HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post fully complete "how to" guides and tutorials here. This is a great place to get feedback on stuff you might put in the wiki.

Moderators: MattKingUSA, khz

User avatar
lilith
Established Member
Posts: 1361
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by lilith »

sadko4u wrote:
lilith wrote: And which slope factor to use in the LSP EQ then and what type of EQ (resonance?)?
You need to use special APO DR filters in LSP equalizer to match the REW settings.
Also, Since LSP equalizer provides APO-compatible filters, I'm currently working on this feature request sent to me as a feedback some time ago:

https://github.com/sadko4u/lsp-plugins/ ... #L196-L231
That would be great!

User avatar
CrocoDuck
Established Member
Posts: 1099
Joined: Sat May 05, 2012 6:12 pm
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by CrocoDuck »

lilith wrote:Is it ok to choose the Generic EQ in REW?
Yes
lilith wrote:And which slope factor to use in the LSP EQ then and what type of EQ (resonance?)?

Settings:
I set up the first equalizer in the LSP parametric EQ as in your settings as an example. See below. The slope control does not affect the APO style of filter.
Image
Hope it helps!
Check my Linux audio experiments on my SoundCloud.
Browse my AUR packages.
Fancying a swim in the pond?

User avatar
lilith
Established Member
Posts: 1361
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by lilith »

CrocoDuck wrote:
lilith wrote:Is it ok to choose the Generic EQ in REW?
Yes
lilith wrote:And which slope factor to use in the LSP EQ then and what type of EQ (resonance?)?

Settings:
I set up the first equalizer in the LSP parametric EQ as in your settings as an example. See below. The slope control does not affect the APO style of filter.
Image
Hope it helps!
Thanks! I thought it's not implemented yet. Wow... the difference between e.g. APO and RLC is quite huge. Is it because of different Q definitions?

User avatar
sadko4u
Established Member
Posts: 765
Joined: Mon Sep 28, 2015 9:03 pm

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by sadko4u »

lilith wrote: Thanks! I thought it's not implemented yet. Wow... the difference between e.g. APO and RLC is quite huge. Is it because of different Q definitions?
This is also because of a bit different filter design:
APO - Digital biquad filters derived from canonic analog biquad prototypes digitalized through Bilinear transform. These are textbook filters which are implemented as in the EqualizerAPO software.
https://lsp-plug.in/?page=manuals&secti ... r_x16_mono
LSP (Linux Studio Plugins) Developer and Maintainer.

User avatar
Capoeira
Established Member
Posts: 1250
Joined: Tue May 12, 2009 1:01 pm
Location: Brazil

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by Capoeira »

lilith wrote:
CrocoDuck wrote:
lilith wrote:Is it ok to choose the Generic EQ in REW?
Yes
lilith wrote:And which slope factor to use in the LSP EQ then and what type of EQ (resonance?)?

Settings:
I set up the first equalizer in the LSP parametric EQ as in your settings as an example. See below. The slope control does not affect the APO style of filter.
Image
Hope it helps!
Thanks! I thought it's not implemented yet. Wow... the difference between e.g. APO and RLC is quite huge. Is it because of different Q definitions?
why not use one of the LSP convolvers?

User avatar
lilith
Established Member
Posts: 1361
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by lilith »

Capoeira wrote:
lilith wrote:
CrocoDuck wrote: Yes


I set up the first equalizer in the LSP parametric EQ as in your settings as an example. See below. The slope control does not affect the APO style of filter.
Image
Hope it helps!
Thanks! I thought it's not implemented yet. Wow... the difference between e.g. APO and RLC is quite huge. Is it because of different Q definitions?
why not use one of the LSP convolvers?
How will this work and is there any advantage?

User avatar
lilith
Established Member
Posts: 1361
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by lilith »

@sadko4u: Just to make sure that it's correct, what I'm doing:

When saving the filter.req file from REW I have to choose the GENERIC filter type, right? When importing this into your EQ it automatically jumps to APO.

In here I have to choose PK, I guess (as written a page above from here):

https://i.imgur.com/wCQ17zN.png

User avatar
Capoeira
Established Member
Posts: 1250
Joined: Tue May 12, 2009 1:01 pm
Location: Brazil

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by Capoeira »

lilith wrote:
Capoeira wrote:
lilith wrote:
Thanks! I thought it's not implemented yet. Wow... the difference between e.g. APO and RLC is quite huge. Is it because of different Q definitions?
why not use one of the LSP convolvers?
How will this work and is there any advantage?
there is in option in REW to export the filter IR. it then can be simply put in a convolver in the audio chain.
I actualy don't know which one is better. I just always did it like that. In theory with convolution you can have infinite filters, but REW is limited to 20 anyways

User avatar
lilith
Established Member
Posts: 1361
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by lilith »

Capoeira wrote:
lilith wrote:
Capoeira wrote:
why not use one of the LSP convolvers?
How will this work and is there any advantage?
there is in option in REW to export the filter IR. it then can be simply put in a convolver in the audio chain.
I actualy don't know which one is better. I just always did it like that. In theory with convolution you can have infinite filters, but REW is limited to 20 anyways
Ah, I found it. Hmmm... When using the EQ I know what I'm doing and I see which frequencies are filtered. When using the IR response it's more nebulous, but I will try this too. ~5 filters or less is enough in my case. Thanks for the tip.

User avatar
sadko4u
Established Member
Posts: 765
Joined: Mon Sep 28, 2015 9:03 pm

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by sadko4u »

lilith wrote:@sadko4u: Just to make sure that it's correct, what I'm doing:

When saving the filter.req file from REW I have to choose the GENERIC filter type, right? When importing this into your EQ it automatically jumps to APO.

In here I have to choose PK, I guess (as written a page above from here):

https://i.imgur.com/wCQ17zN.png
APO filters are the most matching filters for the REW file. Other filters are non-DSP-book designed, so their characteristics do not match REW ones. That's why automatically APO filters are selected when you import REW file.

Sorry for delay. I haven't monitored this topic.
APO - Digital biquad filters derived from canonic analog biquad prototypes digitalized through Bilinear transform. These are textbook filters which are implemented as in the EqualizerAPO software.
Reference to textbook: https://shepazu.github.io/Audio-EQ-Cook ... kbook.html
LSP (Linux Studio Plugins) Developer and Maintainer.

User avatar
lilith
Established Member
Posts: 1361
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by lilith »

sadko4u wrote:
lilith wrote:@sadko4u: Just to make sure that it's correct, what I'm doing:

When saving the filter.req file from REW I have to choose the GENERIC filter type, right? When importing this into your EQ it automatically jumps to APO.

In here I have to choose PK, I guess (as written a page above from here):

https://i.imgur.com/wCQ17zN.png
APO filters are the most matching filters for the REW file. Other filters are non-DSP-book designed, so their characteristics do not match REW ones. That's why automatically APO filters are selected when you import REW file.

Sorry for delay. I haven't monitored this topic.
APO - Digital biquad filters derived from canonic analog biquad prototypes digitalized through Bilinear transform. These are textbook filters which are implemented as in the EqualizerAPO software.
Reference to textbook: https://shepazu.github.io/Audio-EQ-Cook ... kbook.html
Thanks, it's working!

User avatar
Capoeira
Established Member
Posts: 1250
Joined: Tue May 12, 2009 1:01 pm
Location: Brazil

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by Capoeira »

sadko4u wrote:
lilith wrote:@sadko4u: Just to make sure that it's correct, what I'm doing:

When saving the filter.req file from REW I have to choose the GENERIC filter type, right? When importing this into your EQ it automatically jumps to APO.

In here I have to choose PK, I guess (as written a page above from here):

https://i.imgur.com/wCQ17zN.png
APO filters are the most matching filters for the REW file. Other filters are non-DSP-book designed, so their characteristics do not match REW ones. That's why automatically APO filters are selected when you import REW file.

Sorry for delay. I haven't monitored this topic.
APO - Digital biquad filters derived from canonic analog biquad prototypes digitalized through Bilinear transform. These are textbook filters which are implemented as in the EqualizerAPO software.
Reference to textbook: https://shepazu.github.io/Audio-EQ-Cook ... kbook.html
whats your opinion on the EQ vs convolver question?

User avatar
sadko4u
Established Member
Posts: 765
Joined: Mon Sep 28, 2015 9:03 pm

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by sadko4u »

Capoeira wrote:whats your opinion on the EQ vs convolver question?
According to the DSP theory, convolution is another way to present the frequency characteristics of the linear system.
But parametrized filters have two huge advantages:
- parametrized filters can be simply edited by adjusting their parameters; to do the same thing with convolution, you need to add additional filter or pre-process the impulse response with this filter.
- if parametrized filters are IIR (infinite impulse response) filters, then they require much less CPU resources than convolution which works as a FIR (finite impulse response) filter.

The advantage of convolution is the following: if you have a linear system that does not need additional post-processing, then loading a single impulse response that represents this system is much easier than building chain of filters with custom parameters.

So these both are the different methods to do the same thing.
LSP (Linux Studio Plugins) Developer and Maintainer.

User avatar
lilith
Established Member
Posts: 1361
Joined: Fri May 27, 2016 11:41 pm
Location: bLACK fOREST
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by lilith »

Maybe another thing: Boosting frequencies is not recommended and does not work in some cases. So, I like the flexibility of EQ filters.

User avatar
CrocoDuck
Established Member
Posts: 1099
Joined: Sat May 05, 2012 6:12 pm
Contact:

Re: HOWTO: Quick Frequency Response Measurements Using JAPA and Faust

Post by CrocoDuck »

Sorry guys, I did not follow the forums too much in the last few months...
lilith wrote:Maybe another thing: Boosting frequencies is not recommended and does not work in some cases. So, I like the flexibility of EQ filters.
Do you mean using an Impulse Response to boost low frequency? Detailed low frequency filtering can be implemented with FIR filters only if they are pretty long. I reckon that's why it has a high chance of not working: if your impulse response is too short you loose low frequency detail.
Check my Linux audio experiments on my SoundCloud.
Browse my AUR packages.
Fancying a swim in the pond?

Post Reply