Minimum and mixed phase is not on the menu yet. But I'll think about it. the import stuff was loads of work but definetaly worth it, as it is very educational.Downloaded installed and testing a bit now
Nice bit of software but very cpu intensive on my virtual machine.
How do we make minimum and mixed filters here?
Unable to import 1021filt.txt files btw. Nothing happens when I push import.
Please make it doable. "My" filters are welcome to end up in any other collection. No worries.
In fact, many filter brewers including Søren has also shared the .txt files. Almost open source which is one of the very positive things about the dam dac. Very good.
In fact, many filter brewers including Søren has also shared the .txt files. Almost open source which is one of the very positive things about the dam dac. Very good.
Last edited:
Ok, the new version in online.
What you can do:
Import every .skr file ,view the filters, mix and match them.
However you can not change imported filters. If you find that they look horrible in the fixed point Magnitude estimation you can not change the gain or anything. But you can make new and better ones ; )
The filters are printed out as txt for directly running them through mkrom.exe
As for the new dam versions:
At the moment you can not make filters for the new 1.4MHz FIR2 stage. But you can import the original from soekris and add your own FIR1 filters.
The Magnitude Response Estimate is valid for the dam1xxx only as the dam2xxx has 40bit FPGA.
The printed txt file is also rounded to match the 32bit dam1xxx but you can still use the app and run it through mkrom.exe at least it shouldn’t overflow.
Enjoy and report
What you can do:
Import every .skr file ,view the filters, mix and match them.
However you can not change imported filters. If you find that they look horrible in the fixed point Magnitude estimation you can not change the gain or anything. But you can make new and better ones ; )
The filters are printed out as txt for directly running them through mkrom.exe
As for the new dam versions:
At the moment you can not make filters for the new 1.4MHz FIR2 stage. But you can import the original from soekris and add your own FIR1 filters.
The Magnitude Response Estimate is valid for the dam1xxx only as the dam2xxx has 40bit FPGA.
The printed txt file is also rounded to match the 32bit dam1xxx but you can still use the app and run it through mkrom.exe at least it shouldn’t overflow.
Enjoy and report
Too little gratitude her.
Thanks alot for this great tool. Makes me wonder if my earlier filters were optimal..
BTW: For instance 44.1khz filters. How much attenuation to Nyquist is needed to make a good filter? Always thought attenuation should be at least 100dB but notice alot of measurements (from ASR) much less attenuation from the "big boys".
Thanks alot for this great tool. Makes me wonder if my earlier filters were optimal..
BTW: For instance 44.1khz filters. How much attenuation to Nyquist is needed to make a good filter? Always thought attenuation should be at least 100dB but notice alot of measurements (from ASR) much less attenuation from the "big boys".
I‘m happy you’re enjoying the app.
I already put together some new filter Functions that will be included in next release - also a matlab function to make quasi linear phase IIR Halfband Filters. I have hopes for this one 😅.
Min phase is also coming and Parks McClellan. But these are more difficult to design right - there’s not much you can do wrong with spectral windows, they usually just work fine.
Well. A Nyqusit Filter has -6dB at Nyquist frequency .. so maby you don’t actually need that much. That’s why I’m personally aiming for short filters.
Had a chance to look at the magnitude response estimate?
Is there someone who would be interested in a possibility to change the output file to round for the 40bit FPGA Dam versions or to not round at all?
I already put together some new filter Functions that will be included in next release - also a matlab function to make quasi linear phase IIR Halfband Filters. I have hopes for this one 😅.
Min phase is also coming and Parks McClellan. But these are more difficult to design right - there’s not much you can do wrong with spectral windows, they usually just work fine.
Well. A Nyqusit Filter has -6dB at Nyquist frequency .. so maby you don’t actually need that much. That’s why I’m personally aiming for short filters.
Had a chance to look at the magnitude response estimate?
Is there someone who would be interested in a possibility to change the output file to round for the 40bit FPGA Dam versions or to not round at all?
Thanks for clarification! Very misleading what I said. I meant Nyquist L-th Band or Halfband filters like these: https://www.mathworks.com/help/dsp/ug/fir-nyquist-l-th-band-filter-design.html
But I’m not the first one to use the term like this 😅
So been toying with this app for some time now. Made some filters and testing Great work.
Since I am resampling everything to 24/94 I only made filters there. Used janho´s filters for a starter. BTW: Unable to really notice any difference between the 4 banks Actually bank 7 (the minimum filter) which I cannot make in this tool seems ever so slightly more "alive" but the nuances are small.
Great work.Since I am resampling everything to 24/94 I only made filters there. Used janho´s filters for a starter. BTW: Unable to really notice any difference between the 4 banks
Actually bank 7 (the minimum filter) which I cannot make in this tool seems ever so slightly more "alive" but the nuances are small.
Last edited:
Very interesting. Thank you for your work on this and your sharing the app with the community 👍🏻Ok, the new version in online.
What you can do:
Import every .skr file ,view the filters, mix and match them.
However you can not change imported filters. If you find that they look horrible in the fixed point Magnitude estimation you can not change the gain or anything. But you can make new and better ones ; )
The filters are printed out as txt for directly running them through mkrom.exe
As for the new dam versions:
At the moment you can not make filters for the new 1.4MHz FIR2 stage. But you can import the original from soekris and add your own FIR1 filters.
The Magnitude Response Estimate is valid for the dam1xxx only as the dam2xxx has 40bit FPGA.
The printed txt file is also rounded to match the 32bit dam1xxx but you can still use the app and run it through mkrom.exe at least it shouldn’t overflow.
Enjoy and report
Ok, have made a "bunch" of different filters to try. brewing, testing and testing again.
My rpi is connected to the dam1021 the whole time so it is easy to upload filters. Just enable the usb ports to connect to the dam, and disable afterwords.
The filters I try is only for 96khz that is since I resample everything i SOX due to a dsp install.
I really have a hard time noticing any difference in SQ at all, except for one thing.
That minimum filter janho12345 supplied in his filterpack give me a sence of more "ambient/airy" sound in the upper octaves.
I have in the past never really cared about minimum filters, only linear and combining them.
I am not sure what causes this but this is what my ears and brain tells me and they can off course not be trusted
The effect is especially aparent on acoustic "audiophile" material, like Sophie Zelmani - "all about you"
Thanks again for this great tool.
My rpi is connected to the dam1021 the whole time so it is easy to upload filters. Just enable the usb ports to connect to the dam, and disable afterwords.
The filters I try is only for 96khz that is since I resample everything i SOX due to a dsp install.
I really have a hard time noticing any difference in SQ at all, except for one thing.
That minimum filter janho12345 supplied in his filterpack give me a sence of more "ambient/airy" sound in the upper octaves.
I have in the past never really cared about minimum filters, only linear and combining them.
I am not sure what causes this but this is what my ears and brain tells me and they can off course not be trusted
The effect is especially aparent on acoustic "audiophile" material, like Sophie Zelmani - "all about you"
Thanks again for this great tool.
Last edited:
Hi,
the new Version on online now. still the same link, but here it is again: https://jp-hoffmann.de/Files/DamFilterDesigner2_3.zip
Featuring:
-Export of Filters as .skr file
-Fixed DSD Magnitude Response estimate - but for DSD still problematic because these are 1bit data streams and there is no multiplier involved but 4 parallel adders.. I can not simulate this properly in matlab so this is still an estimate, but works much better now!
-More design techniques besides the window approach that include Mixed and Minimum Phase Filters. But the matlab functions I'm using do not like high coefficient counts when designing mixed or minimum phase filters i think maximum here is around 200. But keep in mind, a minimum phase filter does not need as many coefficients as a linear phase filter for same sopband attenuation..
-There is now a feature to upload filters to the dam! either with usb->rs232 serial connection or via telnet!
That means wireless upload of filters to use this feature you need an esp32 that is connected to the dam and my AsyncTelnetSerial library for esp32 that is available on my github account https://github.com/janphoffmann/AsyncTelnetSerial unfortunately I could not test these with a dam - so I will need feedback on this from you!
the new Version on online now. still the same link, but here it is again: https://jp-hoffmann.de/Files/DamFilterDesigner2_3.zip
Featuring:
-Export of Filters as .skr file
-Fixed DSD Magnitude Response estimate - but for DSD still problematic because these are 1bit data streams and there is no multiplier involved but 4 parallel adders.. I can not simulate this properly in matlab so this is still an estimate, but works much better now!
-More design techniques besides the window approach that include Mixed and Minimum Phase Filters. But the matlab functions I'm using do not like high coefficient counts when designing mixed or minimum phase filters i think maximum here is around 200. But keep in mind, a minimum phase filter does not need as many coefficients as a linear phase filter for same sopband attenuation..
-There is now a feature to upload filters to the dam! either with usb->rs232 serial connection or via telnet!
That means wireless upload of filters
to use this feature you need an esp32 that is connected to the dam and my AsyncTelnetSerial library for esp32 that is available on my github account https://github.com/janphoffmann/AsyncTelnetSerial unfortunately I could not test these with a dam - so I will need feedback on this from you!Trying to make som minimal og mixed mode filters just hang when testing here. Have to end the process.
Also, text seems kind of funny
Running win11 as a virtual machine from my macbook. That is win11 arm64 version. Could be a problem?
Maybe I can try on another instance? Have some more virtual machines running on my server.
Also, text seems kind of funny
Running win11 as a virtual machine from my macbook. That is win11 arm64 version. Could be a problem?
Maybe I can try on another instance? Have some more virtual machines running on my server.
