Ideal sound card for multichannel scan and modulations.

[carmangary]

I looked at the Mamba API awhile back. It's pretty straight fowards. It has a few C functions that you have to implement. Basically, you configure your driver to receive points or to receive frames. Then Mamba sends you those via one of the C functions. Pretty elementary actually.

[The_Doctor]

My efforts have found that wxLua is clearly not fast enough, even on a 1.2 MHz C3 processor, to write realtime waves, unless I missed something, but what about wxPython?

That looks like a good system, nice clean layout to the site showing what must be done, except that they miss the one most vital thing, they don't say what it can DO. I imagine it can read/write files, access system clocks/counters, sockets for networks, but can it write fast data streams to wave outputs?

If I can use a language that has the kind of syntax of Lua or Python, but compiles to something approaching C's speed at runtime (hopefully possible, given that these languages are both written in C), it seems the only obstacle beyond a good compiler is the libraries needed for wave output. All of which is beyond me so far, I'd need some patient help to get even the most basic program compiled and spitting out realtime waves, but once I have that I should be ok, if a tad slow.



Edit:
I found this: http://www.wxwidgets.org/manuals/stable/wx_wxsound.html
Seems like a new tool in wxWidgets. Not quite enough, but a step in the right direction for simple developing of ideas. Not sure if wxLua uses a wxWidgets late enough for this one, but wxPython probably does. If anyone can reach further in the direction I'm trying to go, let me know what you find.

[carmangary]

Python is an interpreted language so you aren't going to get the performance you need out of it.

If I had to recommend something it would be VB.NET or C#. You can download the express editions for free and the syntax isn't that hard. I am an expert at C#, and if you chose C# I will gladly help you (via email) through any problems you have. I haven't used VB.NET but it is practically the same and offers the exact same performance. I could probably help you with that, too.

But, messing with any of those oddball interpreted languages you are talking about is going to be a waste of time since they are so slow. But, if you just want to create a wav file from some data and then play the wav file later then those interpreted languages would be fine.

[carmangary]

I was looking into this a little more tonight. Not sure is using the DAC of a soundcard is the way to go since it involves modifying something inside your PC. Instead, I'm thinking a better approach is to use the SPDIF output from a soundcard to drive a SPDIF DAC. As far as software goes, there would be absolutely no difference, though.

[The_Doctor]

That would work well. So long as the software puts its X/Y and analog mod data out of sound ports, all kinds of freedom exist. Building small S/PDIF DAC's or adapting them might be a useful bit of extra business for various laser show parts suppliers. Some people will prefer to buy a cheap sound card and hack their own, or even a pro multichannel one like Layla24 (nothing to that, just bypass output caps). What's nice about the S/PDIF idea is that even the cheapest stuff usually has that, and it's a good way to get a very high signal to noise ratio as well as a sync signal if needed.

There are also cheap USB and FireWire widgets. I bought a couple of C-media IC based USB thingers to try, and that looked like a small mod with an op-amp or three would be enough to remove the DC offset from the DAC output and match the level to scan amp inputs. The C-media CM106 IC has 7:1 outputs, so enough for full X/Y and analog RGB.

I still like the Layla24 best though, it's so easy it's truly painless, it's still fully useable as a pro quality sound unit, and it doesn't need any amplification before direct balanced connection to the scan amps. By the time you add that lot to a new case or DAC, it might be found that buying a Layla24 might have been the cheaper route, as well as one of the best. I imagine pro S/PDIF DAC's exist with full +4dB outputs, but they probably cost a bit.

[Coo-mo-d]

So all this talk got me thinkin 2. I started looking up some older sound cards to check them for modification but even anything way back as the sound blaster series started using DSP mixer chips n such. It's all this extra crap that gets in the way.

I then remembered some ISA bus interfacing schematics I had and dug them up. It would be extremly easy to just build a DAC card for a 16-bit ISA then trying to modify a soundcard, Its basically a glorified faster popelscan but with 12 or 16 bit DACs. Checking the speeds and also the way the ISA bus works thers no reason why you couldn't make it a dual or possibly quad projector card.

I'll be willing to design the card if someone is willing to write the driver. I have rusty and limited programming skills and for me to write them would take forever. Also read Medialas part on how to write a driver and doesn't give much information on how Mamba sends the information, just how to call it, etc.

PCI isn't much different but requires more work on the software and driver side which again is beyond my ability. 16-Bit ISA is a good start and maybe take the PCI step after.

The offers there if anyones up for the challenge.

-Brian

[The_Doctor]

That's going the wrong way. I don't mean going back to ISA, I mean the making of any specific hardware. The point is to avoid that. If you're going to do that, forget sound cards, just make a DAC to fit the interface as a plugin board.

That's not what I want to see though, it's a good idea but it's not solving the problem, it will just make another highly specific solution for those who want it, and can pay, as it will cost more. It will also need its own specialised hardware drivers written for it, and that's HARD, it always raises costs. Expert companies are always struggling to improve drivers and that means taking into account what other current firms drivers might conflict. My idea is to cut that out, to ride on the efforts already done with pro sound cards, to save having to do that work again.

All I want to see is the sound system in the OS being used to stream the data to a sound port. That way the drivers will be those that come with the sound card. Nothing to write. Anything that suggests ANYTHING else has probably been done, is usually a proprietary and expensive and highly specific product. People seem to be making them all the time. No commercial firm seems willing to allow their software to put out through general hardware unless they can profit from the deal themselves, it's as if they'd rather use that as a kind of glorified dongle instead of letting people be free with their choice. Sure, they might lose some software to piracy by allowing any sound card to be used, but they'll sell so many more copies as well if people don't have to buy expensive dedicated hardware that has no other purpose.

I know some cheap sound cards have crap on board and poor control of levels/clipping, but you get what you're willing to pay for. At least, you do if you have the choice! If you're only willing to use the cheapest lowliest Soundblaster, you're probably only likely to be doing spirographics or other stuff that doesn't even need the DC bypass mod. If you're going further, building an RGB system or other moderately complex device, you'd do more to choose a board and make the mod yourself. Or buy from what might become a kind of cottage industry amongst laser people, buying and selling boards known to take DC bypass mods well, already prepared for use. Considering the trouble people seem to have getting DAC level matching and ILDA standard connections and such sorted, a DC bypass mod is trivial, and pro units have enough voltage to match any level a laser system's signals will need. The whole thing is easily controlled in software, with no need to write a driver at all.

Cheap sound cards have dodgy wave mixers, I know, and added low-grade FM synths and such, but these are easily bypassed. Any error in the signal will still be finer than 12 bit res, which most agree is plenty for laser show work. A far bigger problem is the buffering, allowing a steady stream to flow even if the computer is choking on a RAM swap operation or other task. This is why I want the sound port idea to be developed widely, because to avoid that, you just choose a sound card that works well in these conditions. Any tech review mag, online or off, is a source of data to help people choose. Many sound cards can be modified in a way that still lets them work as sound cards too, which for anyone not professional, is a BIG advantage, given that they can also have professional standard signal flow for laser control too.

Any software that can write wave stream data to the OS for sound port output can send signals to DAC's, on many channels, just by duplcating them for whatever the current hardware can support. Considering that a cheap sound card is good enough for hobby use, and a pro unit can do enough for a very powerful colour show, I don't think that turning away from the OS's sound output methods in favour of dedicated ISA and PCI cards is the way to go. It's A way to go, but it totally detracts from the point of this thread.

[carmangary]

Nothing wrong with cheap. Look at the Laserboy website. That was done on a $10 soundcard!

[The_Doctor]

True. They can do well. Actually, there are two kinds of sound card that will be largely free of bells and whistles, and have only what is needed for good basic function. The high end of the market, and the low end. Both are mainly concerned with getting a stream of data out and not interfering with it. If you want low jitter, ability to sync more than one unit, better drivers, full +/-12V out with enough dynamic range for software to do accurate scaling of voltages at user programming level, and the ability to set the sample clock at one Hz accuracy for precisely defined point rates, go for the high end. It still costs LESS than any single DPSS laser worth having. You'll even get MIDI for low-speed automation and remote control, and only one IRQ to worry about, which is vital for system reliability.

[steve-o]

I don't know a damn thing about what you-all are talking about, but I have 2 home recording studios (1 PC based) and the cure to latency in sound-cards is ASIO.

Fwiw.