Easy, Spitsbergenclown
![]()
That too!
Reminds me, I has going to put all Spitsbergen schematics on line. Need to finish that.
Jan
Now THAT is a shameless plug!Linear Audio![]()
At what point? The DAC clock (the only place it actually counts)? That seems... remarkable.
No inter data bit, the way the test is specified. I should mention I was looking at RMS jitter not peak.
I don't listen to the clock. Just the calendar.
Now THAT is a shameless plug!![]()
You asked for it - actually you pretty much twisted my arm for it
Jan
Spitsbergen schematics on line.
Mr Wadsworth may also be pleased
Jan, that wasn't me ..You asked for it - actually you pretty much twisted my arm for it
Jan
the Stuart paper's points that contradict Marsh digital audio word 24/32 bit "requirement"
Bob suggests 20 bit may only be useful at 44.1 to encode down to the human hearing threshold in quiet - without modern noise shaped dither
Stuart's paper:
at higher sample rates he assumes you use dither, noise shaping taking advantage of the inaudible bandwidth and using less bits, down to as few as 11 bits at 96k
and that is with the assumption of coding down to the human perceptual noise floor - not required as many keep pointing out in listening to commercial produced music, made with real world noise limited microphone, played back at "live event" levels in real homes
http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-1026.html#post2485689
Bob suggests 20 bit may only be useful at 44.1 to encode down to the human hearing threshold in quiet - without modern noise shaped dither
Stuart's paper:
This graph still suggests that a well-engineered 20-
bit channel should be adequate, bearing in mind that very few rooms, no recording venues and no
microphones genuinely approach the quietness of the 20-bit noise floor.
at higher sample rates he assumes you use dither, noise shaping taking advantage of the inaudible bandwidth and using less bits, down to as few as 11 bits at 96k
CONCLUSIONS
This article has reviewed the issues surrounding the transmission of high-resolution digital audio. It is
suggested that a channel that attains audible transparency will be equivalent to a PCM channel that
uses:
· 58kHz sampling rate, and
· 14-bit representation with appropriate noise shaping...
and that is with the assumption of coding down to the human perceptual noise floor - not required as many keep pointing out in listening to commercial produced music, made with real world noise limited microphone, played back at "live event" levels in real homes
http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-1026.html#post2485689
Last edited:
/OT
lease take a look here
http://www.diyaudio.com/forums/equipment-tools/271300-pcb-scope-probe-tip-adapter.html#post4256238
/OT off
http://www.diyaudio.com/forums/equipment-tools/271300-pcb-scope-probe-tip-adapter.html#post4256238
/OT off
the Stuart paper's points that contradict Marsh digital audio word 24/32 bit "requirement"
Bob suggests 20 bit may only be useful at 44.1 to encode down to the human hearing threshold in quiet - without modern noise shaped dither
Stuart's paper:
at higher sample rates he assumes you use dither, noise shaping taking advantage of the inaudible bandwidth and using less bits, down to as few as 11 bits at 96k
and that is with the assumption of coding down to the human perceptual noise floor - not required as many keep pointing out in listening to commercial produced music, made with real world noise limited microphone, played back at "live event" levels in real homes
http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-1026.html#post2485689
The over simplification here is comparing broadband noise vs perceptual limits. You can hear signal below the noise floor. So even with a microphone with a noise floor of 20 dB it could record a tone at 3,000 hertz 20 db below its' noise floor and you would still hear it.
Now 32 bits would require significant atmospheric modulation even if it were possible.
Another omission is you may have 20 bits of level accuracy but the timing (jitter) accuracy is ignored.
Also ignored is bandwidth and phase sensitivity.
Scott is perfectly happy as he often states with 16 bits and for what he describes as his listening conditions it is quite reasonable. Dick M. has a system that can play louder and is in a quieter location, it is just as reasonable for him to find limits in 16 bits.
Now the current practical limit is around 21 bits, not to use it just seems silly. If you want to make the size/cost argument then the conclusion should be MP3 is optimum! (Do remember Bell Labs developed this technology with a bit of an eye on reducing telephone system requirements.)
George, while you're at it - check how the clock is generated in this player
Best,
Here it is (and as you know, implementation details matter)
George
Attachments
Here it is (and as you know, implementation details matter)
George
George, what implementation details? No PLL/VCO based oscillator will produce the same low jitter figures as quartz crystal based oscillator. One can't polish a turd, as they say
Audio Asylum Thread Printer
Attachments
At the end, with all the measuring instruments you have, and because only the results matters, why don't you make distortion measurements to compare the two sandard's reproduction quality on your system ?
... while i pretend it is the 96KHz, mostly ;-)
That is what I intend to do.... but have a lot of pans in the fire a.t.t. So, it is taking longer than I would like.... but I have to do this stuff (audio related) between projects of higher priority to me. But I'll get there.... its why I invested a lot of money in test equipment. Now I am retired but still dont have as much time as I would like. Anyway --->
To do camparisons based on specs isnt productive for me.... If I starrt with resolution, and go from there -- INL, DNL, gain error, gain temp coeff, PSRR etc each at 1-2 bits error each max (spec'ed). And then add THD and noise as equivelent bits etc etc etc.... Its just a nightmare. Then there is the ADC/DAC affects of pre=post filtering, jitter reduction schemes to measure etc etc etc.
No, I'm going to buy a really good ADC and let that represent the record side, and run it thru the DAC and see what I get compared to the ultra-pure tone input.... what kind of total junk is on the output. Then multitones. At least using the normal levels used and not ref to max or 0FS output. Like -20 and -30 and -40dB levels and lower. Do it with 16b/44 and 24b/96. And also compare to CD disk playback system.
If this has been done, I'd like to see the results of various systems tested. Sure save me a lot of time in both setup and results.
THx-RNMarsh
Last edited:
George, what implementation details? No PLL/VCO based oscillator will produce the same low jitter figures as quartz crystal based oscillator. One can't polish a turd, as they say
Audio Asylum Thread Printer
OK. On paper not the best in the world, plus audio clocks with this IC have more jitter than the video clock (but the numbers are low).
As Richard’s Sony CD machine is a universal player (with a billion self, auto, manual adjust/test modes and diagnostics
DSP, AV Controller, DSD Decoder, Audio Digital Signal Processor, DAC, all spread across a large highly populated (on both sides) PCB.
Even if it is the best clock in the world in terms of specs, the actual clock jitter, noise ect measured at the input on each of the supplied ICs, will depend on tracks routing, layout parasitics, power supply isolation.
These are the things I called implementation details.
(accidentally, the DAC IC is located the furthest away from the clock, on the other side of the PCB but Richard doesn't uses this DAC)
George
Last edited:
George, what implementation details? No PLL/VCO based oscillator will produce the same low jitter figures as quartz crystal based oscillator. One can't polish a turd, as they say
Audio Asylum Thread Printer
You have to use a PLL in a CD player! The clock is derived from the data on the disc. In theory you could vary the speed of the disc to get the data to match your clock, but inertia, out of roundness etc prevent this approach.
That is why there CD players are not true discrete level/discrete time digital devices. They really are analog in time.
That is what I intend to do.... but have a lot of pans in the fire a.t.t. So, it is taking longer than I would like.... but I have to do this stuff (audio related) between projects of higher priority to me. But I'll get there.... its why I invested a lot of money in test equipment. Now I am retired but still dont have as much time as I would like. Anyway --->
To do camparisons based on specs isnt productive for me.... If I starrt with resolution, and go from there -- INL, DNL, gain error, gain temp coeff, PSRR etc each at 1-2 bits error each max (spec'ed). And then add THD and noise as equivelent bits etc etc etc.... Its just a nightmare. Then there is the ADC/DAC affects of pre=post filtering, jitter reduction schemes to measure etc etc etc.
No, I'm going to buy a really good ADC and let that represent the record side, and run it thru the DAC and see what I get compared to the ultra-pure tone input.... what kind of total junk is on the output. Then multitones. At least using the normal levels used and not ref to max or 0FS output. Like -20 and -30 and -40dB levels and lower. Do it with 16b/44 and 24b/96. And also compare to CD disk playback system.
If this has been done, I'd like to see the results of various systems tested. Sure save me a lot of time in both setup and results.
THx-RNMarsh
Your proposal if you are using a good master clock might take jitter out of the process. Now if you can introduce jitter as a variable that would get very interesting.
ES
In theory you could vary the speed of the disc to get the data to match your clock, but inertia, out of roundness etc prevent this approach.
Jeez... This is EXACTLY how CD players work! One varies the rotational rate of CD. DAC is clocked by the crystal oscillator.
Don't CD players have a small digital buffer to hold samples read from the CD and reclock them to the DAC clock? For sure Walkmans would need this.
BTW, has there even been a CD player that didn't reclock like this, IE if you turned the CD player while in operation the sound would warp like a record player?
BTW, has there even been a CD player that didn't reclock like this, IE if you turned the CD player while in operation the sound would warp like a record player?
In CDP system, disc spin motor speed is controlled in order to keep data buffer half full on average.
The PLL is used to derive the disc spin control voltage.
Audio data is clocked out of the buffer at crystal oscillator determined rate.
The crystal oscillator also provides clock for DAC output stage.
Dan.
The PLL is used to derive the disc spin control voltage.
Audio data is clocked out of the buffer at crystal oscillator determined rate.
The crystal oscillator also provides clock for DAC output stage.
Dan.
..... a universal player (with a billion self, auto, manual adjust/test modes and diagnostics) this clock-generator IC is connected to the following ICs:
DSP, AV Controller, DSD Decoder, Audio Digital Signal Processor, DAC, all spread across a large highly populated (on both sides) PCB.
Even if it is the best clock in the world in terms of specs, the actual clock jitter, noise ect measured at the input on each of the supplied ICs, will depend on tracks routing, layout parasitics, power supply isolation.
These are the things I called implementation details.
(accidentally, the DAC IC is located the furthest away from the clock, on the other side of the PCB but Richard doesn't uses this DAC)
George
One just needs to look elsewhere around here for 'sound cards' (usually external) being used for measuring THD/harmonics and see how they have been improved by rearranging/modifying just the wiring. Theory alone may not be enough. You just gotta measure the final system result and get the whole, big picture of what you are listening to.
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.
- Home
- Member Areas
- The Lounge
- John Curl's Blowtorch preamplifier part II