Sampleing rate...?

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

PRiZ

New member
Does sampleing rate only matter if your sampleing from an external sampler, or is it for software too.
Is their a big difference between 48 and 98 or whatever the numbers are?
I'm thinking about the Lynx one card, but it only has a 48 sampleing rate, I'm using software so blah blah I'm curious...
 
Sampling rate is important. CD's are 44.1, but most modern software/hardware is 96...over twice as fast. The sample rate refers to how fast the "digital snapshots" of sound are taken. The faster, the better. Think of a digital movie. If the sample rate is low, you see flickering...if it's high it looks smooth. Same thing with sound. Right now we can all record at 96, but have to master at 44.1. We're all hoping that will change, but for now it hasn't. If you can afford it, get a card and software that will support 24-bit, 96Khz.
 
But don't forget that twice as many samples = double the amount of storage you need per minute per track. And the hard drive and data buss have to stream twice as much data in the same time. So you'll have lower track counts and fewer plug-ins you can use without dropouts.

Add to this the fact that a 96 kHz sampling rate really only adds the ability for the highest frequencies in your audio material to be up to about 48 kHz rather than the 24 kHz that a 48-kHz sampling rate allows... built where are you going to find microphones that have a frequency response greater than 20 kHz? And all the higher-order harmonics that are in the material are inaudible anyway...

Certainly 96 kHz is overkill for a home studio with inexpensive A/D converters and mics and preamps....
 
AlChuck hit it on the head with this statement:

>Certainly 96 kHz is overkill for a home studio with inexpensive A/D converters and mics and preamps....

The higher resolution of larger sample sizes and faster sampling rates will assist you in reducing artifacts when doing multiple sequential digital manipulations. Such manipulations include volume adjustments, and software effects. For just capture and write to CDR functions the quality of the A/D converters on the soundcard is more important than whether it supports the 24/96 versus the 16/48 thing. Your worries over the Lynx only supporting 48 KHz sampling rates are probably misplaced since this card reportedly has some very nice converters.

How clean can 16/44.1 get?

Just check out your commercial CD collection.
 
Aw, come one guys...I thought bigger was better! :p

I stand corrected. :D
 
hmmm....

c7sus..."I'd like to upgrade to 24bit but I can't really see the neccessity to go above 48k sample rates for my home rig......."
You make it sound like the only thing good about 24bit is it can handle 96khz...????
I'm not really sure about how where sampleing rates are used...?
If I burn a cd and take a sound off it, a "sample", and then throw that into the beat, is this sampleing rate not going to let it sound any better or what...? Or is this only when putting a sound from a peice of hardware, a hardware sampler for instance directly into the soundcard...? I could go on but I'd just be shooting blind, someone explain...?
 
bringing it up to the top...

.........................................................................................?
 
Not that wrong :)

24 bit versus 16 increases file size by 50%

96 kHz versus 48 kHz doubles the file size.

So the diff between 16/48 and 24/96 is 3 times rather than 4...
 
Damn!

I wish I didn't get an Athlon, to late to go back now.
I doesn't really affect me that much does it...it is an Athlon 850?
Computer guy at store said it was a much better deal, because it's cheaper and goes the same speed as a Pentium 933...
"really good at floating-point calcs so bigger numbers means more number crunching......"
I have no idea what floating-point calcs are,what's wrong with crunching, will my computer crunch enough...?
 
The AMD chip you have should be able to do a lot of processing (I'm still running on a PII 400). Just make sure the CPU and motherboard you're using are supported by the card manufacturer before you buy. I think most of the good cards should work with it by now, but best to make sure.

How much of a difference the bit depth (16/20/24) and sample rate (44.1/48/88.2/96) make is really dependent upon what you're going to do. I'll pretty much echo what Doc said: 16/44.1 can sound very good as a master as your CD collection should testify. But if you're going to be doing recording, mixing, adding effects, etc. all via software then 24 bits will help a LOT. I personally think 24/44.1 is fine for most home recording projects. You will get improvements with higher sample rates, but you start reaching the point of diminishing returns.

One final thing, it's my understanding that the noise floor of 24-bit audio is near the level of atomic noise. I think this is a big reason that converters will not give you much accuracy beyond 21 or 22 bits in real life. It's not necessarily that the converters aren't "good", it's that there are physical limitations in the real world. That doesn't mean the other bits are wasted, though. If you do a lot of processing on the audio then the bit depth will help maintain accuracy while performing the necessary mathematical calculations.
 
okay...

I'm not really sure about how? where? sampleing rates are used...?
If I burn a cd and take a sound off it, a "sample" (is this what they mean by sample), and then throw that into the beat, is this sampleing rate not going to let it sound any better or what...? Or is this only when sampleing a sound from a peice of hardware, a hardware sampler for instance, directly into the soundcard, so it gets transfered at 48khz a second because that's all the soundcard can transfer that sampled sound at...?
I asked this before but nobody seemed to understand the question...I'm asking what do they mean by sample? what sample is this? that gets sampled at 48khz? In hiphop, samples are considered peices of sound/music taken from it's origin, a sampled piano, or a sampled drum sound etc... does this make sense? If it does keep reading, if it doesn't tell me and I'll try and explain it better...

In this case, what good would 96khz be if a cd, the cleanest from of music I could sample from was only 48khz...Would I be able to make samples sound better, bring the quality up if I had 96khz...?
I have a feeling nobody's gonna have a clue what I'm talking about, so try and answer it as you interpret it.

Thanks!
 
Digital audio is basically a stream of numbers. Bit width is how big those numbers can be. Sample rate is how fast those numbers are fed throught the system. 44.1kHz = 44,100 samples per second. 96kHz = 96,000 samples per second. Each "sample" in this case means a single number the size of whatever the bit-width is. Not to be confused with the other usage of "sample" such as a cymbal crash or a bass riff. Those would consist of a lot of individual samples. Hope that helps more than it confuses :).
 
yeah...but answer this?

If something is already recorded, will this have any effect on it?
Is this only for real time, or recording...Could I play or record at 96, if the computer had time to configure before...? (or LOAD)
If sound is numbers, then could I make my mix to 96 and then just record it as long as I knows what it going to be doing ahead of time? This is all the same question just different degrees.
 
PRiZ, the main issue with this as related to loops or sample hits and such is that WAVs of different sampling rates and/or bit depth need to be matched together to play correctly. ACID does this automatically. Basically everything gets converted for you to data at the highest bit depth and sampling rate in the project. In other applications, you have to do this yourself. So if you had some hit that was recorded at 24 bit/96 kHz, and you had others at 16/44.1, you would have to convert one or the other to the same settings (using something like CoolEdit or Sound Forge 5) to use them together in the same project.

Note that converting data recorded at 16/44.1 to 24/96 data cannot improve the quality of the original sound -- basically it makes two copies of each sample and plays them twice as fast, and fills in the extra bits with zeroes. That is, any nuances you might have captured had you recorded at 24/96 in the first place are not going to appear like magic when you convert what you actually recored at 16/44.1.
 
PRiZ,

I realize I might not be answering your questions directly, so here's a bit more:

If something is already recorded, will this have any effect on it?
Click to expand...
Anything recorded at a particular bit depth and sampling rate is set that way. You can, however, convert it to other bit depths and sampling rates using most any audio editing software.

Is this only for real time, or recording...Could I play or record at 96, if the computer had time to configure before...? (or LOAD)
Click to expand...

It depends on your hardware and software. Typically what happens is your software lets you select from whatever bit depths and sampling rates are available to the hardware.

If sound is numbers, then could I make my mix to 96 and then just record it as long as I knows what it going to be doing ahead of time?
Click to expand...
Yes, if your hardware and software will support it, and if you have the disk space... remember though that a 96 kHz sampling rate will double your data storage requirements over 48 kHz, and unless you have very very high-end mics and A/D converters and are recording things with lots of high-end detail, you will gain no audible advantage.
 
Few final questions...

"It depends on your hardware and software. Typically what happens is your software lets you select from whatever bit depths and sampling rates are available to the hardware."
..."hardware" meaning soundcard...?

Here's just an idea, what about if I copied an exact replica of a sample (sound) I was using and then doubled them, but put one sample mayby a few milliseconds ahead of the other one.
wouldn't that do exactly what the 96 would do, but mayby even better?...and same for the vocals?
If this would work like this, it would mean it would not work if I place the sample to play parrallel to other sample, because the space where there's no sample wouldn't be filled in?

Sorry if this is too confusing, I'll re-explain it if it is...
If this method does work, I don't think it would help my hardrive...
What do you think...?
 
..."hardware" meaning soundcard...?
Click to expand...
Yes.
Here's just an idea, what about if I copied an exact replica of a sample (sound) I was using and then doubled them, but put one sample mayby a few milliseconds ahead of the other one. wouldn't that do exactly what the 96 would do, but mayby even better?...and same for the vocals? If this would work like this, it would mean it would not work if I place the sample to play parrallel to other sample, because the space where there's no sample wouldn't be filled in?
Click to expand...
What are you sniffing? :)

If you mean doubling a part by mixing one file with a copy that's offset by a couple of milliseconds, that is a technique people use to fatten parts up, but it has nothing to do with sampling rate, it won't change it or anything like that.

If that's not what you mean but you are literally talking about manually copying samples and pasting them in between each other to make a 96 kHz file out of a 48 kHz file... first, why would you want to do this (and what makes you think it would be better?) Second, even if you could do this, think about it. 48 kHz means there are 48,000 separate samples in each second of the file. Even if you were really good at copying and pasting, it would take you a second or so per operation to do this manually one by one. That translates to over 13 hours just to edit one second of a file.

Besides, CoolEdit and Sound Forge and most similar programs will do this for you.
 
OH GAWD!!!

you didn't seriously think I meant double every little millisecond sample...I meant the whole sample like a "loop" that's about two bars, copieing it and then putting another below and playing the copy about a millisecond after the first loop, so there's two loops playing as one...so where yo said:

"If you mean doubling a part by mixing one file with a copy that's offset by a couple of milliseconds, that is a technique people use to fatten parts up, but it has nothing to do with sampling rate, it won't change it or anything like that. "

Yes, it won't change the sampleing rate,...but it will fill in all those little spaces that the 48 isn't thorough enough to fill like a 96, right.

here's the sound files in 48.

2 2 2 2 2 2 2 2 2 2 etc...

here's the sound files in 96.

222222222222222 etc...

here's doubling the sounds in 48.

2 2 2 2 2 2 2 2 2 2 2 etc
2 2 2 2 2 2 2 2 2 2 2 etc

now see how those samples that normally had spaces in between
are being filled like the 96 is, that's how I meant sound the same.

would this work...?
 
To answer your last question simply: no.

The idea of 96k is not to "sound fuller" per se, it's to have more samples per second of the analog signal than lower rates and thus be a more accurate representation of the original sound (especially in the higher frequencies). Converting a 44.1k sample from a CD to 96k will not get you anything... it's only going to have the same quality as the 44.1 source unless you are going to be doing a lot of subsequent processing on the audio. You can't improve the accuracy over the original sample rate.

You should realize that the numbers are constantly changing because a waveform goes up and down. So your illustration with all 2's misses an important subtlety:

96k:
1.2.3.4.5.4.3.2.1.2.3.4.5

48K:
1...3...5...3...1...3...5

The 48k example doesn't approximate the wave with the same level of accuracy. Realize that this is an extremely simplified example, real-life wave forms are much more complex than this.

It sounds to me like 96k will probably be overkill for what you plan to do.
 
Last edited:
PRiZ,

Sorry I misunderstood what you were saying earlier...

pglewis is right.

A file recorded at 48 kHz plays back one sample every 1/48,000th of a second. There are no "gaps" in the data that you can slide a copy into to magically make it be 96 kHz.

Besides, even if what you were thinking was true, you would have to slide the copy 1/96,000th of a second, which is 0.00001 milliseconds, or about 1 thousandth of a millisecond... so a few milliseconds would be way too far...

Anyway, it seems that your grasp of digital audio is a bit fuzzy, from the nature of some of the questions you are asking... have you read anything that gives you an overview? There are a bunch of good primers available on this stuff. One I know of (though I'm not sure how good it is) is at Cakewalks website, called the Desktop Music Handbook:

http://www.cakewalk.com/Tips/Desktop.htm


-AlChuck
 
You must log in or register to reply here.
Back
Top