A new DAC idea - diyAudio
Go Back   Home > Forums > Source & Line > Digital Source

Digital Source Digital Players and Recorders: CD , SACD , Tape, Memory Card, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 17th August 2015, 08:51 AM   #1
vecna is offline vecna  Denmark
diyAudio Member
 
Join Date: Aug 2015
Location: denmark
Default A new DAC idea

Hi have been toying with this idea for a while, and since have joined this forum have decided to publicize ideas , ( actually dont have that long to live so ideas bloody useless to me )

Ideal output of a DAC ( no filtering, over sampling etc ), first attachment


The problem being ( in the frequency domain ) the "spaces" between the samples. And the transition between one bit and the next ( the rise time ) . Both of these create high frequency "images" in the frequency domain. Its this problem that causes the responses of filtering, oversampling etc etc etc.
How about we replace the "spaces" with a fixed frequency sine wave , second attachment.
Am i correct that if the superimposed sine wave is pure, we suddenly have a very narrow band noise ? Instead of a bloody wide band noise ?
And that suddenly filtering becomes easy ???
Attached Images
File Type: png DAC otput 1.png (53.0 KB, 623 views)
File Type: png DAC otput 2.png (8.4 KB, 650 views)
  Reply With Quote
Old 17th August 2015, 09:12 AM   #2
diyAudio Member
 
5th element's Avatar
 
Join Date: Oct 2002
Location: England
I am sure someone more digital world savvy with jump in on this but technically speaking I think that's what dither is all about - adding in random high frequency noise to smooth out the steps between the samples, then filtering out the noise after the reconstruction of the analogue waveform has taken place.
__________________
What the hell are you screamin' for? Every five minutes there's a bomb or somethin'! I'm leavin! bzzzz!
  Reply With Quote
Old 17th August 2015, 09:22 AM   #3
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
You still have the sharp steps. All you have done is replace the flat top. Anyway, where are you going to get the smooth sine wave from? The DAC can't do it as it is clocking at the much slower rate of the edges.

This idea won't work. You need to gain more understanding of sampling, reconstruction etc.
  Reply With Quote
Old 17th August 2015, 10:05 AM   #4
diyAudio Member
 
abraxalito's Avatar
 
Join Date: Sep 2007
Location: Hangzhou - Marco Polo's 'most beautiful city'. 700yrs is a long time though...
Blog Entries: 123
Send a message via MSN to abraxalito Send a message via Yahoo to abraxalito Send a message via Skype™ to abraxalito
Quote:
Originally Posted by vecna View Post
Am i correct that if the superimposed sine wave is pure, we suddenly have a very narrow band noise ?
Seems you've superimposed the DAC's output on a sinewave rather than as it was formerly, superimposed on nothing at all. I can't see that as a step forward myself, substituting a narrow band noise for no noise at all.
__________________
When irony first makes itself known in a young man's life, it can be like his first experience of getting drunk - Robertson Davies
  Reply With Quote
Old 17th August 2015, 11:04 AM   #5
vecna is offline vecna  Denmark
diyAudio Member
 
Join Date: Aug 2015
Location: denmark
Quote:
Originally Posted by DF96 View Post
You still have the sharp steps. All you have done is replace the flat top. Anyway, where are you going to get the smooth sine wave from? The DAC can't do it as it is clocking at the much slower rate of the edges.

This idea won't work. You need to gain more understanding of sampling, reconstruction etc.
The drawing is not so good, but the rise times between the steps should also be part of the sine wave , thus the "edges" will have a sine-wave roll off, removing , i think , the "images" created in the frequency domain by that transition

The sine wave itself can come from a separate "pure" oscillator, it can be introduced by custom front-end for the dac, one can easily reuse the system clock which in most cases is operating at 10x or more the sampling rate, and in most cases is a low harmonic "pure" sine wave before it is turned into square by a latch or comparitor or such. not trivial circuit shape ( to get it right ) but defiantly realizable , the only reason have not produced a working circuit is cos just cannot be bothered, i got my own problems at the moment :-)

The question is "IF" the electronics are realized , would the resultant filters needed be better or worse than what is currently used? It is my understanding that if for example if i were to use a 1mhz sine , i would only have to filter EXACTLY 1 mhz, ( a notch filter for example ) , i could ignore everything else ??
BUT as you pointed out i am a little green in this area, thats why need more expert opinion :-)
  Reply With Quote
Old 17th August 2015, 11:13 AM   #6
vecna is offline vecna  Denmark
diyAudio Member
 
Join Date: Aug 2015
Location: denmark
Quote:
Originally Posted by abraxalito View Post
Seems you've superimposed the DAC's output on a sinewave rather than as it was formerly, superimposed on nothing at all. I can't see that as a step forward myself, substituting a narrow band noise for no noise at all.
Well hopefully i have removed the "edges" between transitions, thus a lot of crap in the frequency domain is gone ?????.

In the real waveform there is LOTS going on in the space between samples, anything we try to put between those transitions is just a guess, and is not that the basic problem with digital reproduction ??

As someone mentioned its like dithering , but in that case we are randomly placing small values in the gaps, thus preventing LSB "sticking" and enriching the sound ( as is done very successfully in digital pictures to enhance the content for human eyes ).
But that produces "white noise" which cannot be filtered, does this solution ( if it is even a solution ) have the advantage of doing the same , if not better, and still allowing one to very easily filter the "noise" out ???
  Reply With Quote
Old 17th August 2015, 11:20 AM   #7
diyAudio Member
 
Join Date: Dec 2003
Location: Sydney
The first photo is really what you should be aiming to achieve, in particular at 20kHz. The least ringing, overshoot, the fastest rise times are all desirable and will approach theoretical limit of a DAC, if it is loaded appropriately. You have probably guessed already that only R2R DAC's could possibly give you such response, and in addition -> only a true NOS topology can get you there. The sine wave, as pointed out by others, will have to be superimposed (generated) "by meanings other than a DAC" - DAC can conduct at a specific clock interval - and at that time -> it will give a specific output (current) value.
  Reply With Quote
Old 17th August 2015, 11:40 AM   #8
diyAudio Member
 
abraxalito's Avatar
 
Join Date: Sep 2007
Location: Hangzhou - Marco Polo's 'most beautiful city'. 700yrs is a long time though...
Blog Entries: 123
Send a message via MSN to abraxalito Send a message via Yahoo to abraxalito Send a message via Skype™ to abraxalito
Quote:
Originally Posted by vecna View Post
Well hopefully i have removed the "edges" between transitions, thus a lot of crap in the frequency domain is gone ?????.
Not from the drawing - those vertical lines between the stair treads do look rather vertical. Perhaps you were envisaging something along the lines of Hawksford's raised-cosine DAC idea (his AES paper, Nov 1994)?

Quote:
In the real waveform there is LOTS going on in the space between samples, anything we try to put between those transitions is just a guess, and is not that the basic problem with digital reproduction ??
In my understanding there's no guessing going on as what's between the impulses is precisely constrained by the bandlimiting at the input. There's always only going to be one way to join up the dots within the band limit constraint.
__________________
When irony first makes itself known in a young man's life, it can be like his first experience of getting drunk - Robertson Davies
  Reply With Quote
Old 17th August 2015, 12:09 PM   #9
diyAudio Member
 
Join Date: Apr 2014
Location: Italy friuli
Default paralleling dac

Quote:
Originally Posted by 5th element View Post
I am sure someone more digital world savvy with jump in on this but technically speaking I think that's what dither is all about - adding in random high frequency noise to smooth out the steps between the samples, then filtering out the noise after the reconstruction of the analogue waveform has taken place.
Yes is best way this is possible by paralleling some number of dac...like the tda1543
schematic
  Reply With Quote
Old 17th August 2015, 12:10 PM   #10
vecna is offline vecna  Denmark
diyAudio Member
 
Join Date: Aug 2015
Location: denmark
Quote:
Originally Posted by abraxalito View Post
Not from the drawing - those vertical lines between the stair treads do look rather vertical. Perhaps you were envisaging something along the lines of Hawksford's raised-cosine DAC idea (his AES paper, Nov 1994)?



In my understanding there's no guessing going on as what's between the impulses is precisely constrained by the bandlimiting at the input. There's always only going to be one way to join up the dots within the band limit constraint.
Like said drawing not very good, will have to do it better i see :-). The Idea is to join all the dots with pure a sine wave ( or whatever fill-in is appropriate, the one with the least harmonics ) so square wave steps no longer exist.

will have a look at that paper, thank you very much.

Well there is defiantly guessing going on, if one were to up the sampling rate by x10 , then one would suddenly have a lot more information to fill in the gaps. Likewise if you compare the original waveform with the sampled one they look different yes ?

If one could "draw" onto an oscilloscope, one could fill in the gaps by hand , and in the case of a simple sine wave would that not be far superior to the series of digital and analog filters used which in effect are trying to do exactly the same thing ???
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
NOS DAC idea... barrows Twisted Pear 6 6th December 2012 12:49 AM
DAC clocking idea preiter Digital Source 4 11th September 2010 12:06 AM
An idea! Fully balanced TDA1541A DAC idea. Please comment! cartman Digital Source 8 17th July 2007 08:18 AM
Anyone can give me idea for DAC of PCM63P-K? audionline Digital Source 16 8th June 2004 02:18 AM


New To Site? Need Help?

All times are GMT. The time now is 09:17 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright 1999-2015 diyAudio

Content Relevant URLs by vBSEO 3.3.2