jaddie
Account deactivated by request.
- Joined
- Mar 28, 2011
- Posts
- 1,253
- Likes
- 91
Quote:
Please supply a bit of backup for that one. And while you're at it, might take a look at the audibility threshold of jitter in general.
Quote:
Ever heard of oversampling? I think it's what you're referencing in the second point. It's found on every CD player in the last 15 years, probably 20, and pretty much takes care of the issues with high-order brick wall filters. Careful with the ringing analysis, though. If you are referring to the visual representation of a square wave, just remember that band-limiting a square wave with even a theoretically perfect filter will display ringing because of the removal of harmonics that make it square in the first place. Just because it's a-ringin' don't mean it's bad. Ringing doesn't exist apart from the cause, which may or may not be audible.
Quote:
Higher bit rates will theoretically increase the dynamic range. Bit rates are the result of two things: bit depth and sampling rate. You can increase one, or the other, or both, and end up at a higher rate than you started with. The dynamic range of 24 bits is at least partly mythological, as there are no non-cascaded ADCs with a real 24 bit dynamic range. Typical ADC performance is around 20 bits, even though 24 bit words are generated. The cascaded ADC architecture results in true 24 bit dynamic range. However, you still can't really hear that, because of the dynamic range of your listening environment. If you had an NC20 room (unlikely unless you are way out in the country and have no HVAC running or equipment in the room with you), if you match the noise floor of 16 bits with your room noise floor, then the theoretical 96dB range of 16 bits would put the max at 116dB SPL, and realistically, it's more like 92dB and that puts the max at 112dB SPL. If you could achieve it, the theoretical 144dB dynamic range of 24 bit, with noise floors matched, puts the max somewhere around pain and irreversible hearing damage.
The real advantage of 24 bits is in processing in post production where DSPs doing dynamics processing and EQ have more data to "chew on", after which down-sampling to 16 bits is fine. 44.1KHz might be a bit low, but there's not much evidence that doubling is necessary.
Quote:
That's pretty much being proven right here. Gab those ABX files and go nuts.
Higher sampling rates generally have lower levels of jitter.
Please supply a bit of backup for that one. And while you're at it, might take a look at the audibility threshold of jitter in general.
Quote:
CD playback requires the use of a brick wall filter to cut off frequencies above 22khz - this filter causes ringing. If you look at a picture of a typical CD impulse, the signal will show ringing before and after the pulse. Higher sampling rates can be implemented with less harsh filters and thus generate less ringing.
Some DACs use different reconstruction filters that eliminate the pre impulse ringing, but these will generate artifacts that are in the audible frequency range. Higher sampling rates with these filters can move these artifacts above the audible range.
Ever heard of oversampling? I think it's what you're referencing in the second point. It's found on every CD player in the last 15 years, probably 20, and pretty much takes care of the issues with high-order brick wall filters. Careful with the ringing analysis, though. If you are referring to the visual representation of a square wave, just remember that band-limiting a square wave with even a theoretically perfect filter will display ringing because of the removal of harmonics that make it square in the first place. Just because it's a-ringin' don't mean it's bad. Ringing doesn't exist apart from the cause, which may or may not be audible.
Quote:
Higher bit rates will theoretically increase the dynamic range.
Now whether any of these effects are audible? Some say they are. Some don't.
Higher bit rates will theoretically increase the dynamic range. Bit rates are the result of two things: bit depth and sampling rate. You can increase one, or the other, or both, and end up at a higher rate than you started with. The dynamic range of 24 bits is at least partly mythological, as there are no non-cascaded ADCs with a real 24 bit dynamic range. Typical ADC performance is around 20 bits, even though 24 bit words are generated. The cascaded ADC architecture results in true 24 bit dynamic range. However, you still can't really hear that, because of the dynamic range of your listening environment. If you had an NC20 room (unlikely unless you are way out in the country and have no HVAC running or equipment in the room with you), if you match the noise floor of 16 bits with your room noise floor, then the theoretical 96dB range of 16 bits would put the max at 116dB SPL, and realistically, it's more like 92dB and that puts the max at 112dB SPL. If you could achieve it, the theoretical 144dB dynamic range of 24 bit, with noise floors matched, puts the max somewhere around pain and irreversible hearing damage.
The real advantage of 24 bits is in processing in post production where DSPs doing dynamics processing and EQ have more data to "chew on", after which down-sampling to 16 bits is fine. 44.1KHz might be a bit low, but there's not much evidence that doubling is necessary.
Quote:
Now whether any of these effects are audible? Some say they are. Some don't.
That's pretty much being proven right here. Gab those ABX files and go nuts.