Filter brewing for the Soekris R2R

[wushuliu]

Hmm, yes I think I incorporated a bit of inter-over protection. Should be 3,1 dB really in that case. It was discussed before I remember... Need to check back on this - I feel a bit uncertain as I don't recall clearly... :-/

//

Pretty sure it was 3 as well. No need to check. I just wanted to confirm this is still the case. Thanks!

 

[TNT]

It was 3,1

//

 

[TNT]

Curious what you might think about this brew. I have tested the file and it plays. It's with DC filter.

Entering:

# filters

should give:

04 Linear Phase TNTmode, 44.1 Ksps,
08 Linear Phase Soekris FIR2, 384 Ksps, 111 Khz -1 db, 192 Khz -130 db
29 DC Blocking IIR, 352.8 Ksps, 2 Hz HP 1st order
30 Deemphasis IIR, 352.8 Khz Samplerate, hiself f=5600 Hz, Q=0.485, gain=-10.1

Appreciate if you try it and report back of any impressions.

//

 

[SebastianL]

Curious what you might think about this brew. I have tested the file and it plays. It's with DC filter.

Entering:

# filters

should give:

04 Linear Phase TNTmode, 44.1 Ksps,
08 Linear Phase Soekris FIR2, 384 Ksps, 111 Khz -1 db, 192 Khz -130 db
29 DC Blocking IIR, 352.8 Ksps, 2 Hz HP 1st order
30 Deemphasis IIR, 352.8 Khz Samplerate, hiself f=5600 Hz, Q=0.485, gain=-10.1

Appreciate if you try it and report back of any impressions.

//

Is this a different version of your "DIYA_1021filt_DC_v2_DC Blocking" from April 2019?

 

[SebastianL]

Appreciate if you try it and report back of any impressions.

//

While still not having had the time to do a thorough listening test my only preliminary observations are regarding the perceived volume (exactly the same as the earlier TNT filter version).

@44.1Khz: F4 lower volume than F5,F6 and F7
@48Khz: F4 lower volume than F5,F6 and F7
@88Khz: F4 lower volume than F5 and F6 while F7 (NOS) has a lot of attenuation!
@96Khz: F4 lower volume than F5 and F6 while F7 (NOS) has a lot of attenuation!

Apart from that the F4 linear sounds good at a first listen.

 

[TNT]

Yes a different versions taking the "concept" even further.

//

 

[TNT]

NB: My change/contribution is only F4/44,1.

//

 

[TNT]

Wish one could tie biquads to a filter bank. Like one set for F4 and an other F5 etc...




//

 

[NETT1969]

Having issues with SPDIF input after upgrading filter pack

Hello all,

I recently joined the diyAudio community looking for information on the Soekris dam 1021 unit.

Thanks to several threats here I was finally able to get the proper USB/RS232 cable connected to the dam 1021. Right after that I uploaded the NOS filter published by Spzzzzkt back in 2015 but after power cycling the unit the SPDIF signal doesn't lock anymore.

I reverted to the original Soekris filter and the signal locks OK again.

Has anyone seen this problem before?

Thanks.

 

[analog_sa]

As this has nothing at all to do with filter brewing you should perhaps ask in the DAM thread. My guess would be either a corrupt file or operator error.

 

[NETT1969]

Thanks for the note. I will check in the other thread as suggested.

 

[Spartacus]

Curious what you might think about this brew. I have tested the file and it plays. It's with DC filter.

Entering:

# filters

should give:

04 Linear Phase TNTmode, 44.1 Ksps,
08 Linear Phase Soekris FIR2, 384 Ksps, 111 Khz -1 db, 192 Khz -130 db
29 DC Blocking IIR, 352.8 Ksps, 2 Hz HP 1st order
30 Deemphasis IIR, 352.8 Khz Samplerate, hiself f=5600 Hz, Q=0.485, gain=-10.1

Appreciate if you try it and report back of any impressions.

//

Is this compatible with DAM1121? If so, would you be willing to make a version with no attenuation? I use one of your older filters which I enjoy, but would like full gain.

 

[TNT]

Yes it was developed using 1121. How much attenuation do you experience? 1 dB, 2?

I remember that Sören changed the gain structure in the filters at one time and compensated for that in the filter multiplication factor. As I recall it, the released filters used 7.999999995 as the factor while mine used 8. I'm would be surprised if the difference is due to that.

But I also incorporated an attenuation in the filters itself to accommodate "inter-over" protection. I think that has a part in the sound quality itself. Don't you have a couple of dB to spare in your rig?

//

 

[Spartacus]

Great. I'm not sure how much attenuation the current filter has, but it is somewhat quieter than the last filter I was using. It's not a big deal, it's just that I'm currently doing attenuation in software, and it would be more ideal to have one calculation rather than two, although in the real world it will probably make no difference.

Thanks.

 

[projectwhite]

IME a filter which minimises pre-ringing while still fully suppressing OOB aliases is the best general purpose filter that can be hoped for for 44.1. You can get better dynamic performance by eliminating ringing (pre, or both pre and post) but at the cost of timbre, due to aliasing.

I was glad to see that the ESS 9038 chips finally have this type of filter implemented, which they call Corrected Minimum Phase Fast.

I can't wait til digital music moves to a higher sampling rate so we can finally stop having to deal with filter tradeoffs.

 

[kenneth_obee]

Curious what you might think about this brew. I have tested the file and it plays. It's with DC filter.

Entering:

# filters

should give:

04 Linear Phase TNTmode, 44.1 Ksps,
08 Linear Phase Soekris FIR2, 384 Ksps, 111 Khz -1 db, 192 Khz -130 db
29 DC Blocking IIR, 352.8 Ksps, 2 Hz HP 1st order
30 Deemphasis IIR, 352.8 Khz Samplerate, hiself f=5600 Hz, Q=0.485, gain=-10.1

Appreciate if you try it and report back of any impressions.

//

Sorry TNT, may I know what's the meaning of "DC filter"?
I have loaded the DIYA_1021filt_DC_v2.skr.zip and that's impressive. I wonder if anyone could tell the difference of that v2 versus this one.

 

[fedde]

The DC filter is an IIR filter that removes DC components in the audio. It is meant to protect the audio equipment fir the (imho unlikely) situation that the audio gets stuck in the PC on a high level. But the filter is audible, so I do not use it.

Fedde

 

[kenneth_obee]

The DC filter is an IIR filter that removes DC components in the audio. It is meant to protect the audio equipment fir the (imho unlikely) situation that the audio gets stuck in the PC on a high level. But the filter is audible, so I do not use it.

Fedde

Thanks a lot!

 

[TNT]

.....

I can't wait til digital music moves to a higher sampling rate so we can finally stop having to deal with filter tradeoffs.

The problem is that the DAC providers will never utilise the greater BW for relaxed filter requirement but rather chase perfect analog performance figures like maximum BW. Like tuning the -3dB analog output for FS=192 khz to 50 kHz but will state that it gors to 96kHz, -0,01 dB.

This is what the "TNTmode" filter is about really. And I think it works and so do some of you apparently.

//

 

[ynmichael]

Someone should try Foobar SSRC linear filter in x8 upsampling with very slow roll-off in ABX. I bet money that it makes a huge difference. Will try to post a more elaborate answer later.