Skip to main content

Hi guys, I'm new to this site as well as the audio engineering scene. :biggrin: I had hope that I could come to this forum for help and guidance so first and foremost, I want to thank whomever gives me advice here and there.

So on to the question, if I wanted a snapshot of a sound wave would my sample rate have to be higher or lower? I'm guessing when recording it should be higher to eliminate the Nyquist Frequencies? :smile:

If I'm wrong, why? If the sound is sampled at a lower rate than that of the original sound, how does that affect the sample? :confused:

Thanks agaiN!

Comments

RemyRAD Sun, 07/31/2011 - 22:13

Most sampling occurs at the same rates as a CD, 16-bit, 44.1 kHz. You might find some more esoteric samples at higher sample rates such as 88.2, 96, 192 kHz and bit depths such as 24 bit, 32-bit float, etc.. If you sample something at a lower sample rate of say 22,500 Hz, 11,000 Hz, 8000 Hz frequency response in the higher frequencies is lost and high-end response is generally about 1/2 of what the sampling rate is. For instance, I have released spoken word recordings down converted to 11 kHz, 16 bit. This sounds much like AM radio without the static, With a high-frequency response cutting off at approximately 5,500 Hz. This is perfectly adequate for spoken word purposes. And even for music purposes, should you wonder how things might sound on AM radio, this more closely approximates AM radio.

So it's all in what you deem to be necessary for your purposes. I will say that most of us do not want low fidelity samples nor do we record at anything less than 44.1 kHz, 16 bit.

Sample rate & bit depth all affect performance.
Mx. Remy Ann David

Boswell Mon, 08/01/2011 - 04:21

The Nyquist frequency is always half the sample rate for base-band sampling, and your system needs to perform low-pass filtering so that everything below the Nyquist frequency is removed from the signal before it is sampled. In practice, with modern converters, this is done in two stages, but the end result is much the same.

There is no such thing as sampling "at a lower rate than that of the original sound", as acoustic sounds are analog in nature and do not have associated sampling rates. What you may mean is that the acoustic sounds can have high-frequency energy that is not captured by a recording device with a low bandwidth, but this is the case whether the device is analog (a tape recorder, for example) or digital (an audio interface attached to a computer). Think of an old reel-to-reel domestic tape recorder with a number of tape speeds. As you make recordings at successively reduced tape speed, as well as hearing various mechanical effects and increased tape hiss, the high-frequency content in the recorded signal drops off.

The CD standard sampling rate of 44.1KHz gives an effective bandwidth of d.c. - 20KHz, allowing 2.05 KHz at the top end for the slope of the low-pass filters before the associated Nyquist frequency of 22.05KHz.

You have to choose a sampling rate that is appropriate for the purpose that you need it for. Unless this is for instrumentation work or for content achival purposes, I suggest you stick with the industry-standard rates of 44.1KHz for CD quality or 48KHz for video compatibility.

Lougarcia Mon, 08/01/2011 - 08:27

Hi There!

If your gear can handle higher sampling rate like 92 kHz It is good! if not, try to record at 48 kHz it is still best. as long as you will not over processed the whole sample and
position all micings in recording correctly, use a good pre-amp, you will still achieve the good sound!

RemyRAD explained it very well!

Hope this one helps!