48kHz vs. 44.1kHz Sample Rate @ 24 Bits

B

Bob's Mods

New member
Has anyone noticed a sonic difference between 48kHz and 44.1kHz mixes?
I notice a slightly warmer sonic texture and a slight improvement in the 48kHz track mix definition over the 44.1kHz mix. There is a 9% increase in samples at 48kHz over 44.1kHz hense a 9% improvement in track accuracy.
I know some on this newgroup have stated they hear no difference between the two sample rates but I am hearing a difference. Its not way huge but its there. It seems to have a more tape-ish feel to me than 44.1

Bob the Mod Guy
 
I think that just getting way up away from any nyquist action, at least with the anti alias filters in a DAW helps, but I NEVER Sample rate convert on my own. That makes things collapse in the end MUCH worse than just staying at 44.1 IMO.

I mix to tape, and I normally run sessions at 88.2k, but I just mixed some stuff that was tracked at 48, and just doing the roughs, I printed straight back into PT (i mix on a console with all outboard stuff).
48k is like a bait and switch sometimes. It sounds cool until you want to actually put it on something like a CD, then you have to bounce to 44.1 and obviously the SRC.... SRC always seems to suck the life out of a mix...

At least at 44.1 I know what I am getting myself into....

Like I say though, I rarely even worry about it because I mix on a console and mix to two track analog. If I dont, I let the mastering people deal with the SRC, because the SRC in PT seems to suck some of the hard earned life out of the two track...

I think any sample rate can sound good. The more I think about it, the more i DONT care about what sample rate something is tracked at...
 
As far as the Nyquist frequency goes, the difference is less than 1 musical step. Hell, going from 44.1 to 88.2 only gets you an octave.
 
I'm with Joel. Unless you have top shelf conversion available to you, the cost of downsampling can be greater than any gross profit that the original 48 might give you beforehand.

G.
 
Last edited:
I'd say it depends on how much mixing's done. If I listen to a stereo master at 44.1 and then the same at 48, I can tell maybe a miniscule difference...little enough that it might even be my imagination. However, I can tell a more noticeable difference if I recorded a song in 48 to begin with as opposed to 44.1, especially if the mix required a lot of processing, despite the ultimate goal being 44.1, 16-bit for CD. Any higher sampling rate will give you that much more flexibility to mash and mangle your waves, should such be necessary. It's just another stat in the signal path, and the higher you can keep the quality of any step on the signal path, the better, I say. I generally record all my bands at 24-bit, 96 KHz (yes, even the metal bands), just on the offchance that they may want to make a DVDaudio mix or something. I mean, why not, with computer memory as cheap as it is nowadays.
 
$a1Ty said:
if you want to record about 44.1 for cd's use 88.2 cos then it converts better
Click to expand...

Interesting point for digital resampling. I'm not a math expert, but if the target rate was 44.1, something like this happens:

88.2 / 44.1 = 2

48 / 44.1 = 1.08843537414965986394557823129252

(or thereabouts)

96 / 44.1 = 2.17687074829931972789115646258503

(or thereabouts)

But if the target rate was a DVD at 96k,

96 / 48 = 2

96 / 44.1 = 2.17687074829931972789115646258503


I guess a way around it would be to record at the target rate, or a multiple or factor of the target rate that gives you a whole number, or else ship the thing through a D/A converter and re-record to the target format.


I think I've read some comments by Dan Lavry somewhere that sampling at more than twice the rate of the frequency that you intend to capture works fine and everything (Nyquist wasn't wrong, which is kind of amazing considering that his sampling theory predates the telephone), but there are problems with getting the low pass filter and anti-aliasing stuff to work for you if the sample rate (eg. 44.1 or 48) doesn't give you that much more room before Nyquist. And that these problems generally start to go away nicely if the sample rate is above around 60 kHz. Something like that anyway. Makes you wonder what 196k is all about if it's still a 24 bit PCM format. The dynamics would be the bottleneck, and the sample rate would just help to chew up system resources.

If the $149.99 DSD DAW ever comes along, maybe this will all become obsolete.


sl
 
Joel Hamilton said:
I think any sample rate can sound good. The more I think about it, the more i DONT care about what sample rate something is tracked at...
Click to expand...

I'm giving you a humble green rep point for this statement.


sl
 
If you record at 88.2, is it really as simple as dropping every other sample to get down to CD? I kinda thought there might be more to it than that. It hardly even sounds like 'dithering'. :confused:
 
Bob's Mods said:
Has anyone noticed a sonic difference between 48kHz and 44.1kHz mixes?
I notice a slightly warmer sonic texture and a slight improvement in the 48kHz track mix definition over the 44.1kHz mix. There is a 9% increase in samples at 48kHz over 44.1kHz hense a 9% improvement in track accuracy.
I know some on this newgroup have stated they hear no difference between the two sample rates but I am hearing a difference. Its not way huge but its there. It seems to have a more tape-ish feel to me than 44.1

Bob the Mod Guy
Click to expand...

What interface/equipment do you have?
 
Dithering refers to reducing the bit depth (or word length) which is how BIG each sample is, usually 16, 20, or 24-bits. We are talking about Sample-rate-reduction (SRC), which is reducing how many samples per second there are. Two different processes, but often done at the same time when mixing down/converting file formats.

PS - It took me a few days to figure out why you had a search box by your username, I thought it was a new BBS feature! :D

noisedude said:
If you record at 88.2, is it really as simple as dropping every other sample to get down to CD? I kinda thought there might be more to it than that. It hardly even sounds like 'dithering'. :confused:
Click to expand...
 
gordone said:
Dithering refers to reducing the bit depth (or word length) which is how BIG each sample is, usually 16, 20, or 24-bits. We are talking about Sample-rate-reduction (SRC), which is reducing how many samples per second there are. Two different processes, but often done at the same time when mixing down/converting file formats.
Click to expand...
Now that you say that, I realise I knew that, of course. :rolleyes: But my question still stands - would the SRC component literally consist of just dropping every other sample?

Oh and my avatar ........ yeah, it's conned quite a few people that way! Sorry! :)

Nik
 
noisedude said:
But my question still stands - would the SRC component literally consist of just dropping every other sample?
Nik
Click to expand...
No. The algorhythm upsamples and then brings it down to the target. It is not just throwing out every other one. That is a myth.
 
Farview said:
No. The algorhythm upsamples and then brings it down to the target. It is not just throwing out every other one. That is a myth.
Click to expand...
That's what I thought. So the 88.2 thing is a myth, yes? :confused:
 
its a matter of connect the dots... if you had a pair of 44.1 samples with values 512 and 128, or in 88.2 with 4 samples: 512, 384, 256, and 128, it would seem likely that the transition made by the ADC would be more accurate.... now which pair of the numbers gets selected is another matter...

since Nyquist theory suggests any number of sine waves can generate/represent a complex signel, then one would summize more points to create more sine waves would result in more accuracy rather than less... I beleive Dan L's point in his article is that the ADC/DAC are constrained in their ability to convert to the desired word size and at around 60K (at the time of his article) you hit the limit of orderly data vs. conversion handling. with the presumption that 96K was probably as far as people should go.
 
gullfo said:
its a matter of connect the dots... if you had a pair of 44.1 samples with values 512 and 128, or in 88.2 with 4 samples: 512, 384, 256, and 128, it would seem likely that the transition made by the ADC would be more accurate.... now which pair of the numbers gets selected is another matter...

since Nyquist theory suggests any number of sine waves can generate/represent a complex signel, then one would summize more points to create more sine waves would result in more accuracy rather than less... I beleive Dan L's point in his article is that the ADC/DAC are constrained in their ability to convert to the desired word size and at around 60K (at the time of his article) you hit the limit of orderly data vs. conversion handling. with the presumption that 96K was probably as far as people should go.
Click to expand...

The white paper is here:

http://www.lavryengineering.com/documents/Sampling_Theory.pdf

I don't think the point is so much that the A/D/A's are constrained at 60k, and he's not mentioning word size.

The paper talks about how sampling at more than twice the frequency of interest can reproduce the waveform accurately, not approximately. This is the basis of Nyquist's sampling theory, which couldn't actually be proven until around 40 years later. There's a bit more to it than that, in that the 44.1 sample rate of CDs can still leave deficiencies in the higher frequencies approaching 20k, because of the filters needed to reject the content above the Nyquist limit from being audible and that these problems are gone above around 60k.

It does seem intuitive that more samples would lead to greater accuracy, but where a faster system would be constrained is in speed vs. accuracy, processing requirements and storage space. I take it that he's talking about 60k as being a point of diminishing returns. Because of the arguments presented, I'm inclined to think that an accurate clock at 44.1 would sound better than a clock that jitters at any sample rate.



sl
 
"44.1 sample rate of CDs can still leave deficiencies in the higher frequencies approaching 20k, because of the filters needed to reject the content above the Nyquist limit from being audible and that these problems are gone above around 60k. "

This is what I was talking about in the difference between 44.1 and 48. The filtering is slightly different, and to say that this tiny difference is not noticeable is like saying that something with a top end response of 10k is no different than one with a response of 12.5k... That can make a BIG difference, but again: The gear used and the operator of that gear makes a much bigger difference. I have done stuff at 44.1 that sounds much better than certain projects I have heard that were done at 96k or 48k or 88.2 or whatever.

Anti aliasing filters that are nice sound better than crap filters, the rest is data... (gross oversimplification).
 
Farview said:
No. The algorhythm upsamples and then brings it down to the target. It is not just throwing out every other one. That is a myth.
Click to expand...

To basically do an accurate conversion following this line of reasoning, the SRC would have to bring 48k up 160 times to bring it down to 44.1 evenly. Converters do exist that can handle this kind of SRC, but I'm not sure that my Delta 44 is up to this task. 88.2 to 44.1 does seem like less of a challenge.


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