CD is actually 44.1 KHz, so it wouldn't be exactly 176...
Have a look at this about how that was chosen:
http://www.cs.columbia.edu/~hgs/audio/44.1.html
and here for an anecdote:
http://stason.org/TULARC/entertainm...digital-data-written-in-44-1-kHz-samples.html
(It's good that Karajan wasn't much of a Mahler conductor,
or to fit *that* ninth symphony, we'd be living with 32 KHz or
have CDs of 10 inches diameter...)
-
Have a look at this about how that was chosen:
http://www.cs.columbia.edu/~hgs/audio/44.1.html
and here for an anecdote:
http://stason.org/TULARC/entertainm...digital-data-written-in-44-1-kHz-samples.html
(It's good that Karajan wasn't much of a Mahler conductor,
or to fit *that* ninth symphony, we'd be living with 32 KHz or
have CDs of 10 inches diameter...)
-
not true at all. You can oversample by non integer numbers, like 384khz result out of both 44-48khz , -Think taking samples out of a FIFO buffer and you only want a single dsp clock. Upsampling term was coined because its supposed to do the heavy lifting before the builtin chip oversampling , and filter jitter as well. Also more refined processes can be tought of than an ASRC, the upsampling possibilities are wide and not by any means the asynchronous manner is necessary .
Last edited:
This does not mean that 192k from 44.1 is right.
Most experienced audiophiles find 176.4 better sounding than 192 k when CD is upsampled. I have had long experience of this with all kinds of gear.
Assynchronous upsampling of any kind is not desirable.
Upsampling from 44.1 to 48, and subsequently other frequencies such as 96 and 192, does not have to be asynchronous since 44.1 is related to 48 by a ratio of 147:160. You want to sit down with a cup of coffee when you first go through the theory, and the filters aren't trivial, but the problem is simple enough that any decent treatment of multirate conversion uses it as an example.
In other words, if the upconversion sounds funky, then something's gone wrong.
another nice tale in the audio-folk-lore but if you throw enough dsp resource at it, its no matter , think about it ,they already did it once going 96->44.1
You and others should use your ears with a range of high quality software and hardware. There is sufficient evidence to show thru measurement and listening tests that differences in the mathematical modelling and programming of resampling dsps matter a lot. Post and or pre filtering implementations can also affect sonics greatly.
The human ear is more sensitive than theoretical performance claims!!!
So are you going to storm to the nearest studio and demand they use 176kHz for recording?
I have no problem with differences in math affecting sonics - that's part of my job - but programming? A filter in C is the same as a filter in machine language is the same as a filter in Ada if the math is the same.
It's not as if filtering is unknown territory, either. Perhaps thirty years ago its audible effects were terra incognita (Philips was still trying to convince the world 14 bits were enough for audio!), but now a large body of practice has grown up to supplement theory. Modern DSPs have more than enough grunt to handle anything one can throw at them, so the end result is mostly a matter of the practitioner's skill.
As for asynchronous converters, current ASRCs are no slouches. The first ones may have been a tad noisy, but Burr-Brown 4192s keep the THD+noise below -140 dB.
There is a long thread here a few years old on ASRC by I believe one of the founders of Silicon labs. Basically it doesn't matter to what frequency you convert. What matters is what happens at the DAC. Some dacs such as the wolfson 8741 can only use the apodizing filters on high-rate data. For those not wanting to read the entire thread, I've summarized it (limited to my understanding) here: http://hifiduino.blogspot.com/2009/06/how-asynchronous-rate-conversion-works.html
Theres a lot of artificial confusion about this ... I can grab a CD, bandlimit all the tracks with a superior minimum phase filter, burn a new copy and say my CDP has an apodizing DAC now, with appetizing 0 prering CD-s. This is just DSP guys ... 
(and by the way the NOS fans should do this to prevent the nice staircase prering they dont want know about)
(and by the way the NOS fans should do this to prevent the nice staircase prering they dont want know about)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.