Why dose Home Studio 2002 need a soundcard in order to export audio?

[nunyabusiness]

My PC is still running Win98 so I’m not able to upgrade from CWHS 2002, which can only export at 48 khz. I do have a lap top that’s running XP though. So I was thinking I would get HS 2004 or Sonar for it (which I will eventually move to my PC when I upgrade). That way I can load the files into my laptop, when I’m done editing, and export at 96 khz. Only problem is that my laptops soundcard won’t support 96 khz. I just did a test and Cakewalk won’t export if there are no soundcard drivers selected. Dose anyone know if there is a good reason for this? I mean it’s just making digital calculations right? Would this be a problem with later versions?

Thanks
NB

 

[dachay2tnr]

First off, why do you want to export a file at 96KHz if the underlying music wasn't recorded at 96KHz? All that is going to do is add some zeros to the end of your file.

Second, I'm not familiar with HS2002, but Sonar will allow you to export the file at any sample rate you want - regardless of the sound card. I don't see why HS wouldn't allow that. As you said, it's just calculations.

Third, if HS won't let you export at 96K, you can always change the sample rate on the file after it's been exported using another program. I believe there are some free sample rate converter programs available on the web.

 

[nunyabusiness]

First off, why do you want to export a file at 96KHz if the underlying music wasn't recorded at 96KHz? All that is going to do is add some zeros to the end of your file.

Well… perhaps I don’t have a full grasp of this stuff (I most certainly don’t) but I figure it like this: I’ve got 11 tracks of audio each recorded at 48 khz. Which is the same as 48000 samples per second (or something like that). When exporting from Cakewalk, the samples from the tracks must get averaged together, or the sample rate of the exported file would have to be multiplied by the 11 tracks, and end up being more like 528000 samples per second. My numbers may be off but you get the idea. The idea may be off too but something’s happening to the audio when I export at 48 because the tracks have good separation when I play them from Cakewalk, but the exported file sounds more blended. It isn't drastic, but it is noticeable.

 

[brzilian]

Well… perhaps I don’t have a full grasp of this stuff (I most certainly don’t) but I figure it like this: I’ve got 11 tracks of audio each recorded at 48 khz. Which is the same as 48000 samples per second (or something like that). When exporting from Cakewalk, the samples from the tracks must get averaged together, or the sample rate of the exported file would have to be multiplied by the 11 tracks, and end up being more like 528000 samples per second. My numbers may be off but you get the idea. The idea may be off too but something’s happening to the audio when I export at 48 because the tracks have good separation when I play them from Cakewalk, but the exported file sounds more blended. It isn't drastic, but it is noticeable.

Ummm, no.

Just because you have multiple tracks doesn't mean your sampling rate increases.

What soundcard are you using? 24bit versus 16bit audio makes more of a difference when mixing down to stereo than the sampling rate. 24bit audio helps reduce the effects of dithering during the summing of the tracks.

None of this matters though if the original audio was recorded at 16/44.1.

 

[c7sus]

If you don't select a soundcard output CW doesn't know where to send the data. You'll want to select whatever digital input/output option you have that's available on both computers.

Whatever differences you are hearing after transfer could be due to having two different soundcards in each computer.

If the card in the laptop has higher crosstalk between channels that could make your mix sound different.

 

[nunyabusiness]

Ummm, no.

Just because you have multiple tracks doesn't mean your sampling rate increases.

So how does that work? It just seems to me that several tracks each playing in 48 khz files would have to sound better than the same tracks squashed into one 48 khz file. Wouldn’t they all be fighting for the same samples?

What soundcard are you using? 24bit versus 16bit audio makes more of a difference when mixing down to stereo than the sampling rate. 24bit audio helps reduce the effects of dithering during the summing of the tracks..

I’m using a digi96/8 pro. I’m keeping every thing 24 bit the whole way through, so I haven’t turned the dithering option on.

 

[nunyabusiness]

If you don't select a soundcard output CW doesn't know where to send the data.

Why doesn’t Cakewalk just send the data to a wav file? What is the soundcard actually doing that Cakewalk can’t do by it self?

Whatever differences you are hearing after transfer could be due to having two different soundcards in each computer.

Actually, the only exports I’ve tried have been with the same computer and sound card. The only difference may have been what I was playing from. I listened to the unmixed files from Cakewalk (I guess that’s obvious), and I listened to the exported file from Wavelab. I’ll try opening up the mixed file in Cakewalk, to see if that makes a difference. Though I can’t imagine Wavelab would be the culprit.

 

[brzilian]

So how does that work? It just seems to me that several tracks each playing in 48 khz files would have to sound better than the same tracks squashed into one 48 khz file. Wouldn’t they all be fighting for the same samples?

It doesn't work in the way you think. The sampling rate is determined by the clock crystal on the soundcard. All those tracks are just playing back the previously recorded infromation in sync with the clock signal of the card.

I’m using a digi96/8 pro. I’m keeping every thing 24 bit the whole way through, so I haven’t turned the dithering option on.

It has nothing to do with turning dithering on or off in HS. Dithering will happen regardless just by the nature of how the computer handles the audio information. This is where your previous assumptions are correct - the computer sums together the tracks when mixing down to stereo. Problem arise if there is not enough information to begin with (the case with 16bit) and the computer guesses (dithers) the result. You see less of this with 24bit audio.

Consider this:

When you are talking about 16bit audio, you have 65536 possible values for the sound's amplitude (2^16 or 2 to the 16th power). With 24bit audio, you are talking about 16.7 million values (2^24 or 2 to the 24th power).

An analogy would be the difference if working with 8bit, 16bit or 24bit image files. More bits mean more colors.

 

[dachay2tnr]

It has nothing to do with turning dithering on or off in HS. Dithering will happen regardless just by the nature of how the computer handles the audio information. This is where your previous assumptions are correct - the computer sums together the tracks when mixing down to stereo. Problem arise if there is not enough information to begin with (the case with 16bit) and the computer guesses (dithers) the result. You see less of this with 24bit audio.

Actually that is not correct. That may be how the term is used in graphic applications, but in sound applications dithering is the process of adding noise to a recording in order to help offset the effects of bit truncation. And, yes, it can be turned on and off. In Sonar you can do this through Options > Audio > Advanced by checking or unchecking the "apply dither" checkbox.

Here is a definition of dithering from the WaveLab help file:

What is dithering?

Dithering is a method for reducing quantization errors in digital recordings. In the case of WaveLab, dithering is applied when reducing the number of bits in a recording, for example when moving from 24 to 16 bits, and when applying processing.

The theory behind this is that during low level passages, only a few bits are used to represent the signal, which leads to quantization errors and hence distortion. To the ear, this is perceived as "graininess" during low level passages in a recording.

When "truncating bits", as a result of moving from for example 24- to 16-bit resolution, such quantization noise is added to an otherwise immaculate recording.

By adding a special kind of noise at an extremely low level, the quantization errors are minimized. Indeed, the added noise can be perceived as a very low-level quiescent hiss added to the recording. However, this is hardly noticeable and much preferred to the distortion that otherwise occurs.

The reason for the dithering block to be last in the Master Section is that the output level must not be changed after dithering a signal.