... until you move a fader. You see it can send audio early, but it can't do anything with a fader move till it sees it. So fader moves don't always appear to be in real time. This delay is often refered to as "latency," and is often affected by the size of audio driver buffers, the trick is to lower latency to a low enough value that you no longer can percieve it.
 
mucho's gracias...

"This only refers to recording and playback." I'm not really sure what you mean by playback, playing what's just been recorded?
How many channels would I need or want, if I was only using a MIC to record vocals...? One right...? or I only need channels to record multiple instruments...?
 
Thanks SPINSTERWUN.

PRiZ, yes playback what you just recorded, as in hitting the play button on a tape recorder, only in the case of a Digital Audio Workstation its just a grahipc of a play button on the screen of your computer. You need one chanel for each thing you want to record *at the same time*. If all you want to record is vocals then you may only need one channel. If you wanted to record one mic on an acoustic guitar and another on your voice as you sing you'd need two. If you want to put four mics on a drum set you'd need four chanels.
 
You must log in or register to reply here.
Back
Top