Hi Everyone,
Sorry if this is asked elsewhere.
I've been using XSim very successfully but DIY'ers have suggested other software as well. What I particularly thought would be a nice feature is crossover optimization. The idea that you can draw or request a type of electro-acoustic shape and the software will make it for you, based on the actual driver.
I think LspCAD has this, though I've never used it. Are there others?
Best,
Erik
Sorry if this is asked elsewhere.
I've been using XSim very successfully but DIY'ers have suggested other software as well. What I particularly thought would be a nice feature is crossover optimization. The idea that you can draw or request a type of electro-acoustic shape and the software will make it for you, based on the actual driver.
I think LspCAD has this, though I've never used it. Are there others?
Best,
Erik
http://www.boxsim.de/download/Boxsim120en.zip maybe not the best tool for Simulation but the xover optimiert works well.i linked the english version and its freeware
Thanks for the suggestion. I got to try it. I found the schematic portion too difficult to use. It reminded me of OrCAD in that way. It was good that I could import data from OmniMic and DATS easily though.
Best,
Erik
First I thought you mean FilterShop instead, but now I see CrossoverShop as well.
Thanks for the ideas!
Erik
Hi Erik,
SoundEasy has quite powerful crossover optimizer. Please read Chapter 10 below.
http://www.bodziosoftware.com.au/Chapter_10.zip
More info is on Bodzio Software website.
Best Regards,
Bohdan
IMO, Soundeasy has quite possibly the worst User interface ever designed. It is a kluge of tools the author has seemingly written for his own use and doesn't have any discernable workflow. There (at least at one point) is a guy who makes $11 a pop for an aftermarket manual on one way (of many) to use it and that is not enough of a hint for the author that his software needs improvement. Spend your hard-earned money elsewhere, is my recommendation.
You can get by, to a great extent, with free tools nowadays. Speaker workshop has a really good crossover optimizer for times when you need one. Probably best used with windows XP, though...
No optimizer, but otherwise, very easy to use. http://audio.claub.net/software/jbabgy/jbagby.html
Thanks!
Why do so many of the software for speaker measurement and design stink. What's worse is even expensive commercial products like Leap seem stuck in the 1980's of operating systems and literature. It's rather scary to think of spending $1,500 on software that doesn't look like it's been updated in 20 years. Same thing for X-Over Pro. I wonder how many software vendors treat their products as pure cash cows with no future at all.
One nice exception to this seems to be Audiomatica's Clio, which has active development and support.
If Windows 10 were mandatory, most of the software we want to use would drop dead. Hahaha.
Erik
Last edited:
I don't think REW stinks at all - it is an amazing piece of technology and easy to use and free. It will do everything you need coupled with a program like PCD or Xsim.
Optimizers: I assume you are saying something that will parametrically and automatically massage a XO circuit component values to find the optimal solution? Kind of like auto Room EQ wizards? I think you are far better off manually applying strategic filters - by turning blocks on and off and manually tweaking values while watching the visual prediction of the outcome. There are too many variables for an auto solve using method of simulated annealing to arrive at a good solution. The brain is sometimes the best for doing this quickly and with as few filters as possible.
Optimizers: I assume you are saying something that will parametrically and automatically massage a XO circuit component values to find the optimal solution? Kind of like auto Room EQ wizards? I think you are far better off manually applying strategic filters - by turning blocks on and off and manually tweaking values while watching the visual prediction of the outcome. There are too many variables for an auto solve using method of simulated annealing to arrive at a good solution. The brain is sometimes the best for doing this quickly and with as few filters as possible.
I'm not big on optimisers, usually finding the topology is the hard part, not tweaking the values in -- which isn't too hard and is kind of fun. But if you do want to go that route, look for an optimizer that can be made to restrict to standard discrete values. Otherwise, youll have to spend more time and $$ coming up with odd things like 3.72uF caps or 5.38mH inductors.
I like REW too! It has a lot of things going for it, including:
- Free
- Cross-platform (due to being Java based)
- Actively being developed (Last release 12/15/2015)
That's what I do now.
I'm just wondering if the AI is better. Now you are invoking a phrase I haven't heard outside of genetic programming.
It doesn't have to do it all, but at least take a part of a circuit, and try to come up with better solutions. For instance, tweaking the 30 parts to try to get better phase matching. Best,
Erik
Last edited:
I am a happy Audiomatica Clio owner, err, only my version is the
vintage ISA card from the 90's that works in DOS. Came with
a Clio mic for a total of 1,000.00 DM which I bought in Germany.
Clio Pocket on PE is a superior upgrade, worth the money asked, IMO.
There is a an Excel plugin for genetic algorithms. Anything you can model in Excel, it lets you pick the variables you want it to tweak and you set the calculation for what you feel is the optimal - for example low phase shift. It then generates mutations of solutions based on genetics (parents offspring etc) and then applies and checks if your "goodness" parameter gets better. The genetic mutations are what allow out of the ordinary solutions that otherwise might not be tried. I can see that this with PCD would be the perfect pair. Genetic algorithms are more robust at finding optimal solutions than simulated annealing. The problem is PCD is "locked" down and I am not sure you could get the GA plugin to be able to work with it.
Thus it uses the users' know how of what variable is important and which ones we can tweak to automatically solve and hopefully converge on a solution that is more optimal than what we can do by hand.
Thus it uses the users' know how of what variable is important and which ones we can tweak to automatically solve and hopefully converge on a solution that is more optimal than what we can do by hand.
Last edited:
That's pretty cool, but Excel is too proprietary, it's Perl or nothing for me.
What is PCD?
Erik
True, but how difficult would it be to add an API to XSim to allow others to write their own? The API would have to present a list of components that can be changed, their current values, allowed ranges and the results.
The plug-in would change the values, request a new computation, and repeat.
We can call the new version ..... XSin II
Erik
That's pretty cool, but Excel is too proprietary, it's Perl or nothing for me.
What is PCD?
Erik
Passive Crossover Designer. It is Jeff Bagby's personally developed Excel spreadsheet that does real time modeling of crossovers and PEQ's. It takes as input FRD and ZMA files. It auto generates your chosen filter function target overlay so you can work towards a goal as you tweak values in real time. It looks intimidating because it is so big visually. You need a 24in widescreen to see the whole spreadsheet. I really like it. I have been able to use it and take one set of raw measurements and then model the DSP PEQ's and the XO I need for a certain topology. Like the Harsch for example which needs a BW4 and Bessel 2 acoustically. So you need to pick the electrical and PEQ's that combine with FRD files to produce the correct target function. It is so accurate that I just load parameters from the PCD simulation into my miniDSP and it is correct the first time. No more tweaking in real time.
Jeff Bagby's Software Page
If PERL is what you like you will have to write your own XO simulator program and your own genetic algorithm code and your own data acquisition code for measurements because no one uses it except for web applications and maybe TV set top boxes.
Last edited:
What kind of crossover are you targeting?
passive, active, digital IIR, digital FIR?
"Automatic" corrections should always be used with caution, as they put a lot of pressure on the quality of the measurement. OTOH manual correction might be more tedious at first, but let you use the best possible analysis tool to get the right result: your brain and knowledge of the measurement context.
The important part is the measurement anyways, because corrections end up being quite simple most of the time (unless one is using sh**ty drivers, or use them out of their intended range).
That's pretty cool, but Excel is too proprietary, it's Perl or nothing for me.
rePhase is written in Perl
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.