highest fidelity

[ohthatguy]

which is more important to fidelity, bit rate ( 24 vs 16 ) or sample rate ( 44khz vs 48khz vs 96khz)? would a 24 bit 44khz recording sound better than a 16 bit 96 khz recording or vice versa? would a 24 bit 96 khz recording have an audibly better sound than 24 bit 44 khz one? thanks

 

[Farview]

It is easier for me to hear the difference between 16 and 24 bits, than 44.1 to 96K. Your milage may vary.

 

[ohthatguy]

thanks. am shopping daws and am wondering if 24 bit 44khz will do the trick, or if i need to go to a 24-96 machine. fidelity is more important than drive space. dvds sound much better than cds. for now i will prob only burn to cd but would like to have a recording for down the road when i can burn to something of higher quality. by the way if i save data to disk (cdrw) will it save in whatever sandard i recorded in (24 bit) or only in cd quality?

 

[earworm]

believe me, 48khz and 24 bit sounds CHRISTAL CLEAN !!

you don't need 96khz and 96 bits, how much do you know about dithering ? you might kinda fuck up your sound when downsampling everything when you wanna burn your stuff to CD.

if you're not an absolute pro, and your signal path isn't incredibly clean, then you shouldn't be recording at 96kz and 96 bits or whatever

i tried to hear differences between all those "settings" and its hard to heard with pretty cheap equipment ;.. i'm recording nice stuff now, all thanks to miking techniques and stuff, not thanx to 96 bits ;...

and my PC can handle the 48 khz and 24 bits perfectly,
it might start to protest at higher bitrates and stuff

..but be warned;.this is only MY 2 cents

cheers

 

[imissthewar]

Why not just record at 44.1/24? It seems that anything you gained by going to 48 is just going to get lost when you downsample.

 

[parkbirds]

"Why not just record at 44.1/24? It seems that anything you gained by going to 48 is just going to get lost when you downsample."

Um, not really. The information above 22kHz (Nyquist theorem) effects the timbre of the sound event. It's best to choose a high sampling frequency and stay there as long as possible. For example the effects or eq you might use during mixdown would be reacting to an inferior picture of the first sound event.

This sort of thing you learn in audio 101.

 

[imissthewar]

"Why not just record at 44.1/24? It seems that anything you gained by going to 48 is just going to get lost when you downsample."

Um, not really. The information above 22kHz (Nyquist theorem) effects the timbre of the sound event. It's best to choose a high sampling frequency and stay there as long as possible. For example the effects or eq you might use during mixdown would be reacting to an inferior picture of the first sound event.

This sort of thing you learn in audio 101.

Sorry to offend you with my stupid question.

 

[mshilarious]

Um, not really. The information above 22kHz (Nyquist theorem) effects the timbre of the sound event. It's best to choose a high sampling frequency and stay there as long as possible. For example the effects or eq you might use during mixdown would be reacting to an inferior picture of the first sound event.

This sort of thing you learn in audio 101.

What do you learn about anti-aliasing filters in audio 101?

 

[omtayslick]

What do you learn about anti-aliasing filters in audio 101?

Ha ha ha! There's more to audio 101 than meets the eye.

 

[Recording Chick]

I always record at 24bit 48 khz. There's 2 cents for ya, or mabey less.

 

[mshilarious]

I always record at 24bit 48 khz. There's 2 cents for ya, or mabey less.

Welcome, Recording Chick! That song on your site, was that a homerec? That sounds good even on the built-in PC speaker in the office. Very refreshing to hear a vocal that isn't compressed to death.

 

[cominginsecond]

The difference between 24/44.1 and 24/48 is unnoticeable to my ears. Therefore, to prevent the necessity of downsampling, I choose 24/44.1.

 

[Rev E]

Personally, I can hear a difference between 44.1 and 48 kHz (in the high end). I also can hear a difference between 44.1/48 and 88.2/96 kHz. I agree that there is a bigger sonic difference between 16 bits and 24 bits. BUT, there is an audible difference between 44.1 kHz (or 48 kHz) and 88.2 kHz (or 96 kHz). Plus, plugins really seem to shine at 88.2 and 96 kHz.

One of the things that doesn't get mentioned in the discussion of whether to use higher sampling rates (88.2 kHz or 96 kHz) is that the brick wall filter (that approximates to 1/2 the sample rate) is moved higher. As a result it affects less of the recorded signal. Even though human beings (except babies maybe) cannot hear at 22 kHz (which is 1/2 the sample rate of 44.1 kHz), signals up there can affect the audible range (lower than 18 or 19 kHz). So by recording at higher sample rates, you reduce the indirect effect that the filter will have on the signal you recorded. If you've got the hard drive space and plug in or computer power, you can't lose by recording at 88.2 and 96 kHz. I usually record at 88.2 kHz so that the sample rate conversion back to 44.1 is just dividing by 2... should be real easy for a machine to do that.

 

[mshilarious]

Even though human beings (except babies maybe) cannot hear at 22 kHz

As a parent of three (soon to be four), I can attest that a major source of adult hearing loss is due to those babies screaming at 22kHz

 

[boingoman]

One of the things that doesn't get mentioned in the discussion of whether to use higher sampling rates (88.2 kHz or 96 kHz) is that the brick wall filter (that approximates to 1/2 the sample rate) is moved higher. As a result it affects less of the recorded signal.....(clip)......So by recording at higher sample rates, you reduce the indirect effect that the filter will have on the signal you recorded.

Yeah, you are right, it seems. The filter actually isn't a brick wall, though. The signal is rolled off starting at 20k and is out by 22k. That much level dump in that small of a range causes phasing and ringing. Here is something I found about aliasing and why we use 44.1, and a link to the page.

ALIASING

if a 25 kHz waveform is sampled at 44.1 kHz (which has a Nyquist value of 22.05 kHz), the Nyquist rule is broken. 44 kHz - 25 kHz , results in a 19 kHz waveform which is heard as distortion. This is also known as 'foldover'


You can capture a 20 kHz simply by sampling at 40 kHz to satisfy Nyquist, plus 10% more for the guard band, plus 100 Hz to lock to video. 40 + 4(10%) + 100 Hz = 44.1.

"Now we have to build these anti-aliasing filters [low pass filter] to cleanly pass 20 kHz, but be out (-90 dB) by 22 kHz". So the extra 2k is the space needed to allow the filter to cut the signal to zero, and 44k avoids aliasing this data from 20-22k, the guard band.

" Truth is you can't dump that much level in that little frequency band without huge phase problems in the analog or digital domain. Therefore phase shift and high frequency ringing are common. 48K is smoother than 44k because of the extra headroom (10%). The problem with 48 k is it uses more media and is another standard"

Some digital stuff

 

[mshilarious]

Truth is you can't dump that much level in that little frequency band without huge phase problems in the analog or digital domain. Therefore phase shift and high frequency ringing are common. 48K is smoother than 44k because of the extra headroom (10%). The problem with 48 k is it uses more media and is another standard

I've love some input from a converter designer. Seems to be either filter is gonna be pretty steep, and I don't specifically know if 48kHz converter designs really use the extra little of bit range or not. My 48kHz converters are switchable to 44.1kHz. Does the filter change between the two settings? I don't know.

The higher sample rates (88.1 or 96) should allow a comparatively gentle filter, fourth order or so (I doubt -90dB at 48kHz is necessary, since there's gonna be very little info there to start with). What do they use? Again, I don't know. Do they use a steep filter to enable ultrasound response? Somehow I doubt it based upon the published frequency response specs, but it'd be interesting to know.

If the audio 101 chap is still around, this issue should be of interest.

 

[Farview]

If you are preserving the stuff above 22k with the higher sample rate, wouldn't that same steep filter cause the same problem when you down sampled?

 

[Rev E]

If you are preserving the stuff above 22k with the higher sample rate, wouldn't that same steep filter cause the same problem when you down sampled?

I don't think so... or at least my sense is that the effect isn't as great. Here's how I think of it. This may not be completely technically accurate, but it's how I understand it. If you record at higher sample rates initially each track is recorded free of the audible effect of the aliasing filter (or at least the effect has less effect on the audible signal).

If, instead, you recorded every track at lower sample rates, each track is affected in a greater way by the filter's effect. So when you're mixing down a track where all the files are affected by the filter, the end product is a combination of 'compromised' tracks.

IF you record at higher sample rates and wait till the end to convert the sample rate back down to 44.1 kHz, you record a signal that's less affected by the aliasing filters (i.e. more pure; more accurate; less distortion). When the aliasing filter does finally affect the signal, most of the signal is preserved and the track comes out with all of the highs in tact (at least the highs that we can hear or perceive).

P.S. this is not to say that everyone should record at higher sample rates. There's a lot more to it than that. You've got to take into account the increase space requirements, whether your computer can handle processing these files (if you're using a native program) and whether your soundcard is capable of the higher sampling rates.

 

[Recording Chick]

Welcome, Recording Chick! That song on your site, was that a homerec? That sounds good even on the built-in PC speaker in the office. Very refreshing to hear a vocal that isn't compressed to death.

Thanks, That was a home recording. Full Details and pic are on my studio page. We try to track vocals with no more than 2:1 ratio compression, can't remember the threshold. In that case my engineer really wanted to keep the vocal as close to the cieling as possible as it was a 16 bit 44.1 recording. I think 16 bit 44.1 can sound good. You just really need to keep the levels up there way above that ugly digital noise floor. Since then I've upgraded to the Roland 2480. It's always a relief to hear that from somebody that has ears.

 

[mshilarious]

It's always a relief to hear that from somebody that has ears.

It was kind of funny, since I'm at work I can't listen to dynamic music without disturbing the neighbors, so I had to manually compress by riding the Windows mixer fader . . .