|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.255.52.139
In Reply to: RE: Digital improving -final posted by J. Phelan on April 12, 2016 at 12:46:21
"Clock noise. Coming from oscillators and quartz-crystal resonators. Co. are now making (on-board) custom clocks, to reduce this. We had our choice of outboard-clocks, for a high price. 'Jitter' applies here -but this was actually noise, not timing imperfections, that did the damage."
The clock noise isn't nearly as bad as the DAC's bit-switching noise from its transistors switching on to off or off to on, in generating the music signal. This is how the RF emissions/interference is generated during active digital audio playback. In my opinion, this is the single biggest problem with digitized audio playback, especially at higher resolutions.
"Linear-phase filtering /asynchronous sample-rate conversion. Either (or both) of these caused 'pre-ringing' and aliasing errors."
Asynchronous conversion might be in the form of a pre-ringing (linear phase) filter, but the conversion method is in itself not responsible for the pre-ringing. Whether the conversion method is synchronous or asynchronous, any digital filter can be programmed for either pre-ringing or non-pre-ringing (minimum phase).
The problem with asynchronous conversion is that the signal itself is compromised because the it is no longer "bit perfect." If the timing of the independent clocks (used in asynchronous conversion) are not precise, noise will also get embedded into the audio signal.
"**16/44 never should have been attacked, being we couldn't hear its potential !!"
I agree about the potential.... It's attacked because getting it to reach its potential has been excruciatingly difficult.
Follow Ups:
Alright - so before 2010, most DACs were programmed wrong -to pre-ring. But how audible is this ? Is there proof ?
I believe there's bit-switching noise, but I thought digital measured very good. Has anyone measured the RF ?
Good measuring does not mean good sound. R2R dacs don't measure as good as delta sigma but I still like there sound better
Alan
"Alright - so before 2010, most DACs were programmed wrong -to pre-ring. But how audible is this ? Is there proof ?"Right or wrong really doesn't apply.... It's a matter of personal preference.
Linear phase measures better, especially "phase error".... But I think minimum phase is truer to the original pre-digitized audio waveform..... There are fans and detractors of both. All things equal, I prefer minimum phase.... (I think minimal pre-ringing sounds more pleasant than no pre-ringing, but there should be more post-ringing, relatively speaking.)
"I believe there's bit-switching noise, but I thought digital measured very good. Has anyone measured the RF ?"
There should have been a log with such data put forth back since day one of digital audio playback..... I think it would impact people's decisions on what products to attain.... All I know is that putting an AM radio near an active DAC, the radio station is often drowned out by the noise from the switching in the DAC. (This is why I never liked PCs for digital audio playback. They're RFI monsters relative to dedicated DACs.)
Edits: 04/16/16
"..drowned by the noise". But it could be the power supply, no ?
"But it could be the power supply, no ?"
Only if it's a switching power supply.... Some of those things are nasty...... But most good DACs don't use them.
Most audio gear still use linear supplies..... Poorly designed linear supplies will make an AM radio emit 60/120 Hz "buzz" out the speaker..... But not enough to drown out a station.
Could be the clock - how do we know it's the switching ?
Scanning around, I found an old Analog Devices spec-sheet (for a 20-bit DAC) which claimed noise of -120db. But only if you employ the extra (recommended) capacitors. In reality, it was probably -115db. Still low, but some are now claiming -140db, in a complete D/A device.I then found a research paper by TI that talked about clock-noise. But the report also mentioned 'intrinsic' DAC-noise. So, you're right -the question is how much...
Edits: 04/17/16
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: