Support
Looper's Delight
|
Support |
|
|
Ahh, I see what you mean.. 1bit converters works quite differently than older parallel converters.In the old parallel converters, all bits were level set in parallel at the same time. Those were also called successive approximation converters, and was build on a ladder resistor network, each step in the ladder being calculated, and lazer trimmed to hairfine tolerenses, to power-of-2 voltage steps.
A 1bit converter will look at the level difference in the analog signal from each convertion sample to the next, and set the next word to the determined new value. In other words, it's based on voltage/level comparators.
And no, it doesn't change one bit only at a time ;) andy butler wrote:
Whose on first base? I didn't mean that.rather to question what the *numerical* bit depth of the 12.288Mhz sampling would be.I was prompted to doing that by looking up "how does an oversampling ADC work?"and reading that the sampling was at 1bit. Now, I'm easily convinced that sampling at 16bit 12.288Mhz and digital filtering to 44kHz would rid of aliasing very easily, ....but not so convinced (yet) knowing about the 1 bit sampling. andy Charles Zwicky wrote:Bit depth is independent of sample rate. The bit depth simply determines the number of discreet amplitude levels which can be quantified at each sample. The number is expressed as an exponent of 2 (because each bit is binary - a 1 or a 0). 16 bits = 2^16 = 16,535 amplitude values, 24 bits = 2^24=16,777,216...! Sent from my iPad Nano On Dec 16, 2011, at 4:42 AM, andy butler <akbutler@tiscali.co.uk> wrote:Charles Zwicky wrote:256 x 48khz = 128 x 96khz = 64 x 192khz = 12.288MhzIn other words, the input is sampled at rate of 12.288Mhz independent of the system sampling rate. For one thing, this means that aliasing is nonexistent even at a "44.1khz" sample rate.Interesting, what's the bit depth at the 12.288Mhz rate? andy
-- rgds, van Sinn