It's obvious most devices nowadays will go to 96k and for the mosy part 48k wich is half that is typical for video.
But the again half of 88k is 44 which ulimately is the end product for audio not associated with video.

I would thin a half coversion is better than an odd number conversion like 96 to 44.1 in staed of 88.2 to 44.1.
There is no question tha the bit rate is a must at 24 but is the sample rate that big of a deal?

If I'm recording at 48k to start with. Should I just record at 44.1 to make the conversion smoother to 16 bit?

Depends on the quality of your audio hardware (ADC/DAC) and the target 'resolution' (Audio-CD: 16/44.1 - DVD-A 24/192) ... by recording using 'native' samplerate you don't need SRC then later (SRC changes the data in any case). If you get much better quality by using higher samplerates then just use it and remember use as good as possible SRC software then for to convert to the target samplerate.

As you already mentioned, 24-bit is recommendation in any case.

In your own situation, you can actually find this information quite easy by searching some measurement data for your card or by measuring it with RMAA software.

Here is one good example of the difference coming from bit-depth:



and here the difference coming from samplerate:



The card in measures were SB Audigy 2 which is not very usable for other than this type testings..


Junski

Simple answer. About half of what I do ends up in a video project. So right now 100% of my projects are 24/48. I was a bit shocked a few weeks ago to find the SRC issues with Sound Forge so I have to do my SRC either in Sonar or elsewhere now.

-D

Yeah, you need to decide or find out what your target final file format has to be, and base it off of that, if possible. The main rule of thumb is that you want to do SRC (sample rate conversion) as RARELY as possible, preferably NEVER. SRC is BAD in general and should be avoided until absolutely necessary. And if you do HAVE to do SRC, then do your homework on a good program to do it... there was a thread recently on this topic in this fine forum so do a search. Not all SRC programs are equal, some of them are quite terrible to your audio. Fortunately, Sonar 6 appears to have a half-decent SRC algorithm. But you still should avoid it when possible.

I'd recommend devoting a few sessions to experimenting, ideally with a friend who can help you set up some blind tests.

No scope, no theory can beat your own ears.

The best way to do this is to record a collection of masters, all of the same sound. One common way to do this is to record a passage from a recording as it plays through your stereo system; a passage of classical music including a variety of instrument sounds and moving from soft to loud is the best (Stravinsky's "Rite of Spring" offers plenty of passages suitable for this purpose.) The concern here isn't to produce a great recording, but to compare recordings to originals and to hear the changes that different techniques impart. Listen to whether you can hear any difference among the sample rates and bit rates.

Then, make a set of sound files performing various sample- and bit-rate conversions.

Finally, wait a week or two then have a friend play you the files without telling you what you're hearing. Trust whatever your ears tell you. There are two quite different tests to do: (1) the fidelity test -- which version sounds the most like the recording playing through your sound system, and (2) the ear training test -- without regard to fidelity to the original, how to the various recordings sound? It's best to separate these tests by a week or two to give your ears time to forget.

For years, engineering types succeeded in convincing a lot of people that digitally-recorded CD's sounded better than analog recordings on vinyl. Musicians used to trusting our ears knew immediately this was crap, but apparently more people believe what they read over what they hear.

Oh, and forgot to mention. Sample rate (48KHz, 44.1KHz, 96KHz, etc.) is obviously different than bit depth (16- 24- 32-bit, etc.). As for bit depth. Do all your projects in 24 bit at least. Then in the mastering stage for CDs, dither it down to 16 bit. For the mastering stage for other medium, you may want to deliver a different bit depth. Find out what the target final file format requires. But start with at least 24 bit.

If you are confused by the terminology, read up things like this:

http://en.wikipedia.org/wiki/Sample_rate
http://en.wikipedia.org/wiki/Audio_bit_depth

Good luck!

If you use a good SRC like R8Brain then it really won't be much of an issue--Sonar's SRC is not bad either. But the best thing to do is to record at the native sample rate of the final format your project will end up at (unless you need the high fidelity of recording acoustic instruments at higher sampling rates). Go to This Website and compare each application's SRC by looking at the different Test Results (Sweep, Passband, Phase, etc). HTH.

Good info guys.
Now I have read somewhere that acoustic instruments benefit the most at higher sample rates.
I do a lot of acoustic mixed with hard rock gut wrenching distortion guitar. :-) But in a very melodic way.

But my point being would I benefit recording my acoustic stuff at 88. nor 96 and the rest at 44.1?
Or should I recordd the acoustic at higher rates first and convert it by itself and then reimpliment it back into the tracks?

I know I read that Sonar will support different sample rates in one project but the project only reads on sample rate at the botton task bar. Right?
Or did I misunderstand that feature of Sonar. I use 6PE.

i did an album project at 24/48 a few years ago, now i'm 24/44.1 for the most part. just the fact that bouncing all the tracks & converting took too long to make rough mixes.

i think there is a bit of sparkle from a 48khz recording that's missing from 44.1. many years ago ( 93 ish or so ) my band recorded a project at a studio and while we were waiting for the CDs, the engineer made a direct-to-cassette copy from the DAT master. that cassette had a nicer top end than the final CDs did, so either the 44.1 sucked off some shine or i had a super-double-hi-fi tape deck in my crappy car and at home. i tend to think that the DAT at 48khz just had more high freq information on it.

i tend to think that the DAT at 48khz just had more high freq information on it.

Well that makes sense. Basically what happens is that whenever you record an instrument at 44.1 KHz you're capturing frequencies at half that frequency (Nyquist frequency) which is 22.05 KHz. At 48 KHz the highest frequency captured is....you've guessed it....24 KHz. So you've gained more top end even though it's in an inaudible range.