PCD Windows app interest?

[dlr]

I'll try to put up a page at my site with some screen captures of what's working to date.

I've put up a page with links to full size screen shots of the first passive sections (tabs). The missing parts of these are the calculations related to the crossover. I think that I have everything else that's needed. I hope to get to doing that part soon.

WinPCD

dlr

[Pete00t]

You can sign me up for the beta testing. Have you also thought in the future testing it using WINE? I'm sure there are many users with the love for Linux. Cross platform ability would be a nice thing.

[dlr]

You can sign me up for the beta testing. Have you also thought in the future testing it using WINE? I'm sure there are many users with the love for Linux. Cross platform ability would be a nice thing.

Cross platform would be nice, yes, the problem is the .NET framework. I believe that Windows has a (limited) version for OS-X and I see that WINE shows "Beginnings of support for .NET mixed assemblies". I'm using .NET4 and not anything much beyond basic classes in Winforms, with one exception, so it's possible that it may run on these. The exception is the GPL 2-D graphics package that might be a sticking point.

dlr

[neildavis]

Cross platform would be nice, yes, the problem is the .NET framework. I believe that Windows has a (limited) version for OS-X and I see that WINE shows "Beginnings of support for .NET mixed assemblies". I'm using .NET4 and not anything much beyond basic classes in Winforms, with one exception, so it's possible that it may run on these. The exception is the GPL 2-D graphics package that might be a sticking point.

dlr

You can go to http://mono-project.org and download the compatibility checker for converting the .NET CLR code to Mac and various UNIX/Linux platforms.

[neildavis]

You can go to http://mono-project.org and download the compatibility checker for converting the .NET CLR code to Mac and various UNIX/Linux platforms.

Nevermind...This is for PSD_Lite, which uses no external libraries and no database code. Looked like some of the errors could be circumvented with different code constructs, but the sheer number of errors here is not a good sign.

[scottq]

If I were closer to Boston, I'd gladly drop off a few cases of Mountain Dew or some wine or beer... to help motivate the creative process. Thanks again for all your efforts!

[dlr]

Nevermind...This is for PSD_Lite, which uses no external libraries and no database code. Looked like some of the errors could be circumvented with different code constructs, but the sheer number of errors here is not a good sign.

That definitely has a foreboding look to it. Developing in .NET4 probably increases the likelihood of missing methods.

I just took a quick look, it appears that it's for C# only on .NET. Despite the CLR, yours being VB may be the issue.

dlr

[Jeff B.]

I've put up a page with links to full size screen shots of the first passive sections (tabs). The missing parts of these are the calculations related to the crossover. I think that I have everything else that's needed. I hope to get to doing that part soon.

WinPCD

dlr

Dave, I love the look. It makes a smooth visual and intuitive transition for the thousands of PCD users (literally) that already exist, which I am sure would make it very attractive to them.

I believe Neil offered some of his passive circuit code. Have you checked it out yet? Is there any way you can implement it using your interface?

Jeff

[arlis_1957@yahoo.com]

david, i agree. its nice to see the driver area as well as the sum with phase at the same time.

[neildavis]

I believe Neil offered some of his passive circuit code. Have you checked it out yet? Is there any way you can implement it using your interface?

Jeff

He's got the entire PSD_Lite project source code.

Obviously, it's up to Dave to decide, but my code could help get over a fairly tough hurdle. About 3-1/2 years ago I sent you a mock-up of the user interface for WB&CD using .NET to show you how it would look. That turned out to be fairly easy. The hard part was taking all of those Excel equations and converting them to code. The biggest problem is understanding the dependencies. In Excel the cells are calculated using the results from other cells, and Excel keeps track of what needs to be updated in the right sequence. When you write your own code, you need to manage those dependencies yourself. Also, you need to develop your own data model because variable names like B19 are too difficult to keep track of. I think PCD is going to be a lot simpler to convert than WB&CD because that Benson model is very large, and it had lots of interim calculations to keep track of. But still, it's going to take time to do the conversion for PCD, and maybe using my code would be a lot quicker. I even documented how the equations are derived so that someone could extend them someday--there is a link to an online page in PSD_Lite on the Crossover Help menu. But again, that's Dave's call and I don't want to interfere with his decision.

My only "hot button" about this project is that I would really like to see some reuse of the object model I used in PSD, as it is the key to having these audio tools interoperate. The downside of using this object model is that some of the the variable names become very long. For example, the variable for the DC resistance of the inductor in the first parallel leg in the woofer circuit is: Main.PCD.Crossover.Woofer.P1.L_DCR, where PCD is the instance of the loudspeaker object. But once you get used to using this model, it becomes more and more intuitive and "logical". If you look at the model from the Help menu in PSD_Lite you can see how extensive this model is. I can use it for all of the data in PCD and also for WB&CD, the Baffle/Diffraction Simulator, and even completely model the internal workings of a miniDSP (including the configuration of all of the biquads, channel mappings and volume settings for up to 8 channels). Another benefit of using this object model is being able to save the entire program state with a single XML_serialize command. The PSD project save file is an open-format self-documenting XML file--pretty cool.

Anyway, the best nugget in that source code is the Phase Extractor that is part of the Response Editor (select "Tools" from the Driver menu). It uses the method described by Julius Smith: http://www.dsprelated.com/dspbooks/f...mum_Phase.html. The code is easy to read, easy to understand, it is well documented, and it works. Paul Verdone's code in RM also works, but it is the most cryptic code I've ever tried to follow, and there is almost no documentation (very few comments, and his user manual doesn't help much). I wasted quite a few days trying to understand his code before deciding it was best to move on to a fresh design. If Dave wants to add phase extraction to PCD, I would strongly recommend we discuss using this code instead of Paul's.

[dlr]

I believe Neil offered some of his passive circuit code. Have you checked it out yet? Is there any way you can implement it using your interface?f

I doubt it. For one I've got the schema for the objects the way I intend to use them. The issue on which I still have no handle is that my code is in C# and his is in VB. They both can be used due to the CLR, but I don't know how I could use it directly. I don't think that there's a way to directly use his classes. I'm also doing this as an exercise in working through all aspects of design in this development environment.

He's got the entire PSD_Lite project source code.

Yes, but I've yet to go into it. Between another side project (that I hope will end up a compensated gig) and advancing the PCD, I haven't had time.

The hard part was taking all of those Excel equations and converting them to code.

This may be where our approaches differ greatly. I decided very early on that it would take a lot more time to try to duplicate the Excel code. Mine is totally from scratch, with two exceptions. One is the UI look and fell, the other is the basis for the interpolation/sample points and Jeff's algorithm for that. The calculations in the Excel code don't directly translate to the objects as I created and use them. I strive for the most flexible scheme that I think may prove easily expanded/reused later. My "tree"hierarchy is broad, not deep so-to-speak. The dependencies are likely very different than yours. I like the exercise anyway.

Another benefit of using this object model is being able to save the entire program state with a single XML_serialize command. The PSD project save file is an open-format self-documenting XML file--pretty cool.

This is an area that interests me. I haven't used this, have no experience (yet) in it, but would like to. I'll have to look into using this as well.

Anyway, the best nugget in that source code is the Phase Extractor that is part of the Response Editor (select "Tools" from the Driver menu). It uses the method described by Julius Smith: http://www.dsprelated.com/dspbooks/f...mum_Phase.html. The code is easy to read, easy to understand, it is well documented, and it works. Paul Verdone's code in RM also works, but it is the most cryptic code I've ever tried to follow, and there is almost no documentation (very few comments, and his user manual doesn't help much). I wasted quite a few days trying to understand his code before deciding it was best to move on to a fresh design. If Dave wants to add phase extraction to PCD, I would strongly recommend we discuss using this code instead of Paul's.

This is the most likely part that may work out in some form. I think it would a good addition to the PCD. However, not for v1.

One thing that just occurred to me is that Visual Studio Express does not allow development in two languages in the same instance of VS. You work in C#, C++ or VB, not all three in the same instance. That alone probably prevents easily using any VB code even if it could be accessed directly, something about which I am unsure.

dlr

[neildavis]

I doubt it. For one I've got the schema for the objects the way I intend to use them...

Probably too late in the design phase to address commonality...

This may be where our approaches differ greatly. I decided very early on that it would take a lot more time to try to duplicate the Excel code. Mine is totally from scratch...

This is the right way to do it. It tried converting WB&CD several years ago and I swore I would never go that route again. My version of WB&CD even has some of Jeff's original Excel equations as comments to allow readers to "follow the trail". But as the program got closer to completion, more and more of the Excel heritage got driven out, and it became clear to me that it would have been *a lot* quicker to just start from scratch with the equations. That's what I did with all of the other modules in PSD.

One thing that just occurred to me is that Visual Studio Express does not allow development in two languages in the same instance of VS. You work in C#, C++ or VB, not all three in the same instance. That alone probably prevents easily using any VB code even if it could be accessed directly, something about which I am unsure.
dlr

Right--you would have to convert the VB stuff into C#. But in the last few years all of these languages have evolved along similar paths. VB.NET has undergone the largest amount of evolution, and I think it is fair to say that it can be as object-oriented as the others. There are still the syntax differences, but translating a modern VB.NET app into C# should be fairly easy.

[mdocod]

Another +1 here

I use linux, so, windows apps are no problem under wine.... excell: no thanks

[TAJ]

I have to add my 2-bits and some encouragement here, but I'm really surprised that you guys would want to use any proprietary platform anymore when you can build a generic web app using javascript and HTML 5. It (any current web browser) seems ready for prime time to me.

A couple years ago I created a web app that plotted some basic T/S parameters onto a nice looking graph, and it worked great. I used the Flex language from Adobe at the time, which loaded the app into Flash on a web browser. It worked very well. But there are a bunch of reasons not to use Flex/Flash now (ipads, iphones, etc.) so I've been meaning to re-code my app into javascript and HTML. (Flex has the same roots as Javascript, so porting is probably effortless.)

Anyway, my main point here is: Create a web app! Don't bother building a desktop app in this day and age. Web apps are useful for desktops, tablets and even phones. No downloading, no installing, just go to the address and start using it!

Check out Sencha.com for an open source coding framework (or 2 or 3) that will completely simplify the project. Caveat: Their main web app framework (ext4) seems to be missing a log scale x-axis function in their charting package, but the mobile framework (Touch) does support it.

Think web and mobile, not desktop.

~ Todd J.

[jsr]

Just adding my support for an app version of PCD, especially since my home computer does not have office and it's driving me nuts that I can't run WBCD and PCD now as there are some designs and sub sims I really want to do (I'm off work for a month so no access to my work PC which has PCD).

[johnnyrichards]

I vote against Java or Flash gaining any more traction in this world. Leave it for little games and low-rez videos.

[dlr]

Anyway, my main point here is: Create a web app! Don't bother building a desktop app in this day and age. Web apps are useful for desktops, tablets and even phones. No downloading, no installing, just go to the address and start using it!

Too late. V1 will without question be a desktop app.

Think web and mobile, not desktop.

If you look at the size and real estate required, you may think differently about wanting this mobile! Mobile "flashed" through my mind up front, but I just don't see this app working in that mode.

A Windows app also does not require web access to use. I would not want to be dependent on an internet connection at all times. However, I suspect that a web app will be considered, but at a much later date, meaning no time soon. This is my first app like this of any kind, for the record.

I vote against Java or Flash gaining any more traction in this world. Leave it for little games and low-rez videos.

I've read that Flash is on the way out. I would not go that route at this point.

dlr