 |
|
You are not logged in! F.A.Q Log in Register |
 |
|
 |
 (nobody)...and 323 guests Last 5 registered Oplandisks nothingstar N_loop yipe foxtrotromeo Browse members... |
 |
Members 8025Messages 2614155 Today 0 Topics 127544 |
|
|||||||||
| |||||||||
|
|
Capturing the Sound Let’s talk about all the business that goes on before a sound ever gets to your computer’s memory. Sound in the air is continuously changing, and when it gets converted to an electrical signal the changes are still continuous. Your computer, however, can only store numbers using a limited number of digits or precision. Continuously varying sound is called an analog signal. Once the computer grabs the sound, it doesn’t have enough precision to store all the information about the sound in order to perfectly reproduce it. What the computer has stored is called a digital signal representation. Your sound card captures information about an analog sound signal by measuring its intensity at a given instant. This corresponds to one single point on the waveforms we’ve been looking at. In order to capture an entire waveform, the measurement process must be repeated at a high rate, usually thousands of times a second. Since the hardware has limited speed and memory capacity, there are only so many points it can capture. Any information between those points is lost forever. This process of capturing the sound in small intervals is called sampling. To play back a sound, we just reverse the process and convert the digital samples back to an analog signal. Of course, the new signal will probably retain some of the staircase effect, so the reproduction won’t be perfect |
||||||||
|
|
|||||||||
| |||||||||
|
|
Sampling Rate So just how many points do we need? If you look at audio specs much, you’ve seen CD sampling rates of 44.1 kHz, or 44,100 samples per second. That’s a lot of points! A well-known signal processing theorem (Nyquist Theorem) says that to accurately reproduce a signal, you have to sample at a rate at least twice the highest frequency component in the signal. So the CD sampling rate of 44.1 kHz will capture frequencies up to 22 kHz (a little past the 20 kHz spec that most audio components try to achieve). Of course, most personal computers have trouble doing something a few thousand times a second, let alone tens of thousands, so high-end stereo quality is usually reserved for the fastest machines and most expensive sound cards. Fortunately, many applications don’t require a wide frequency range to get the reproduced sound "good enough". Human speech, for example, contains some frequencies in the 10 kHz range (needing a sampling rate of 20kHz), but even at 4 kHz (8 kHz sampling), voice is perfectly understandable. Telephone systems rely on that fact; if they had to handle hi-fi audio, few people could afford the price of a phone service. You might be wondering what happens if you don’t sample at a high enough frequency. Well, what you get is something called aliasing. This sinister sounding term just means that since the sample points aren’t close enough together, it looks as though you sampled a lower frequency that really wasn’t part of the original signal. Alias frequencies are like ghosts - poltergeists really - you can’t see them but they make a lot of noise. So by sampling at too low a rate, not only do you miss some of the high frequencies, some of them get thrown back into the mix as unwanted guests at lower frequencies. They are audible as background noise and distortion. |
||||||||
|
|
|||||||||
| |||||||||
|
|
Sampling Resolution Along with the sampling rate, the other key factor in audio sampling is resolution. Personal computers are designed to work with chunks of data in 8-bit bytes. Because of that, it’s convenient for sound cards to use a single byte to represent one sound sample. But because the original sound is a continuous analog signal with an infinite range of loudness levels possible, something’s got to give. After all, the 256 possible loudness levels that an 8-bit byte can represent are a lot less than infinite, so you end up with the staircase effect you already know about. This step effect means that for the time intervals of one sample, we’re assuming that the waveform was flat, instead of whatever it might have been doing in reality. Squared off waves are legitimate waveforms, but if those sharp edges existed in nature, they would be produced by some pretty nasty high frequencies in the sound. So when we play back the sound we’ll be creating those new frequencies, and they’ll sound like background noise. Noise is the main effect of using a low number of bits to represent sound samples. In audio terminology, we talk about the signal to noise ratio, or SNR of sound equipment. This is a number you get when you divide the maximum sound level bu the noise level. You want that number to be as high as possible indicating that the noise level is very small. The SNR is measured in united of decibels (dB). Decibels are like the Richter scale for measuring earthquake intensity - each step represents a much larger increase than the last. Below is a table that shows how the number of bits used to store a sample relates to the signal to noise ratio. This table is based on the approximation that each bit is worth 6 dB. Number of Bits : 4 6 8 10 12 14 16 SNR (db) : 24 36 48 60 72 84 96 Don’t forget that decibels do not increase at a constant rate. The difference between 8 and 16 bits is not just a simple doubling of |
||||||||
|
|
|||||||||
| |||||||||
|
|
are you trying to talk Lee into sleeping with you? |
||||||||
|
|
|||||||||
| |||||||||
|
|
HI! |
||||||||
|
|
|||||||||
| |||||||||
|
|
sucking my knob would be just fine... |
||||||||
|
|
|||||||||
| |||||||||
|
|
Point in this cut and paste jobby is what? The implication you made was that I sampled specifically from autechres work and were particularly negative about it. I was simply stating the fact that the track was completely sequenced out of individual drum hits and not taken from another specific source. That's all. |
||||||||
|
|
|||||||||
| |||||||||
|
|
no problem mate...but it's too familiar that's why i commented like that. what did you use? |
||||||||
|
|
|||||||||
| |||||||||
|
|
If you are implying that ecnadniarb (I don't think I know him well enough to call him "Lee") sampled the autechre song in question, you're wrong. He used a similar technique, but you can cleary hear that it was a unique creation, not an audio sample. |
||||||||
|
|
|||||||||
| |||||||||
|
|
Cubase with Kontakt and about 20 drum samples. |
||||||||
|
|
|||||||||
| |||||||||
|
|
proove me that and i'll believe you. |
||||||||
|
|
|||||||||
| |||||||||
|
|
I don't care if you believe me. |
||||||||
|
|
|||||||||
| |||||||||
|
|
Good threads have purpose. This thread has a frightening lack thereof. |
||||||||
|
|
|||||||||
| |||||||||
|
|
sorry? |
||||||||
|
|
|||||||||
| |||||||||
|
|
Is your face burning with shame? |
||||||||
|
|
|||||||||
| |||||||||
|
|
no...should it be cos i'm doubting his beatz? |
||||||||
|
|
|||||||||
| |||||||||
|
|
Well I found all this interesting regardless of how well it fits in the whole "Dide Birandunce sample teh ortectors or not, lol" argument. I think this is a good time for me mention the bizarre sounds that stomachs make... It's like a videogame sound or something, actually sort of similar to that one boss in ghouls and ghosts (original) that is like a big rock you stand on and it shoots up worms and you have to kill the worm nests on it's body... I can't really think of any mechanism, natural or artificial, that could possibly make this sound. Maybe an octopus could make it. It kind of goes like this: baruwabaruwabaruwabaruwabaruwa, but that doesn't really describe it well. Anyway I think music scientists should study that stomach to make a cool musical instrument that emulates it. |
||||||||
|
|
|||||||||
| |||||||||
|
|
I've sampled a snare hit from Lexaunculpt, does that make everyone sad? |
||||||||
|
|
|||||||||
|
Messageboard index
|
|||||||||
