should i use 48khz for my tunes?
 

i've allways done my music in 44khz 16bit, and have never questioned it. But is it better to make them in 48khz nstead? btw,my technical understanding of these things is minimal- i usualy know enough to get by as far as quality, mastering type stuff is concerned. i use cubase sx3

I would say no. I don't think you will gain much by going to 48khz over 44.1 Where you will gain quite a bit is by going from 16bit to 24bit (or 32) - if your audio card can handle it. I could actually hear the difference in my audio when I recorded my external instruments/vox in 24bit vs 16bit !

If you are going to go higher than 44.1, don't bother tracking at 48 because there is going to be little difference in the frequency response and, if anything, your audio is going to be compromised when you do bounce it down to 44.1 for mastering.
If you really want to record at a higher sample rate, try 88.2. Even then, you aren't going to notice much ov a difference and its going to take up twice as much room on your hard drive.
The only reason a sample rate ov 48 was ever popular is because there was a time when people were still tracking with A.D.A.T. and other digital 8-tracks whose host format was 48.
I'd recommend getting a book by Bob Katz called "Mastering Audio" or something to that effect.
Here is the book I'm referring to, its a wealth ov valuable information on digital audio: http://books.google.com/books?id=EBC...with-thumbnail

You can't hear the difference of the sampling... but you can hear aliasing phenomena. So the highest you sample, less aliasing you hear. 48 IS better than 41 and if you downgrade to 41 with a good aliasing suppressor you'll gain some quality. No one can hear over 22 KHz so for the well known theorem you have to use double KHz in sampling. Every KHz more will make you a little more happy.
For acoustic purpouse I always use 48KHz and I'm pretty happy about it when downgrading to 44100... for other kind of recoring I always use 44100 and 24 bit as mentioned before.
Regards, Lorenzo

48kHz is still a standard in the broadcast/media industry...

I have the option of running 88,2 that some say is better than 96 kHz if you plan to downmix to CD quality (44,1 kHz, 16 bit) because half of 88,2 is 44,1 it will be devided by a nice even "2" instead of the more uneven devision of "2,176870748299319727891156462585". This will in your case mean that you devide by "1,0884353741496598639455782312925" instead of just doing it in 44,1 kHz from the start. It makes sence to me but still this is pure theory and i really don´t know if you will be able to hear the diffenrence. Still Doc is right when saying go to 24 or 32 bits instead of 16. This will give you a lot more than 48 kHz will...

Yes this is the theory... but I can assure you that 96 KHz is far better than 88.2, you practically have no aliasing cos there are only a few instruments with harmonics at those frequencies... and probably the aliasing would fall over the 22 KHz so you can't hear it.

so what exactly are you talking about ? Samples ? Or recording your full mixdown ?

when i'm using samples (what is very unusually for me, but sometimes it's just necessary) 44khz 16bit is enough for me, in the most cases - but when it goes to recording the full mixdown or a live-set on stage 96Khz 24bit is the better choice imo.

48Khz will lead no real benefit. You'd be far better off using 24bit 44.1khz....

If you have a cut down version of cubaseor something....it'll probably let you use 24bit, 48khz......but dont get obsessed by these figures. Whats really important is that you have decent quality converters. If your on a real budget then I highly recommend either the Emu 0404 or Audiophile 24/96 sound cards for audio recording.

DS

you can't say this, bit and khz will affect different sound "components":
more KHz = less aliasing
more dB = less digital noise (related to the amp of the sound)

so, you can sample at 44100 but you should put an anti aliasing (low pass filter) at the half freq (in this case 22 KHz) but not after the Digital conversion of the sound... so we should ask our audio device manufacturer if it has a filter related to the sampling frequency, if not you'll have aliasing.
The higher (in this case 48 ) KHz we will use the less liasing we will have, than converting, every software I know has an antialiasing plugin to digitally correct the problem when downgrading to 44100 ... 48KHz downgraded to 44100 after a aliasing filter is ALWAYS better than the straight 44100 recording.

I'm not saying you are right or wrong here but can you provide some evidence to support this? Perhaps a link to an article or something along those lines?

Do you need the aliasing theory here or some test with software? cos in the first case there are plenty of page in the net that explain the aliasing phenomenon... (starting from here http://en.wikipedia.org/wiki/Aliasing), else, if you want to know if a software is good or not in removing aliasing when it downgrade the sample frequency well it's just a filter... I guess it's not a difficult algorithm to implement in any software so I trust i.e. motu if they say that their anti aliasing filter is working... maybe it's difficult to find and remove aliasing AFTER the downgrade cos you should find the aliasing noise... but this isn't our case.
regards, Lorenzo

It would seem as though I was wrong. I found an interesting quote from Bob Katz, a mastering god, who had the following to say about sample rates:

"Usually 48K sounds more "open" to me. And 96K sounds "purer and warmer" while retaining the openness of 48K. I highly recommend 48K over 44.

HOWEVER, yes, I've encountered situations where 44 sounds better than 48! If the material sounds better a little more closed in and less revealing... It all goes back to the choice of compromises and the nature of the original recording. I try to upsample to 96K for any digital processing, so at the end of that chain I get to listen to both the 96 and the 44 and I'd hate to admit it (lose my audiophile license)--- for SOME material, especially the hard rock, the 44.1 reduction takes away some of the ugliness or softens some of the distortion. Most times, though, I terribly miss the 96K.

This is a dilemma, should we use the more open, revealing format only to like the reduction at the end, or get our sound at 48K that we like and hope for the least "degradation" when reduced to 44K? I say, do what works best under the individual circumstances, and get to know your medium's limitations---or even advantages."

I guess I'll be giving 48kHz another try.

you weren't wrong... many people just can't hear 20 KHz and many times the aliasing phenomenon isn't that bad due to the fact that we usually don't produce with our instruments frequencies so high... I think that "open" "purer" "warmer" even used by Katz do mean NOTHING. I've seen a video of a meeting of great audio technician here in Italy where the speaker let them hear to some recordings done with the same microphone and different pres: the first list was in order so "this comes from a UAD, this one from a Neve, this from a Focusrite" and so on and people sitted there swear to recognize each sounds due to the "crisp" or the "flat" and other kind of BULL$HIT. Then he mixed the files and say "this is file 1, 10 seconds ago you saied that it is obviously an ... what is the pre?" EVERYONE GAVE DIFFERENT ANSWERS at this point and he started laughing. This prove many things but above the other there is only one... brains are "influenceable" (hope this is the correct word)...
So I think that it's not important the frequency or the bit rate... but just the fact we DO like or not the recording...

I agree partly. Higher bit rate is going to give you more headroom to work with but I'm starting to thing that sample rate is a subjective thing. I'm sure some sound cards and some types ov music sound better at different sample rates.

I found this example in the net... play a sine wave at 34.1 KHz and record it at 44.100... you'll hear an awful sound at exactly 10 KHz... this isn't subjective eh eh eh! (ok, 34.1 KHz is a really high freq but many harmonics can reach that HZ)

I use a ceative x-fi, but the only audio i record usualy is from the virus via usb. is this method still affected by convertors, am i oosing quality?

can i change to 24bit on the Cubase songs that i've done allready, aslong as they're just midi and virus tracks (no audio yet)?

il get hold of that bobkatz book as most this stuff is above me.

thanks

Logically, the higher the rates, the better the quality. The biggest determining factor is surely the quality of your A/D converter. I use an RME Multiface at 44.1 24 bit and cannot tell any difference between 44.1 or 96 running 16 tracks of audio. The difference between 24 and 16 bit is, however, discernible.

Yeah, honestly, I think a person is better off buying a good converter like an Apogee Rosetta or something rather than tracking at a really high sample rate.
With my own music, I've never been mixing and thought to myself, "ya know, this sounds great and all but I'm getting some aliasing from those cymbals thats really messing things up." Its just never happened to me.
Now this is something I notice in different synthesizers. But when it comes to recording, I've never really ran into that problem.

Yeah, if your fundamental is 34.1kHz. . . .
I see your point though.

this is just cos you can't hear difference between an aliasing artifact and digital noise... but digital noise produced by bit and by aliasing are different... maybe you should record some classical music then you'll hear that not only the bit rate is important but the frequency too. As saied before maybe this depends on the music stile you're recording.
Regards, Lorenzo

is there much more strain on the cpu when going from 44khz to 88 or 96khz?

Haven't tried it personally, but I believe so, yes. 88.2KHz would double CPU usage on plugins and stuff?

It's generally advised that increasing the bitrate (from 16 to 24bits) rather than the frequency samplerate will be more beneficial to achieving better sound quality whilst trying to make best use of CPU usage.

If you're doing pop- or dance-music or similar, then no-one's going to hear or admire the nuances of higher sample rates anyway!

My ears are fucked, so I can't tell either way, hoho, but I've heard this being said more than once: Some people CAN notice a difference between 48KHz and 96KHz, etc, but they can't often say which one they prefer.

Something else I thought of, when it comes to aliasing, Korg's workstations ALL interally use native 48KHz for all the samples and mixing, but no-one's ever said "hold on, this keyboard sounds shit when recording for CD (44.1KHz)".

And Korgs sound mighty fine too with their 48KHz! Mind you I'm not entirely sure that every rom sample is 48KHz on the mid rangers like Triton or M3. But I do know that my 44.1KHz sampled drums collection sound better on the Korg than some other 44.1KHz workstations. Definitely more top end 'air', whatever that means :) But after recording it at 44.1KHz, can I still claim it sounds better? Dunno, its different. And in the end its all synthetic, there is no 'right'.

Cheers,
B

They are nice and bright, "airy". Certainly I can't recall ever hearing any aliasing, but then again my ears are shot!

In the case of the Korg Trinity with PBS-TRI expansion (custom 8MB flash-ROM sampler) if you uploaded a 44.1KHz sample, the Trinity automatically converts it to 48KHz before storing it on the 8MB chip, as it can't deal with any other sample rate. Reason I know this, is that uploading samples to the Trinity and then downloading them back off of it the file sizes have increased as they have been converted to 48KHz. Or the 8MB chip couldn't load 8MB of 44.1KHz samples, as it would require more than 8MB to store them as 48KHz. So any samples I upload to the Trinity I record them natively at 48KHz before uploading them, to avoid at least one lossy conversion.

That's the big question: if you load a 44.1KHz sample into ram and play it, at what point does it get converted to 48KHz? While the samples are being scanned? Before sculpting by the filter and amp? Before FX?

And the whole problem of converting the sample rate, how is that possible without audible artifacts? Obviously it can be very good, but surely changing sampling rates several times in the signal chain must have an impact.

Cheers,
B

Higher sampling rates are used for DVD / TV media, stick to 44100 at 24 bit