
Home  Forums  Rules  Articles  diyAudio Store  Blogs  Gallery  Wiki  Register  Donations  FAQ  Calendar  Mark Forums Read 
Digital Line Level DACs, Digital Crossovers, Equalizers, 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 

Thread Tools 
18th February 2021, 08:28 AM  #291 
diyAudio Member
Join Date: Mar 2005
Location: Blaricum


18th February 2021, 09:05 AM  #292 
diyAudio Member
Join Date: Mar 2005
Location: Blaricum

I also did the test with the two time shifted sample streams, because I had my doubts about the removed first mirror at 24.1Khz, and indeed, it's not the case.
See image below of two added samples time shifted 10usec sampled at 50Khz. The 30Khz first image is still there, alive and kicking. If you prefer, I could repeat this with 44.1k instead of 50K. Hans . 
18th February 2021, 10:04 AM  #293 
diyAudio Member
Join Date: Mar 2005
Location: Blaricum

I tried to fit a Sinc envelope around the two time shifted signals.
On one hand the original one at Fs 50Khz should still be there and at the same time since the sample width has halved, a Sinc with it's first zero at 100Khz was expected. Well, it seems to be a bit of both, as becomes clear in the image below for the two 10usec time shifted and added samples. The Sinc with it's first zero at 50kHz touches the spectrum at 20K, 80K, 120K, 180K etc, while the Sinc with it's first zero at 100Khz, touches at 30k, 70K, 130K, 170K etc. A very complex envelope ! The right image is for exactly the same sample but now on it's own without a second time shifted sample. In this case the Sinc with its first zero at 50Khz fits perfectly. Hans Last edited by Hans Polak; 18th February 2021 at 10:06 AM. 
18th February 2021, 10:28 AM  #294 
Account disabled at member's request
Join Date: Dec 2015

I am referring to post 258 non oversampling dac to be a hoax that they could by two DACs shift the 24khz image to much higher frequencies. Your way is always by filtering the image 30khz by a brick wall and not by shifting.

18th February 2021, 11:16 AM  #295 
Account disabled at member's request
Join Date: Dec 2015

I am looking a way to interpolate digitally the midpoint between two samples by sinc/Hamming window. for 20khz, I have 10 cycles for 1 beat 2khz, as each cycle is 3 samples, I need 30 samples to interpolate the midpoints. It will output a true 2xoversampled. Explained here http://www.audiomisc.co.uk/ArchiveMa...Upsampling.pdf
As sinc function is symmetrical, I need 15 constants representing sinx/x * Hamming variable to apply on each sample and sum to obtain the value of the midpoint. Maybe such interpolator does already exist. Last edited by U130421; 18th February 2021 at 11:22 AM. 
18th February 2021, 12:13 PM  #296 
diyAudio Member
Join Date: Mar 2005
Location: Blaricum


18th February 2021, 01:18 PM  #297 
diyAudio Member
Join Date: Mar 2005
Location: Blaricum

I’ll have to read the article, but sinc convolution in the time domain is the same as brick wall in the frequency domain.
In fact what I saw in a quick glimpse is exacly like the Fir filter that I used. But I increased Fs a factor 4. Hans 
18th February 2021, 01:40 PM  #298 
diyAudio Member
Join Date: Mar 2005
Location: Blaricum

Here are at the left the filter coefficients of my FIR, and when taking it's FFT , you get the FR.
Hans P.S. I used two in series to get twice the damping. Not the best way, but good enough for the test. Last edited by Hans Polak; 18th February 2021 at 01:44 PM. 
18th February 2021, 03:27 PM  #299  
diyAudio Member
Join Date: Mar 2003
Location: Haarlem, the Netherlands

Quote:


18th February 2021, 03:51 PM  #300 
diyAudio Member
Join Date: Mar 2005
Location: Blaricum

Yes, Marcel you ate right.
The guy inserted one zero after each sample instead of the three that I did. For some strange reason he did not mention this zero stuffing. So indeed, it is exactly what I did in the previous simulations. How to find the coefficients for your FIR filter is up to the designer, but it has to be long enough and the calculation has to be accurate. What I mentioned earlier is that you will need a fast DSP to do the math. Hans 
Thread Tools  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Philips tda1305 (hybrid multibit/bitstream DAC with interpolation)  1z2a3y  Digital Line Level  37  22nd January 2021 12:02 AM 
High quality parts used deltasigma dac vs cheap parts used r2r ladder dac  Esotar01  Digital Line Level  2  6th May 2016 12:50 AM 
Question about DAC interpolation  cuallito  Digital Line Level  1  28th April 2011 11:11 PM 
Deltacurrent linearinterpolation DAC  wimms  Digital Source  86  4th November 2004 02:05 AM 
New To Site?  Need Help? 