Latest Thread Images

20231007_135719.jpg
BNC cables.JPG
s-l400-1.jpg
430055964_930277665502616_6955730380570727475_n.jpg
20230507_182323.jpg
IMG_4284.jpeg
Screenshot_20240309-190756_Samsung Internet.jpg 20231117_092737.jpg IMG_0965.jpeg F8B0C1F6-051F-4A11-A233-DD38DCC0A3D6.jpeg PXL_20240305_210208075.PORTRAIT.jpg 20231111_154656.jpg 20231117_092631.jpg IMG_0065.jpeg IMG_0066.jpeg 20230513_145127.jpg 1708074082619.png
See full thread gallery...
CHORD ELECTRONICS DAVE
Jun 27, 2015 at 5:46 AM Post #137 of 25,845
The H
rob watts said:
  Yes the half band filter has much poorer timing of transients accuracy - it is in part due to the poor suppression of aliasing images as 22.0501 kHz is -6dB down, whereas an ideal infinite tap length FIR sinc filter would have infinite attenuation at 22.0501 kHz, and these errors degrade the time domain accuracy which manifests itself as poorer transient timing accuracy. But its much more complex than this!
 
The WTA filter is as you call full band filter - but as the tap length increases, it converges onto an ideal sinc filter.
 
The 16 bits refer to the accuracy of the coefficients against the ideal sinc function - so if we want to get an impulse response that has 16 bit accurate coefficients (that is the truncation of the coefficients occurs when the coefficient is below 16 bit in level), then you need about 1M taps. But what level is audible? Hugo is about 12 bits accurate against the ideal, Dave is about 14 bits - and there is a substantial difference in SQ between the two. But at what point does increasing tap length make no difference? Nobody knows the answer to that, but I will find out! There is also the issue of oversampling level, but that's another story...
 
Rob 
Click to expand...


Dear Mr Watts,  what does 12 bits accurate against the ideal mean? 
 



The ideal impulse response, the sinc function gradually converges to zero, so the future and past samples no longer have a significant effect. With the 12 bits accuracy, then one can say that the WTA impulse response is roughly the same as the ideal sinc function to a 12 bit accuracy, as the coefficients are now smaller than 12 bits. There will come a point when the coefficients are so small they no longer have an audible effect, and I certainly have not hit that point yet. 
 
Rob
 
Jun 27, 2015 at 6:49 AM Post #139 of 25,845
sapientiam said:
  I see in the specs you mentioned 'no measurable noise floor modulation' - how did you do the measurement and with what stimulus?
Click to expand...

Good question. Noise floor modulation is very difficult to measure, as ADC's (even the very best) have very significant noise floor modulation. Firstly, what is noise floor modulation? When a sine wave signal is used in a DAC, you get different types of distortion - harmonic distortion (distortion of integer multiples from the sine wave fundamental) enharmonic distortion (distortion products that are non integer) and changes to the noise level. So for example you may have a DAC that produces noise at -120dB with -60dB sine wave (traditional dynamic range test) but the noise with a 0dB sine wave maybe -115dB - thus the noise has increased by reproducing a higher level sine wave - in this case the noise floor (seen by doing an FFT measurement) would increase by 5dB.
 
Now noise floor modulation is highly audible - it interferes with the brain's processing of data from the ear - and immeasurably small levels of noise floor modulation is audible. I know this as I have listened to noise floor modulation at around -200dB - these numbers are derived from simulation - and heard the effect when the noise floor modulation mechanism was switched on and off.
 
The problem with noise floor modulation is measuring it, as ADC performance is far worse than Hugo's, and certainly worse than Dave's. With my old Audio Precision (AP) I used to use a fixed frequency passive notch filter to remove the fundamental, then fed the residual into the AP. From this one could determine noise floor modulation, but the AP was not good at resolving small noise floors. But around October 2014 AP launched the APX555, and this had a clever system to enable more accurate measurements of distortion and noise floor. What this instrument does is the employ two ADC's per channel, and an automatic notch filter, so one ADC uses notched out fundamental, and another ADC for the fundamental. The instrument then stitches the two plots together in the digital domain.
 
It also had a very high purity analogue oscillator - the system has residual THD at 2.5v of -150dB. Since I need a high purity analogue source to test the pro ADC project, and since Dave at that time exceeded the old AP measurement capacity, once AP launched the APX555 I went out and purchased one.
 
So we can now see the performance of Dave using the APX555:
 
 

 
Here we have a 1kHz signal at 2.5v RMS and distortion is below -150dB (blue trace). The no signal is the red trace, and you can see that the noise floor is identical in both cases, close to -180dB. 0dB is at 6v RMS, the maximum output of Dave. Clearly, there is no measurable noise floor modulation at all, even with the noise floor close to -180dB.
 
Rob
 
Jun 27, 2015 at 7:25 AM Post #141 of 25,845
From memory it was 1.2M FFT length, 48kHz sample rate. Total noise is about -130dB A wt. The benefit of resolving -180dB noise floor is that it exposes small errors that are normally buried well within the noise floor.
 
Rob
 
Jun 29, 2015 at 11:35 AM Post #146 of 25,845
The most impressive statistic to me is that Rob stated (in a video interview) that he was able to continue to lower DAVE's digital section S/N until he achieved an amazing -350dB, and he continued to hear improvements in depth of the reproduced signal until he reached that point. Even though the analog section does not meet that spec, he was still able to hear the improvement in the digital section by realizing that amazing spec. This just shows the levels that Rob goes to in order to extract the finest sound quality, even though typical logic might have prevented other engineers from even attempting such low levels of noise as to be virtually unmeasurable by conventional means. I'm counting the days until DAVE becomes available!
 
Jun 29, 2015 at 4:26 PM Post #147 of 25,845
audiobill said:
  The most impressive statistic to me is that Rob stated (in a video interview) that he was able to continue to lower DAVE's digital section S/N until he achieved an amazing -350dB, and he continued to hear improvements in depth of the reproduced signal until he reached that point. Even though the analog section does not meet that spec, he was still able to hear the improvement in the digital section by realizing that amazing spec. This just shows the levels that Rob goes to in order to extract the finest sound quality, even though typical logic might have prevented other engineers from even attempting such low levels of noise as to be virtually unmeasurable by conventional means. I'm counting the days until DAVE becomes available!
Click to expand...

 
Yes, amazing!
 
Jun 30, 2015 at 11:22 PM Post #150 of 25,845
Dave has such a wide OP level and ultra low noise, the gain of the power amp is unimportant.
 
Dave is SE as that's the best for SQ; but to do so is much harder, as differential gives measured benefits that hides SQ problems. Going SE means that you no longer hide the problems.
 
Rob
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 15 (members: 0, guests: 15)
Back
Top