Filter brewing for the Soekris R2R

Jump to Latest
Ken Newton

Ken Newton

Member
Joined 2003
spzzzzkt said:
Currently putting together a blog to house future filter developments.
I'll post up a link once I have something that is worth looking at...
Click to expand...

Hi, spzzzzkt,

I applaud all of the custom reconstruction filter evaluation work you are doing. While I don't have a DAM, I am interested in the area digital audio reconstruction filters. I wonder if you've yet evaluated a brickwall SINC filter having a sub-Nyquist stop-band? For example, with a stop-band frequency fully beginning by 20kHz?

The objective of such an in-band cutoff is to remove any alias products encoded into the audio band as a result of the violation of Nyquist which is permitted by A/D half-band anti-alias filters widely in use. Such filters feature a stop-band which typically begins approximately at 0.546 * Fs = 24.1kHz, for CD. Which then means that any orginal signal content from 22.050kHz through 24.1kHz will alias in-band to 20kHz through 22.050kHz. While such HF aliases are ultrasonic and so are themselves very likely inaudible, Bruno Putzeys claims that they audibly blur timing cues, especially with instruments producing harmonics above Nyquist. At playback reconstruction, even an Nyquist stop-band frequency will not fully remove the potentially aliased band.

If this intrigues you, I would very much be interested in reading the subjective listening evaluations of the DAM experimenter community regarding a SINC filter that's fully into it's stop-band by 20kHz. Of course, wth such a lowered stop-band it would be desirable to have a filter transition-band as narrow as can ne made possible by the number of FIR taps available via the FPGA hardware. Hence, the preference for a SINC brickwall response.
 
Last edited:
oneoclock

oneoclock

Member
Joined 2010
Ken Newton said:
Hi, spzzzzkt,

I applaud all of the custom reconstruction filter evaluation work you are doing. While I don't have a DAM, I am interested in the area digital audio reconstruction filters. I wonder if you've yet evaluated a brickwall SINC filter having a sub-Nyquist stop-band? For example, with a stop-band frequency fully beginning by 20kHz?
Click to expand...

spzzzzkt stopped intervening in this thread. He continued in moreDAMfilters

Last filters made do what you indicate:

"EQHQv5, A remake of the EQHQv2. I've taken advantage of the longer passband filters so the edge is now above 20 kHz, and stop band edge is at 22.5kHz."

An externally hosted image should be here but it was not working when we last tested it.
 
Ken Newton

Ken Newton

Member
Joined 2003
oneoclock said:
spzzzzkt stopped intervening in this thread. He continued in moreDAMfilters

Last filters made do what you indicate:

"EQHQv5, A remake of the EQHQv2. I've taken advantage of the longer passband filters so the edge is now above 20 kHz, and stop band edge is at 22.5kHz."

An externally hosted image should be here but it was not working when we last tested it.
Click to expand...

Hi, oneoclock,

Thanks, for repsonding. Judging only from the excellent chart you provided, and looking only at the blue 44.1kHz curve, it doesn't appear as though the associated filter is performing as I was describing. I intend a low-pass filter that's fully at it's stop-band by 20kHz. Meaning, if the filter has 100dB of stop-band rejection, then a 20kHz tone would be suppressed fully by 100dB. Yes, a bit of the pass-band upper edge necessarily gets sacrificed, the impact of which is minimized by narrowing the transisition-band. However, please do correct me if I'm misinterpreting your chart. Thanks!
 
Last edited:
vdi_nenna

vdi_nenna

Member
Joined 2000
Paid Member
Do the filters have to match the current version of the soekris firmware?

Let me rephrase this because I can see that it will only generate more questions.

I'm using the built-in Soekris filters. I listen to all 4 filters and settled on Soft or F7.
I wanted to see if anyone has made or modified filters that will work with firmware 1.06.
 
Last edited:
vdi_nenna

vdi_nenna

Member
Joined 2000
Paid Member
I couldn't wait to hear them, so I plugged in the filters tonight.
I wanted to share my impressions of them while they are still fresh in my head. Please understand that these are my impressions and someone else may have other results.

I liked them in order of F4, F5, F6 and F7. They all had amazing vibrancy to them. Everything was as big as can be. Great detail with all 4 types. Very powerful sound.

They all had a phaser effect quality to them in the upper mids and highs. F7 was virtually un-listenable because vocals sounded as if there were a 5 inch cylinder in front of the vocalist. The effect was reduce with F6, F5 and F4, respectively. There was also less separation between instruments. This may be because everything sounded so big and powerful. Maybe too big for my room.

I loaded the the Soekris 1.06 filters and thing were back to normal. I tried F6 again and I think I kind of like it. It has finer high frequency detail.

Again, these are my impression. Your mileage may vary.

Thank you for sharing the filters. Great work on a difficult topic.

Vince
 
oneoclock

oneoclock

Member
Joined 2010
A

arteom

Member
Joined 2016
Hello,

I seem to have gotten myself into a jam and not too sure how to get out, hoping someone here can help. As stated in a previous post, I tried to send the DAC older filter after upgrading to the latest firmware version. Now the DAC seems to be stuck. I only have SPDIF input, and when connected the device just returns "L044", if I power it down and back up it shows -

R1.06
I3
L000
F5
v+00
I3

but the UI manager does not load. I was hoping to get it to that point and then try to again update firmware, but no such luck.

I am fairly new to all this, have mostly been following guides posted. Am using PuTTY to communicate with the DAC. Anyone have any ideas on how I can get myself out of this and get the spdif input working again?

Thanks,
 
You must log in or register to reply here.
Top Bottom