It wasn’t planned to use it that way so F7 at 352.8kHz is clipping – magnitude should be corrected to match on all frequency. Apart of that do you have any comments? At the end all filters goes thru IIR. If you have any interest in continue to test this kind of filter and have will to resign from true NOS it is possible to build set with much higher gain.
Let me tell you, I'll stay with the 1021filt77 filter a little longer. That sounds very interesting. I used upsampling in the roon Labs application.
It is a pity that the 44.1 does not sound well, because I have a CD transport and you cannot do upsampling there.
However, I cope in a different way.
I have to listen to different musical genres longer. In some time I will report any comments.
Remember that I judge the sound "by ear" and not from measurements.
It is a pity that the 44.1 does not sound well, because I have a CD transport and you cannot do upsampling there.
However, I cope in a different way.
I have to listen to different musical genres longer. In some time I will report any comments.
Remember that I judge the sound "by ear" and not from measurements.
Last edited:
I think it can happen in very rare occasions when you have a faulty file storing the PCM data or a really wacked up recording. They weren't there in the initial release if I recall correctly. Many seem to run without them with no problem. But for sure, a normal CD has no DC out. As it is a digital filter it has absolutely nothing to do with DC out of the hardware when no song is playing.
//
//
Because I didn’t quite enjoy the Soekris filters, I’m upsampling everything to 352/384 outside the DAC. But the DC filter is still in play, correct?
By removing the IIR filter, the DC filter is removed, am I right?
How do I remove the DC filter?
Do I need to unpack the 1.21 ROM into a text file, edit it, use mkrom to generate the binary, then flash it back? Need some guidance here...
I suppose the DC filter is useless in preventing any DC offset generated by the buffer stages...and possibly modify the data stream unnecessarily...
By removing the IIR filter, the DC filter is removed, am I right?
How do I remove the DC filter?
Do I need to unpack the 1.21 ROM into a text file, edit it, use mkrom to generate the binary, then flash it back? Need some guidance here...
I suppose the DC filter is useless in preventing any DC offset generated by the buffer stages...and possibly modify the data stream unnecessarily...
Last edited:
I was asking about my filter where anti-aliasing is done by cascaded IIR filters so there is plenty of them.
DC filter is high pass filter. In fact it’s not needed from perspective of DC protection if at last one IIR of any type exist, by design IIR use feedback so never goes to zero.
As HP filter it has quite big influence on sound which manifest for example as slope on square wave test or steep response shape.
Tonight I switched from F5 back to F7 / "NewNOS++" and immediately caught on again. I'm just loving the dynamics and grit of this filter. So it's decided for me: F7 over F5. @gumisb happy to test any further improvements you might cook up. Great job so far.
Headphones / pre-amp stage is a b22 with very high bandwidth (0-2,5 MHz +0/-3 dB). In that respect very different from tubes. So do keep 'em coming in the "vanilla flavour" !
Headphones / pre-amp stage is a b22 with very high bandwidth (0-2,5 MHz +0/-3 dB). In that respect very different from tubes. So do keep 'em coming in the "vanilla flavour" !
" I set filter to F7.Not possible with the current version of the DAC firmware. Its unpredictable drift (now officially named "wander") prevents anything approaching precise measurements with DeltaWave. Don't ask me how many times I've tried this...
New IIR anti-aliasing/NOS upsampling filters set for evaluation
What’s new
- Filters are divided into two groups F4,F5 with gain ca.-10dB and F6,F7 with gain ca.-1dB if we compare to stock firmware. (Increasing gain by V+ command above 0 introduce clipping on both groups, so better stay below or at 0dB.) So this is important to pay attention when changing filters.
- F4 filter is new stack of NOS and smoothed NOS evolution filter introduced earlier as F5 filter, now used as FIR2.
- F5 filter is “my classic” stack. (only at 44.1/48kHz)
- F6 filter is C128dp as FIR1 and NewNOS as FIR2 stack (only at 44.1/48kHz)
- F7 filter is previous NOS/NewNOS stack called by @roderickvd NewNOS++.
- All input frequency over 44.1/48kHz are served on FIR1 stage by gain matched NOS filters, identical no matter what filter was chosen, final gain will differ and FIR2 filter according to group.
- New IIR all pass filter tuned differently. Should match most expectations but we will see.
What’s new
- Filters are divided into two groups F4,F5 with gain ca.-10dB and F6,F7 with gain ca.-1dB if we compare to stock firmware. (Increasing gain by V+ command above 0 introduce clipping on both groups, so better stay below or at 0dB.) So this is important to pay attention when changing filters.
- F4 filter is new stack of NOS and smoothed NOS evolution filter introduced earlier as F5 filter, now used as FIR2.
- F5 filter is “my classic” stack. (only at 44.1/48kHz)
- F6 filter is C128dp as FIR1 and NewNOS as FIR2 stack (only at 44.1/48kHz)
- F7 filter is previous NOS/NewNOS stack called by @roderickvd NewNOS++.
- All input frequency over 44.1/48kHz are served on FIR1 stage by gain matched NOS filters, identical no matter what filter was chosen, final gain will differ and FIR2 filter according to group.
- New IIR all pass filter tuned differently. Should match most expectations but we will see.