FRD Response Blender and Minimum Phase Extractor is now available

[charlielaub]

I finally posted the FRD Response Blender and Minimum Phase Extractor spreadsheet on my web site, on the software page:

http://audio.claub.net/software.html

I've also created a tutorial with a couple of examples (as usage instruction). Since I haven't created a web page for the Response Blender, I'm posting a link to the tutorial here:
FRD_Blender/FRD_Response_Blender_Tutorial_v1.0.pdf

This is an invaluable tool for creating phase-accurate FRD files - especially useful if you are planning to design crossovers with them!

-Charlie

[[email protected]]

Re: FRD Response Blender and Minimum Phase Extractor is now available

excellent charlie!

[Jeff B.]

Re: FRD Response Blender and Minimum Phase Extractor is now available

I am quite impressed with Charlie's implementation of Excel in this file. His smooth blending of the frequency response for near and far field response data is much nicer than my hard splice, and his method of minimum phase extraction is quite novel as well. This file, while it may have a bit of a learning curve, is a nice tool to have in the arsenal.

Jeff

I finally posted the FRD Response Blender and Minimum Phase Extractor spreadsheet on my web site, on the software page:

http://audio.claub.net/software.html

I've also created a tutorial with a couple of examples (as usage instruction). Since I haven't created a web page for the Response Blender, I'm posting a link to the tutorial here:
FRD_Blender/FRD_Response_Blender_Tutorial_v1.0.pdf

This is an invaluable tool for creating phase-accurate FRD files - especially useful if you are planning to design crossovers with them!

-Charlie

[dcibel]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Thanks, I'll give this a shot later tonight. Previously I've used Frequency Response Combiner, which works well and will apply some smoothing to the transition, but it quite cumbersome and fiddly to use.

[charlielaub]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Thanks, I'll give this a shot later tonight. Previously I've used Frequency Response Combiner, which works well and will apply some smoothing to the transition, but it quite cumbersome and fiddly to use.

I'm hoping that you and others will find this to be a reliable tool for routine processing of FRD files. Once you get the hang of it, things move along quickly and I am finding it to be extremely useful. Please let me know if you find any steps of the process are difficult to understand, etc. or if it's not clear how to use the controls for blending, scaling, or changing the slope of the tails.

I could easily implement a similar kind of blending where the tails join the "data". With the current version, there is an abrupt transition to the tails from one point to the next, and this might lead to a discontinuity. Also, instead of a "start" and "stop" frequency for blending, I could implement a "blend center frequency" and "width" if that is easier to understand for the user...

I think that the FRD Blender is less painful when it is paired with the ACD Extension called the "File Import-Export Helper". I highly recommend that you use these together - the Helper to get FRD files in and out easily, and the Blender to merge different FRD files and extract the minimum phase response.

-Charlie

[dcibel]

Re: FRD Response Blender and Minimum Phase Extractor is now available

This is just great. Very simple and does what I want it to. Thanks!

[[email protected]]

Re: FRD Response Blender and Minimum Phase Extractor is now available

I guess im just not familiar enough with excel to get it to function right.
how about a video?

[charlielaub]

Re: FRD Response Blender and Minimum Phase Extractor is now available

I guess im just not familiar enough with excel to get it to function right.
how about a video?

Arlis,

If you are having some difficulties using the blender, please contact me by PM. I'll send you my EMail address where you can send me the spreadsheet with your data in it and then I can help you.

Have you read through the tutorial?

-Charlie

[[email protected]]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Charlie, I have. Sometimes I'm just dense.

[Jeff B.]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Charlie, I have. Sometimes I'm just dense.

I use it all the time now. However, I use a modified version that I tweaked for myself, because once again I just wasn't smart enough to work with something the way it was made. I need to send my latest to Charlie and get his two cents worth on it, because when I do the design seminar I will probably use this little tool since it's become a part of my standard design process now. I'll want to make sure Charlie is OK with what I have added before I start sharing it.

Jeff

[charlielaub]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Thanks to Arlis for helping me find a small "bug". When a user imported only File 2 and File 3, but not File 1, the File1+File2 data did not show up in the "Magnitude Responses" plot on the control worksheet. It's just a plotting problem. I typically use this tool with all three files populated with data, so I didn't catch this issue.

I figured out where this glitch is happening and it should be a quick fix, although I will need to run it through some tests to check it out afterwards.

I should have a new version posted to the web page tomorrow sometime, and I will confirm here with a follow-up post.

Jeff, feel free to send me something. Please let me know if you feel that a part of your custom version should be integrated into the tool, if that would be helpful for others and you would like to see that happen. Of course, you are more than welcome to incorporate it into your presentation.

-Charlie

[Jeff B.]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Thanks to Arlis for helping me find a small "bug". When a user imported only File 2 and File 3, but not File 1, the File1+File2 data did not show up in the "Magnitude Responses" plot on the control worksheet. It's just a plotting problem. I typically use this tool with all three files populated with data, so I didn't catch this issue.

I figured out where this glitch is happening and it should be a quick fix, although I will need to run it through some tests to check it out afterwards.

I should have a new version posted to the web page tomorrow sometime, and I will confirm here with a follow-up post.

Jeff, feel free to send me something. Please let me know if you feel that a part of your custom version should be integrated into the tool, if that would be helpful for others and you would like to see that happen. Of course, you are more than welcome to incorporate it into your presentation.

-Charlie

That's odd. I've never seen this and I only tend to use files 2 and 3 and leave 1 blank. I use it primarily to blend my response measurements with a box model plot. I'll have to double check this, but I see all plots on the Control screen just fine.

[charlielaub]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Jeff - let me know what version you are using.

Also, while I am in there I might as well do other stuff. Here is my idea - looking for feedback:

I typically use this program to blend together measured impulse-response derived frequency data (e.g. 300Hz and up) with a Thiele-Small box model + diffraction response model. This results in a very nice, smooth low frequency roll-off. However you can also use a nearfield measurement in place of the box model and in that case the response may not be very smooth around the knee of the roll off and below. Additionally, the high frequency roll off of the driver is often full of peaks and dips (e.g. not smooth). In the current version of the FRD Blender, the tails are simply spliced on to the data at a single data point. This can have the effect of introducing a sharp discontinuity. Sharp artificial discontinuities in the frequency response introduce disturbances in the phase, as if millions of voices suddenly cried out in terror and were suddenly silenced. In other words, it generates an undesirable phase artifact.

To address this, I have been thinking of implementing some blending between the tails and the adjacent data, possibly within a user definable zone and with the possibility of "no blending" like it is done now. This would allow the user to have control over the transition into the tails and smooth it out when it's appropriate/desired.

Another change I will likely implement is in the blending function. I currently use an error function to blend things together. I thought that this would make for a nice smooth transition but now feel that a simple linear blending function would be better, especially at the extremes where the ERF is only slowly changing. The middle of the ERF is like a linear combination anyway, and this would make it easier to define the zone within which the blending takes place. The user likely won't notice this change and the blending effect will be very similar.

If you have thoughts on these things, and/or suggestions for new features this would be the time to let me know.

-Charlie

[Jeff B.]

Re: FRD Response Blender and Minimum Phase Extractor is now available

Jeff - let me know what version you are using.

Also, while I am in there I might as well do other stuff. Here is my idea - looking for feedback:

I typically use this program to blend together measured impulse-response derived frequency data (e.g. 300Hz and up) with a Thiele-Small box model + diffraction response model. This results in a very nice, smooth low frequency roll-off. However you can also use a nearfield measurement in place of the box model and in that case the response may not be very smooth around the knee of the roll off and below. Additionally, the high frequency roll off of the driver is often full of peaks and dips (e.g. not smooth). In the current version of the FRD Blender, the tails are simply spliced on to the data at a single data point. This can have the effect of introducing a sharp discontinuity. Sharp artificial discontinuities in the frequency response introduce disturbances in the phase, as if millions of voices suddenly cried out in terror and were suddenly silenced. In other words, it generates an undesirable phase artifact.

To address this, I have been thinking of implementing some blending between the tails and the adjacent data, possibly within a user definable zone and with the possibility of "no blending" like it is done now. This would allow the user to have control over the transition into the tails and smooth it out when it's appropriate/desired.

Another change I will likely implement is in the blending function. I currently use an error function to blend things together. I thought that this would make for a nice smooth transition but now feel that a simple linear blending function would be better, especially at the extremes where the ERF is only slowly changing. The middle of the ERF is like a linear combination anyway, and this would make it easier to define the zone within which the blending takes place. The user likely won't notice this change and the blending effect will be very similar.

If you have thoughts on these things, and/or suggestions for new features this would be the time to let me know.

-Charlie

Charlie,
I am using version 1.91.

Well, here's a brief summary of what I did for my personal version. I really like how you blend the response, add tails, set the frequency ranges, and extract the phase. I thought all of them were very well implemented (much nice than I have done). And I know you intentionally designed the spreadsheet to operate without the use of macros and VBA. However, since I was using it for myself I decided to "automate" mine a bit to speed things up and make it a lot quicker and easier to use, especially since I was using it very often on a lot of my measured data. So I......

Added frd import and export macros for all sections,
Added spinners for the controls so changes could be dialed in quickly,
Added a macro to run the phase extraction,
Added a box model with the option of copying its result in for blending rather than importing an frd file (which you can still choose to do if you want).
Added a diffraction model with the option of copying its result in for blending rather than importing an frd file (which you can still choose to do if you want).
Added the ability to also change the SPL level of the final response curve before exporting it.
And made some small changes to a couple of the graphs.

So, if you make many changes to your program's function I doubt I will be able to keep up and reimplement all of these features. I understand your original intention, but I bet a lot of people would like the convenience of the things I added.

Jeff

[[email protected]]

Re: FRD Response Blender and Minimum Phase Extractor is now available

That could make it a bunch easier for guys like me. Please collaborate.