• Get involved.
    We want your input!
    Apply for Membership and join the conversations about everything related to broadcasting.

    After we receive your registration, a moderator will review it. After your registration is approved, you will be permitted to post.
    If you use a disposable or false email address, your registration will be rejected.

    After your membership is approved, please take a minute to tell us a little bit about yourself.
    https://www.radiodiscussions.com/forums/introduce-yourself.1088/

    Thanks in advance and have fun!
    RadioDiscussions Administrators

Concept of PCM/Wave sampling rate

C

ChiefOperator

Folks,

I have a question about the concept of the sampling rate of PCM/wave. I understand that the sampling rate is related to the frequency response, but does it affect basic sonic quality?

For example, lets say I rip a CD at the standard 16, 44.1 (PCM). Now, let's say I rip the same CD at the lower sampling rate of 32k. I understand the CD ripped at 44.1 will have a frequency response of around 20k and the one ripped at 32 will have a response of about 16k. Question: Will the two CDs have the exact same audio quality at all frequencies BELOW 16k? Or will this lower rate degrade the audio quality?

If the two are similiar, it seems that an FM station could save disc space and not sacrifice quality by using 32, not 44.1. Likewise, an AM station would not sacrifice quality by sampling at 22, not 44.1

Am I correct??

Thanks....
 
ChiefOperator said:
Folks,

I have a question about the concept of the sampling rate of PCM/wave. I understand that the sampling rate is related to the frequency response, but does it affect basic sonic quality?

For example, lets say I rip a CD at the standard 16, 44.1 (PCM). Now, let's say I rip the same CD at the lower sampling rate of 32k. I understand the CD ripped at 44.1 will have a frequency response of around 20k and the one ripped at 32 will have a response of about 16k. Question: Will the two CDs have the exact same audio quality at all frequencies BELOW 16k? Or will this lower rate degrade the audio quality?

If the two are similiar, it seems that an FM station could save disc space and not sacrifice quality by using 32, not 44.1. Likewise, an AM station would not sacrifice quality by sampling at 22, not 44.1

Am I correct??

Thanks....
Click to expand...

Ripping a higher quality audio file to a lower sampling rate will affect the overall freq.....not much but the problem in ripping wasnt quality, it was HD size.....compression like MPEG 2 helped saved much disk space when HDs were expensive compared to lower sampling rates...NOW? Why bother.............You dont want to degrade the audio..so thats why CDs are ripped at 16bit, 44.1 which is their normal recording format......ripping it at 22kHz will cause some issues with the audio....theoretically, it shouldnt but it does; MPEG 2 kept the high end but the low and midrange would suffer....once you digitzed audio, its best to keep it at the same rate, etc all the way through...besides HDs are CHEAP now..........so why bother throwing away quality? If a decent digital system comes along, you want your audio the best it can be and not have to redo it all over again....
 
Okay, thanks, that makes sense.

How about an live, original recording of instruments or vocals (not ripping a CD). Will one sampled at 44.1k and one sampled at 32k sound identical at all frequencies below 16k? There is really no practical application for what I'm asking; I simply started thinking about these concepts.....

Thanks again....
 
The higher sampling rate will always give you a higher quality sample. The minimum number of samples required to get the frequency of a sound right is 2 samples. But, those two samples will not show you the shape of the wave, and may not get the amplitude right. A higher sampling rate will help you better define the shape of the wave, and better represent the original amplitude.

Sampling rate usually creates two effects. First, lower frequency sounds will be better represented than higher frequency sounds because they will benefit from more samples to create the waveform. Secondly, lower amplitude sounds may be better represented than higher amplitude sounds. It's entirely possible that audio may spike between samples, and that spike may be missed.
 
SirRoxalot said:
The higher sampling rate will always give you a higher quality sample. The minimum number of samples required to get the frequency of a sound right is 2 samples. But, those two samples will not show you the shape of the wave, and may not get the amplitude right. A higher sampling rate will help you better define the shape of the wave, and better represent the original amplitude.

Sampling rate usually creates two effects. First, lower frequency sounds will be better represented than higher frequency sounds because they will benefit from more samples to create the waveform. Secondly, lower amplitude sounds may be better represented than higher amplitude sounds. It's entirely possible that audio may spike between samples, and that spike may be missed.
Click to expand...

You are 100% correct! Though in theory you only need two samples to reconstruct audio, this assumes that the (low pass) reconstruction filter is perfect-that it has zero amplitude and zero phase anomalies right to the cutoff frequency then drops to zero amplitude just past cutoff. Such a filter only exists in books. Real filters have ripple in the amplitude response and also have phase shift as you get closer to cutoff. They also don't fall off to infinity past cutoff. Thgough digiital fliters are much better then the older analog ones, they are still far from perfect. It's much better to get the abnormalities up past our hearing (using our ears as a low pass filter). This is why many report that sampling at 48 kHz sounds even better then sampling at 44.1, when conventional wisdom says it shouldn't make a difference. What's happening is that the audio within our hearing passband is more linear because it's not in the 'bite' of the filter.
 
I am so happy to see wisdom being propogated here. I remember this exact issue and arguments about it back in the 80's.
I argued that two measurements could only assume sine waves as the limit was reached, and that the amplitude became
ambiguous. I was usually outnumbered by people blinded by "digital quality", who could not even grasp the concept.

At the upper frequency limit, the info may be present, but any "rate of change" in the info cannot be related, as that would require more points of mesaurement to define other than a sine wave.

I wish understanding of this were more widespread.
 
Sampling Simulator

A lot of this goes back to Nyquist's Theorem for transmission of multiplexed telegraph signals developed in 1927, and published in "Certain topics in telegraph transmission theory" in 1928. Note that the original Nyquist Theorem had nothing to do with sound or sampling of sound. His theory was cited by Claude E. Shannon in 1949 in his breakthrough work on sampling audio called "Communication in the presence of noise".

For an interesting simulation of sampling and its limitations, check out http://www.vias.org/simulations/simusoft_nykvist.html. It's a little program that you can download to display the results of sampling different wave forms at different rates. It's really eye-opening.
 
SirRoxalot, thanks for the program. You're right--it is certainly eye opening and informative.

All, thanks for clarifying this subject. Also, if you haven't seen the program that SirRoxalots mentioned, take a look.....it's definitely worth playing around with.
 
Status
This thread has been closed due to inactivity. You can create a new thread to discuss this topic.


Back
Top Bottom