Hi Charlie, Is the requirement for miniDSP or beringer so that the filters can easily be simulated for testing to check the results match the programs output?
I've got the measurement gear, and I've got an active filter project (which I haven't worked on for a while but want to get back into) so it's relevant to me
Tony.
That's right. It's one thing to play around with the software and create a nice crossover, and another to then implement the crossover and measure the loudspeaker frequency response. People can do whatever they want with the first release, but at this point I am aiming just to make sure that there are no major bugs or glitches that I haven' caught, and to make sure that other can understand how to use the tools and that documentation is in place for that.
If you want to be in at the testing stage and can implement the crossover, send me a PM. Thanks,
-Charlie
I would be happy to create output formatted for other IIR DSP units such as the netlist that you mentioned. I don't have time to contribute to the hardware development, but maybe that effort will gain some momentum. People working in that area are welcome to contact me.
-Charlie
You can't implement a proper crossover with EQ alone. That said, yes you can use my tools to model how the EQ will effect the response.
If you want to be a tester, please send me a PM. Thanks,
-Charlie
Forgot to mention...
I was running through the instructions I wrote for the trial FRD data sets that I included in the zip file to all the testers and I realized that:
I FORGOT TO MENTION THAT THE TWEETER POLARITY SHOULD BE SET TO NORMAL WHEN DETERMINING THE ACOUSTIC OFFSET
This is actually nice demonstration of just how sensitive this way to determine acoustic offset is. If you try it with the tweeter polarity reversed, you can't get a good match between the measurement and the model of the response. It's only possible to get these to agree when the polarity of the tweeter phase is normal. I inadvertently left the polarity set to "reversed" in the worksheet I sent out. Check cell B12 in the tweeter driver response spreadsheet. The value should be set to "1" for "normal polarity".
Sorry for the confusion!
-Charlie
P.S. I have also cleaned up some of the instructions about loading the FRD files that were unclear and replaced the file in the zip archive.
I was running through the instructions I wrote for the trial FRD data sets that I included in the zip file to all the testers and I realized that:
I FORGOT TO MENTION THAT THE TWEETER POLARITY SHOULD BE SET TO NORMAL WHEN DETERMINING THE ACOUSTIC OFFSET
This is actually nice demonstration of just how sensitive this way to determine acoustic offset is. If you try it with the tweeter polarity reversed, you can't get a good match between the measurement and the model of the response. It's only possible to get these to agree when the polarity of the tweeter phase is normal. I inadvertently left the polarity set to "reversed" in the worksheet I sent out. Check cell B12 in the tweeter driver response spreadsheet. The value should be set to "1" for "normal polarity".
Sorry for the confusion!
-Charlie
P.S. I have also cleaned up some of the instructions about loading the FRD files that were unclear and replaced the file in the zip archive.
Preview
I was answering some questions in a thread on another forum, and I put together this preview of the interface for one of the tools. This should give people an idea of what they will be getting into..
The pic below is the interface where you specify the filters and EQ to be applied to the response. There are two tables, one for the filters and one for EQ. The other stuff is located above the filter table.
Let me illustrate how to use this with a little example:
Let's say you want to use a 4th order High-Pass Linkwitz-Riley filter at 3000Hz for your tweeter. An LR4 is comprised of two second order filter stages, each having Q=0.707. The filter corner frequency for all LR filters is the same as the crossover frequency. Also, what we are entering is the Fo and Q for the POLE of the filter, so that is Fop and Qp in the table. Thus we enter for Filter 1:
type = 22
Fop = 3000
Qp = 0.707
and, similarly, for Filter 2:
type = 22
Fop = 3000
Qp = 0.707
Now let's say the tweeter is a little hot compared to the woofer, for instance 3.5dB too hot. So we need to reduce the gain somewhere. Let's just choose the gain for Filter 1 to be -3.5 dB. To do that, you enter -3.5 in the cell for the gain for Filter 1, which is cell C26.
Now the filter table looks like this:
A plot is provided in the worksheet so you can make a visual check of the filter response:
The IIR magnitude (red line) is not really visible because the analog filter response (blue line) lies on top of it everywhere. This is usually the case, the exception being LP filters near the Nyquist frequency, which tend to have an increased roll off as the response approaches 20k Hz. Note that the maximum passband level is at -3.5 dB. The phase response of both IIR and analog filters is essentially the same.
Hopefully this interface is simple enough for people to use efficiently, once they learn a couple of things about filter Q.
-Charlie
I was answering some questions in a thread on another forum, and I put together this preview of the interface for one of the tools. This should give people an idea of what they will be getting into..
The pic below is the interface where you specify the filters and EQ to be applied to the response. There are two tables, one for the filters and one for EQ. The other stuff is located above the filter table.
Let me illustrate how to use this with a little example:
Let's say you want to use a 4th order High-Pass Linkwitz-Riley filter at 3000Hz for your tweeter. An LR4 is comprised of two second order filter stages, each having Q=0.707. The filter corner frequency for all LR filters is the same as the crossover frequency. Also, what we are entering is the Fo and Q for the POLE of the filter, so that is Fop and Qp in the table. Thus we enter for Filter 1:
type = 22
Fop = 3000
Qp = 0.707
and, similarly, for Filter 2:
type = 22
Fop = 3000
Qp = 0.707
Now let's say the tweeter is a little hot compared to the woofer, for instance 3.5dB too hot. So we need to reduce the gain somewhere. Let's just choose the gain for Filter 1 to be -3.5 dB. To do that, you enter -3.5 in the cell for the gain for Filter 1, which is cell C26.
Now the filter table looks like this:
A plot is provided in the worksheet so you can make a visual check of the filter response:
The IIR magnitude (red line) is not really visible because the analog filter response (blue line) lies on top of it everywhere. This is usually the case, the exception being LP filters near the Nyquist frequency, which tend to have an increased roll off as the response approaches 20k Hz. Note that the maximum passband level is at -3.5 dB. The phase response of both IIR and analog filters is essentially the same.
Hopefully this interface is simple enough for people to use efficiently, once they learn a couple of things about filter Q.
-Charlie
Impressions?
So, for anyone testing out the software, how is it going?
Can you post your impressions, and comment on any problems or successes?
I have made some updates to both the tester instructions and the MiniDSP filter functions that I would like to post, but I would fist like to know about any other issues that people might discover so that I can fix them before the next update...
Thanks
-Charlie
So, for anyone testing out the software, how is it going?
Can you post your impressions, and comment on any problems or successes?
I have made some updates to both the tester instructions and the MiniDSP filter functions that I would like to post, but I would fist like to know about any other issues that people might discover so that I can fix them before the next update...
Thanks
-Charlie
Hi Charlie,
Still having a hard time getting started. I've copied and pasted drivers into their respective spreadsheets in the right order but the only response plot I see is a combined, no filter response in system. Most of the intermediary response lists are blank (#NAME?) after pushing F9. How does data get from spreadsheet to spreadsheet?
I can work out the filter aspects once I get past this initial hurdle.
David
Still having a hard time getting started. I've copied and pasted drivers into their respective spreadsheets in the right order but the only response plot I see is a combined, no filter response in system. Most of the intermediary response lists are blank (#NAME?) after pushing F9. How does data get from spreadsheet to spreadsheet?
I can work out the filter aspects once I get past this initial hurdle.
David
I will try to post some screen shots and instructions here later. Sorry for the confusion. The #NAME error sound like some cell or range reference has been deleted or changed. Here are some ideas on that:
It's possible that when you are pasting the FRD data in to the driver spreadsheets you are accidentally deleting the named ranges in Columns A, B and C. If you just open the woofer spreadsheet (alone) you should be able to paste in the FRD data for the woofer, hit F9 and see the data show up on the first worksheet in the plots. If not, then there is something going on with the pasting. I copy and paste from Notepad, which is text only (no formatting) and the data pastes in to columns A,B and C directly.
If pasting in to the driver response spreadsheets works, then the next step is to check the links between the driver response spreadsheets and the system response spreadsheet. Remember, you need to export all of these spreadsheets together in to the same directory when you export the zip file contents. There are links between these spreadsheets that must be preserved on order for the driver responses to show up in the system response. It's too complicated to explain here, however.
It would be best if you sent me a PM with your EMail address so that I could communicate with you that way. Anyone else having issues, please PM me with your Email address and I will be happy to help you.
It would be good to figure out where people are getting stuck. I have heard from some people who did not have any problems, so it might just be that I need to be more specific in the instructions on how to do certain steps. I'm sure that we can get this worked out.
Thanks,
-Charlie
NEW VERSION NOW AVAILABLE
I have the next version of the active crossover design (ACD) tools ready for release. This will probably be the last "test" release before I post it for everyone.
If you would like to get the latest version, which has an improved manual and a better now-to tutorial example, please send me a PM or Email, and include your EMail address so that I can communicate with you privately.
Thanks, and happy holidays!
-Charlie
I have the next version of the active crossover design (ACD) tools ready for release. This will probably be the last "test" release before I post it for everyone.
If you would like to get the latest version, which has an improved manual and a better now-to tutorial example, please send me a PM or Email, and include your EMail address so that I can communicate with you privately.
Thanks, and happy holidays!
-Charlie
UPDATE - OO Calc version
Here is an update in case people are interested:
STILL RECRUITING TESTERS:
I'm still looking for people to download, and test the software, and then GET BACK TO ME to chat about successes and failures. I have managed to do this with a couple of people, and have discovered some things to improve or fix. This kind of feedback is very valuable to me, so if you are interested please send me a PM with your Email address so that I can send you the files and then get in touch with you later. Please read the first post in this thread before you contact me:
http://www.diyaudio.com/forums/mult...ource-excel-based-crossover-design-tools.html
Thanks!
NEW OO CALC VERSION:
I have now created a separate version for OpenOffice Calc. A few key things are handled slightly differently in Calc, so it made sense to make a separate version for it.
The Calc version is ready for people to try. You can run it on whatever operating system you want, e.g. Linux or Mac O/S, as long as you can run OpenOffice on your platform.
3-WAY AND 4-WAY SPEAKER CROSSOVER DESIGN TOOLS:
I also can provide 3-way or 4-way versions for people, once they get used to the tools so that you are not limited to 2-way speakers.
-Charlie
Here is an update in case people are interested:
STILL RECRUITING TESTERS:
I'm still looking for people to download, and test the software, and then GET BACK TO ME to chat about successes and failures. I have managed to do this with a couple of people, and have discovered some things to improve or fix. This kind of feedback is very valuable to me, so if you are interested please send me a PM with your Email address so that I can send you the files and then get in touch with you later. Please read the first post in this thread before you contact me:
http://www.diyaudio.com/forums/mult...ource-excel-based-crossover-design-tools.html
Thanks!
NEW OO CALC VERSION:
I have now created a separate version for OpenOffice Calc. A few key things are handled slightly differently in Calc, so it made sense to make a separate version for it.
The Calc version is ready for people to try. You can run it on whatever operating system you want, e.g. Linux or Mac O/S, as long as you can run OpenOffice on your platform.
3-WAY AND 4-WAY SPEAKER CROSSOVER DESIGN TOOLS:
I also can provide 3-way or 4-way versions for people, once they get used to the tools so that you are not limited to 2-way speakers.
-Charlie
new testing underway!
After a few months away from this project, I am back at it.
I developed a new approach that should solve some of the issues that people were having as a result of links not working correctly when the spreadsheets were sent via EMail. Now the user can set up the ACD spreadsheet tools for a loudspeaker with up to 4 drivers. This nicely matches the capabilities of the MiniDSP products.
If anyone wants to try out this new version, please contact me.
Thanks,
-Charlie
After a few months away from this project, I am back at it.
I developed a new approach that should solve some of the issues that people were having as a result of links not working correctly when the spreadsheets were sent via EMail. Now the user can set up the ACD spreadsheet tools for a loudspeaker with up to 4 drivers. This nicely matches the capabilities of the MiniDSP products.
If anyone wants to try out this new version, please contact me.
Thanks,
-Charlie
news...
Quick update: this new approach with templates for the tools is looking good.
Hopefully I should be able to do a public launch by next weekend. I am still looking for people who want to jump in now, in case they find any last glitches.
Thanks to those who have contacted me.
Remember, this software can model just about any active filter/crossover and can quickly implement the entire crossover network in the MiniDSP products so you can make a change and hear it almost immediately.
For now, here is a link to the version 0.3 technical manual, which covers background material about filter design, and how to use the tools to design a crossover from measurements/ FRD files:
Active Crossover Design Technical Manual version 0.3 (PDF)
-Charlie
Quick update: this new approach with templates for the tools is looking good.
Hopefully I should be able to do a public launch by next weekend. I am still looking for people who want to jump in now, in case they find any last glitches.
Thanks to those who have contacted me.
Remember, this software can model just about any active filter/crossover and can quickly implement the entire crossover network in the MiniDSP products so you can make a change and hear it almost immediately.
For now, here is a link to the version 0.3 technical manual, which covers background material about filter design, and how to use the tools to design a crossover from measurements/ FRD files:
Active Crossover Design Technical Manual version 0.3 (PDF)
-Charlie
Drop me a private message with your EMail address and I can send yo the links to the files.
-Charlie
today's update
Thanks to some recent communication from one of the testers, I have identified a couple of minor things that need some attention. I thought I would discuss these issues here in case other people are running into the same difficulty.
Issue # 1. There is currently a limitation on the number of data points that can be supplied in the FRD file to about 860. If your data set is larger than this, the data will be dropped for all rows greater than 860.
Issue # 2. Each driver response spreadsheet has its own frequency grid onto which the FRD file is interpolated. This is necessary to insure that the various internal calculations to work properly using the limitations of Excel. If you supply an FRD file with a start frequency higher than the start frequency of the frequency grid, the calculations for all points between the frequency grid start and the FRD min frequency will generate an error because there is no FRD data to operate on. You can avoid this problem by making sure to set the frequency grid minimum (cell B14 in the ResponseBuilder worksheet) to a value equal to or greater than the lowest frequency in your FRD file.
I have some fixes that I am working on for these minor annoyances. Otherwise, I have been getting very positive feedback from people. I have been using these tools myself, and it's really a nice way to do active crossover development.
Also, I want to re-iterate that the "tools" are completely extensible. If you want to modify a chart, add a new one, add calculations, etc. you can do whatever you would like within the limits of what can be done in Excel.
-Charlie
Thanks to some recent communication from one of the testers, I have identified a couple of minor things that need some attention. I thought I would discuss these issues here in case other people are running into the same difficulty.
Issue # 1. There is currently a limitation on the number of data points that can be supplied in the FRD file to about 860. If your data set is larger than this, the data will be dropped for all rows greater than 860.
Issue # 2. Each driver response spreadsheet has its own frequency grid onto which the FRD file is interpolated. This is necessary to insure that the various internal calculations to work properly using the limitations of Excel. If you supply an FRD file with a start frequency higher than the start frequency of the frequency grid, the calculations for all points between the frequency grid start and the FRD min frequency will generate an error because there is no FRD data to operate on. You can avoid this problem by making sure to set the frequency grid minimum (cell B14 in the ResponseBuilder worksheet) to a value equal to or greater than the lowest frequency in your FRD file.
I have some fixes that I am working on for these minor annoyances. Otherwise, I have been getting very positive feedback from people. I have been using these tools myself, and it's really a nice way to do active crossover development.
Also, I want to re-iterate that the "tools" are completely extensible. If you want to modify a chart, add a new one, add calculations, etc. you can do whatever you would like within the limits of what can be done in Excel.
-Charlie
I have re-coded some sections of the tools and the above limitations have been removed. I will be checking over some other parts of the tools this weekend to see if I can make similar improvements. I'm hoping to put together a web page for downloading the files as well.
-Charlie
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.