Filter brewing for the Soekris R2R

[avr300]

Is there any volunteer that can test run my file before forum upload?

Yes

 

[TNT]

Yes

PM!

//

 

[TNT]

Seems like avr300 vent to bed 🙂

Here is the file along with the mkrom run.

No DC filter!

No warranties as have not had time to test the .skr myself.

//

 

[TNT]

Maybe it should read "Paul" in there also to honour the designer of some of the filters...

//

 

[kray]

Seems like avr300 vent to bed 🙂

Here is the file along with the mkrom run.

No DC filter!

No warranties as have not had time to test the .skr myself.

//

I'll give it a go! give me 30-45 min to confirm working

 

[kray]

Updated and working fine.

 

[kray]

@tnt

Been listing to your filter most of the night and easily my favorite so far. Great job! Looking forward to you getting it working with 88/96/192

 

[TNT]

Good to hear!!

Thanks for the report.

//

 

[avr300]

Seems like avr300 vent to bed 🙂

He did 😀

 

[analog_sa]

No DC filter!

No warranties as have not had time to test the .skr myself.

//

Listened briefly to F4 and F6. Did not like F4 much. Overdetailed but too polite for my taste. Weird tonality. F6 has more gain and even at -3db still clips on one side. Looks like at -5db it is fine.

Some quick and dirty measurements.

F4 F6 F7

 

[kray]

Listened briefly to F4 and F6. Did not like F4 much. Overdetailed but too polite for my taste. Weird tonality. F6 has more gain and even at -3db still clips on one side. Looks like at -5db it is fine.

Some quick and dirty measurements.

F4 F6 F7

Interesting. I guess I find F4 more pleasing because I have a overall warmer system and enjoy the extra detail. I've never found it fatiguing
I found F5,F6, F7 to all be a little hot on gain, wouldn't mind those reduced.

Could you compare F5 with graph?

 

[analog_sa]

The impulse response looks identical as the F4, apart from the higher gain.

 

[TNT]

What is that stimuli?

//

 

[analog_sa]

The shortest possible pulse for the sampling rate. I guess 22uS for these graphs. Unless it's twice that 🙂

As a comparison, here is Paul's filter i have been listening to for the last 6 months. And not because it's pretty.

 

[TNT]

Just wondered if it something that could come through a red-book CD played on a CD player via a s/pdif connection?

//

 

[analog_sa]

I use Wavegene >usb> XMOS > i2s DAM.

Don't see a reason why a similar stimulus cannot come off a CD via spdif.

 

[TNT]

The shortest "pulse" at 16/44,1 is what comes out of an ADC when fed with 22,049 khz sinus with a sharp 22,05 filter. No dv/dt faster than that of a sine at fs/2 need to be handled by the DAC - feeding it something else is not relevant.

This means that one should not create, in digital space, a sequence of 16 bit words that could not be produced by an ideal ADC (filtered according to the sampling theorem) fed by analogue input (of any sort really).

//

 

[kray]

The shortest possible pulse for the sampling rate. I guess 22uS for these graphs. Unless it's twice that 🙂

As a comparison, here is Paul's filter i have been listening to for the last 6 months. And not because it's pretty.

Which of Pauls filter is that?

The last pack TNT posted was his filter at F4 then F5,F6,F7 were all Pauls filters

(F5 mixed ) EQHQ_lpbr_b4 Linear Phase filter, Paul.
(F6 minimum ) "quasiNOS" C128dp Linear Phase filter, Paul.
(F7 soft ) New NOS filter, Paul.

 

[analog_sa]

Sorry, forgot to mention this: it is the C128. Not sure if porting the filters had an effect on gain or anything else. The original C128 seems to avoid clipping on firmware 119, when volume is set to -2db

The picture below illustrates the outcome at 0db. One would assume that while porting the old filters one would take care of the gain to avoid clipping.

 

[TNT]

Porting? Just to make it clear from my side... I changed nothing in other sections than 04 44,1.

//