sample rates

Lt. Bob

Spread the Daf!
Hello everyone.
Glad I stumbled onto this site!
I'm mixing down to a DAT that will be professionaly mastered. Should I set the DAT to 48kHz or 44.1kHz
Thanks!
 
Before you mix anything to any particular format, you'd better contact the folks you plan to bring it to afterwards and ask them what they'll accept. For most versatile studios this may be no big deal, but why waste your time because of an over optimistic assumption. A good friend runs an auto repair shop (he's easily the best in the biz in his niche in his area) and he'll accept cash or your personal check or a money order or even barter in some cases, but he has no facility to accept plastic. Many of his customers are taken by surprise at this policy. Why? They didn't ask in the first place. One man's standard is another mans exotica.
And of course 48KHz will provide some more fidelity. How much more? I dunno, I've never heard any difference. That doesn't mean there isn't any. But if you're paying to have it mastered I'd give them the best source files I could.
 
Yeah, I don't hear a difference between 44.1k, 48k, and 96k sample rates. However, I go with 48k because psycologically, I'm getting better quality with 48k than with 44.1k and with just a little more disk or digital tape space unlike the 96k storage eater. But, I soon want to focus on using the extreme noticeable difference of 24 verses 16-bits. If your converters and software can handle it; and you have a common 24-bit storage format found in most major mastering facilities, then I'd send that 48k not yet resampled and that 24-bits not yet dithered to the mastering house. If you don't have 24-bit storage capabilities but have a DAT, then I'd buy an Apogee Rosetta 48k, 24-bit stereo A/D converter and use its UV22 dithering process to 16-bits, digitally out to either your computer or your DAT.
 
Thank you for your quick replies.
"Recording egineer", I have no clue what dithering is but I assume that that's because I'm not using a computer but a Tascam 564. I don't think that knowing this would alter your responses but if it does please let me know.
Thanks again
 
Dither is when you cram 18, 20, or 24 bits onto a 16-bit medium (like a DAT or CD). I'm considering buying the Apogee Rosetta 48k, 24-bit stereo A/D converters which dither 24-bits to 16-bits. They cost about $1200, but that's a real nice price for top of the line converters. Your DAT recordings would be taken to a whole new level!
 
Can someone here explain the relationship between: sample rate-44.1, 48 etc..and bit resolution: 16 bit, 24, etc...?
Are there systems that allow you to bump resulution as well as sample rate? If I'm close with this it seems to me that even if the sample rate is increased to 96khz @ 16 bit res, theres still no more data being captured, but only more "slices" like a pie: if its only 10 inches around, I can cut it into 44 slices or 96 slices, but its still worth only a 10" pie. Is this an acurate analogy? Please help me get this right.
 
There's definitely more data being captured when using higher sample rates and larger sample sizes. Just compare file sizes of sound bites of the same length in seconds.
The question really is: Is this just redundant data? I don't think so, but the extra resolution and dynamic range available with these fatter formats cannot really be distinguished from CD quality without some pretty fancy playback equipment. And a really good set of ears. That pizza analogy doesn't quite work, because an audio waveform isn't uniform from slice to slice, especially as the size of the slices get smaller. Just open a .wav editor like Sound Forge and zoom in as far as it goes to get a graphic representation of what I'm talikng about. Some slices might have no sauce, others no crust! And some might just be a wad of greasy pepperoni. It wouldn't fly with a pizza pie.
 
As I understand it, sampling rate has to do with the frequency range that is captured (half of the sampling rate = the highest frequency that can be captured). Bit depth has to do with amplitude: the higher the bit depth, the wider the amplitude range, meaning you can have quiet passages that are as clear as the loud ones, allowing for more dynamic music.

In answer to the original question, I think it's kinda pointless to record at 48kHz, since 44.1 kHz is already well beyond the frequency range of human hearing, and since it'll have to be resampled to 44.1 by the mastering house anyway if you want to put it on a CD.

Ryan
 
As Sonusman has pointed out- there is some promising research into the premise that frequencies present in a recording that can't be heard on their own do contribute to the overall sense of the perceived sound. I'm more convinced that a 50% increase in sample size (24 bit from 16 bit) will improve the sound more noticeably than a similar expansion of sample rate when compared to 44.1KHz, 16 bit sounds.
But the standard rates I've heard of progress from 44.1KHz to 48KHz to 96KHz.
Ratios of: 1:1.088435:2.176871 or 0.459375:0.5:1.
But if I'm paying some third party to master my mix, and I have the capability to store it for transport in a higher resolution format, and they can use that format, then any extra cost for the media (for example 3 CDRs instead of one) is inconsequential.
Hopefully the mastering house will have worthy sample rate/size converters. That really can be the Achilles Heel of this whole operation, which is why I haven't utilized the 48 KHZ recording capability on my sound card. In defense of these higher sample rates/sample sizes, if you're going to apply a series of software effects, it makes sense that they would sound better if the transforms were done with the sample represented with higher resolution. That way any approximations that occur at the edge of digital resolution are further rounded at the end of the chain, but the full resolution is used in the effects calculation, much like reading a value to +/- 0.01, but reporting the value via digital display to +/- 1. This "smoothes" the digital signal.
 
As far as bit resolution goes, I think everyone will agree it's best to sample at 24-bits and dither to 16-bits; as long as you use a relatively decent logarithm and relatively decent A/D (and if needed) D/A converters; and a mastering house will most likely have better logarithms and converters than we can afford.

When it comes to higher sample rates than 44.1k, I think it's a matter of personal opinion if the difference is worth the storage space. For me, it's not worth all that space for 96k when compared to 44.1k. However, I feel the very little more storage of 48k compared to 44.1k is worth it even if the difference is all in my head.
 
Thanks to all who answered in here. I'm sorry I kind of stole the thunder from the original Question.
I dont think I would have quite gotten "it" with only one reply. But the collection of responses paints a fairly clear idea of whats going on.
I hate to bring up my pie bit again, but I'd like to press the subject: the thing I have been fixating on is that in order to "really get more data", you need a bigger pie (Bit count) not just splitting that pie up either 44x or 96x.
Is this just out there?
Thanks for your patience
 
Back
Top