"Matlab to Scilab for Dummies" anywhere?

Hello All,

After finding some possibly useful routines to calculate filters (wave digital in this case) I got stuck because these are for Matlab. I only have Scilab and next to no experience with it. The files are mostly *.m, some *.p and a few graphics GUI files *.fig. Scilab doesn't understand any of these.

Is there a text somewhere, along the lines of "Matlab to Scilab file conversion for dummies"?

The brief instructions for routines like "loadmatfile" on Scilab's site are simply not verbose enough for an analog guys like me :-(

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg
Loading thread data ...

"Joerg" schrieb

Octave

formatting link
might be closer to MATLAB than Scilab.

HTH Martin

Reply to
Martin Blume

Hi Joerg,

I dont think its a matter of conversion, a .mat file is a bianary used by matlab, you can use it to port data around to other applications, but not much else... The .m files are ASCII. .fig are either gui tools or graphs. I think you will have to get matlab to use its filter design tool (a gui that gives you the filter coefficients for a dsp,) or the dsp toolbox, etc. Matlab is an excelent tool if you plan to start doing dsp.

good luck,

John

Joerg wrote:

Reply to
john

Hello Martin,

Thanks, I'll give it a try.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Hello John,

Yes, but also very expensive. There is the usual discrepancy, really high prices for individuals and academia pays less than 1/3 of that. At least it used to be that way. It's hard to justify to shell out almost $2k just to try something out since I normally don't do much work that would need it.

BTW, this is for a uC so it'll be one of those shoe horn and sledge hammer jobs.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Academic pricing is based on the presumptions that (1) getting the tools into the hands of future engineers will lead to future sales and (2) academic institutions often don't have that much money around anyway, so some sale is better than no sale. The student edition of Matlab is quite cheap -- $100!

formatting link
.

It is unfortunate that a lot of tools are clearly not priced for the "casual commercial user"... we just purchaesd a copy of Matlab to the tune of $11k (that included a bunch of toolboxes and Simulink), and while we can justify that based on the contract we'll be using it for, it definitely stings. On the other hand, it's not *too* bad when you compare its utility (if you can use it on a regular basis) to that of, e.g., a $100k network analyzer...

Whatever happened to those companies where the idea was that you'd just logon to their Windows terminal servers (or the Unix equivalent) and effectively just "rent" the tools for the time that you actually needed them? I suspect you'd be willing to pay, say, $10 to use Matlab in such a fashion, no?

---Joel

Reply to
Joel Kolstad

Hello Joel,

Future sales is what companies believe will happen. But then the student-turned-engineer files a cap-ex and his new boss throws it right back at him with a big red X across it.

Maybe I should sign up for studies of philosophy or whatever and then order? But that wouldn't be righteous.

Some EDA pricing clearly has spiraled out of sanity. OTOH that inspires new companies to jump in. When OrCad went to >$1.7K just for the schematic package I simply pulled the plug on them and switched to Cadsoft. Haven't regretted it. Even if Cadence now came back with a really sweet teaser offer it'll be too late.

Yeah but, you can often find a used analyzer that's good enough for a small fraction of that. I have done that a lot, bought stuff when companies closed shop, needed to upgrade and so on. In fact, more than half my lab grew together that way. Works fine. There is some gear that I bought used and it looked like it had been used once, if that. Actually, there is some equipment that I could not have bought any other way because HP/Agilent quit making them.

With SW you can't usually do that because when companies upgrade they are typically not allowed sell their old licenses.

Sure. Just for kicking the tires there won't be any confidential data in design files so it could all be run via the web. The perfect biz model would be similar to prepaid cell service. They should give you free access to play with it using their demo stuff. Then when you feel you can wing it you pay for the minutes of runtime.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Hi Joerg,

Depends a lot on the company; I believe that The Mathworks probably has a large number of copies being used at all the Big Electronic Design Houses (Intel, Tektronix, Skyworks, Motorola anyone doing DSP, etc.)

I've thought of doing that just to get IEEE conference price reductions. :-) (Are you an IEEE member, Joerg? It seems to me that the prices of most everything they do these days is high enough that it's seldom justifiable unless you're working for those big guys... or in academia...)

There's a lot of overpriced EDA software out there, yet plenty of companies (cough... CADENCE... cough... MENTOR GRAPHICS... cough!) seem able to to get others to buy it on a continual basis.

Certainly true, although I think it does largely depend on what you're doing. Something like, say, a $50-100k spectrum analyzer today will be able to perform IQ demodulation and many automated measurements on various complicated standards such as WiMax or CDMA-2000. Although you can often find something like an HP 8590-series spectrum analyzer where the raw RF performance is just about as good (other than IQ demodulation, since that requires two mixers), realistically few people are going to be able to sit down and use such an instrument to make measurements on any modern digital communications system.

Slight change of topic here:

The success rate of product design when the engineering is contracted out almost seems independent of how much the engineers are being paid... in fact, it might even be higher for lower dollar contracts, because only guys who really *know* they can do the job or are really desperate for any job and will keep at it until it works will bid! On the other hand, with the high dollar contracts, there's at least usually money left over to call in someone like you when things don't work... :-)

Given your skills and reputation, I would imagine you can pretty much "cherry pick" which contracts you're willing to accept these days?

---Joel

---Joel

Reply to
Joel Kolstad

Hello Joel,

Yes, I am a member (UFFC, EMBS).

It's not just EDA. Got a dentist bill in the mail today and almost blew a gasket. A whopping $400 (four hundred!) for well under an hour's work. Should have become a dentist. They can bill whatever they want, they aren't regulated, clients are almost hostage once in the chair, it's a gold mine.

True. But it's almost a prerequisite that there is some large government contract or grant behind it. Or lots of VC.

In my experience you get what you pay for. But you are right in that a high rate does not guarantee quality of design. Had to redesign a lot of stuff that was done by "big shots". It always worked on the bench but not in production.

It's nearly all word of mouth. This means that I am still in the med electronics biz because that is where I started 20 years ago. Slowly I'd like to migrate out of that though, can't get PL insurance from any US carrier for this field anymore.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Goodness, didn't you get an estimate, or a second opinion, before your dentist began work?

About a year ago, we switched dentists (bad service); our new dentist wanted to do about $500 worth of major crown work for my wife. Decided to get a second opinion. Second opinion said the crows weren't necessary; only a filling. We switched dentists again to that one ASAP.

Michael

Reply to
mrdarrett

Wow, open source? That's really neat. I'll have to try it when I get home...

Anyone know of an open-source clone of Mathematica? Or is that too much to hope for... ;) ?

Michael

Reply to
mrdarrett

This isn't a direct answer, but at least in Matlab a very large percentage of the functions are themselves just script files that you can edit and try to glean an algorithm from... for complex algorithms there's often a pointer to a paper as a reference.

Unfortunately, I expect that Matlab/Mathematica/MathCAD/etc. are going to start heading the direction of the "Numerical Recipies in C" book did some years ago now -- cleaining that, while the algorithm there's aren't proprietary (since, after all, they usually came from public references!), the particular implementations are... which leaves you in somewhat murky legal waters when realistically a given algorithm pretty much implies 95% of the implementation. In particular, I noticed that MathCAD 6 back in the late '90s still came with a manual that provided plenty of algorithm/paper references... whereas by the time MathCAD 2002 was released, this documentation had vanished as far as I could tell... that's just one step away from deciding the algorithms themselves are the company's own IP.

That's a good reason to hope that OCtave keeps evolving!

Reply to
Joel Kolstad

schrieb

Don't know, but maybe Maxima

formatting link
might fill the bill?

Regards Martin

Reply to
Martin Blume

Ooh, I'll have to try that, thanks!

In school, we had Mathematica workstations in several of the engineering computer labs. One of the more memorable tasks I had a workstation do was to solve a system of nonlinear equations - about 100 of them (for sizing a distillation column).

Thanks,

Michael

Reply to
mrdarrett

Octave is quite compatible with MatLab.

I like the SciLab syntax and environment better, but it's _not_ friendly to beginners (but then neither are MatLab or Octave).

SciLab does have routines to translate a matlab script into a scilab script. Typing 'help matlab' at the command line will get you quite a few hits. The conversion covers the *.m files, but I'm not sure what the *.p or *.fig files do, or if SciLab can handle them. There is a SciLab newsgroup on which you can post your questions -- comp.soft-sys.scilab.org.

Absent considerations of being compatible with industry, I recommend SciLab over MatLab - even if both were free. If you're going to do a lot of work with customers who want m files then you're kind of stuck with MatLab or Octave, but I haven't had that happen to me more than once in the last three years.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google?  See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html
Reply to
Tim Wescott

Hello Michael,

It's rework on something done in Europe and there aren't too many that would even touch it.

Estimates with dentists are often elusive. When you are sitting there and the pain is shooting up to the cranium there ain't much motivation to shop ;-)

Well, we did shop around and for most work like root canals the others were even higher. Plus our dentist is really good. It's just that sometimes the bills make you cringe.

If we would just live closer to the Mexican border this burden would be much less heavy.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Hello Tim,

That's what I had tried, using the "loadmatfile" routine. I tried several m-files and got errors and aborts on all of them.

In the old days things were easier. Universities would compile everything into a nice DOS executable like the program by Prof.Mildenberger's team. It has a wave digital filter section which is what I am after. I had bought this program from them around 1990. Fair price and most of the money went to good cause. Later it became shareware but now he is retired and his site is down.

Nowadays they often post vast piles of files, either for some obscure operating system system or for a math program that they happened to have. I really don't understand why they don't walk the last mile and create a nice stand alone package. They could make money from that and most faculties are in dire need of funds. That would also serve as a teaching tool to students and Ph.D. candidates so they learn project management, proper documentation, and what it takes to create a saleable product.

Well, mostly I am stuck with industry standards. For example, I must create my docs in Word. In this case I'd just like to try out some filter design software but I am finding that the efforts to create an environment for these to run might outweigh the benefits. Might as well go back to notepad, sliderule and HP calculator. Plus maybe some Excel spreadsheeting.

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Email me the files -- I'll give it a whirl and either let you know or send you working ones.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google?  See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html
Reply to
Tim Wescott

Hello Tim,

Thanks. I'll do that tomorrow. Have to find out which ones are important. The docs are, well, skimpy and it seems that dozens of are are being called from the GUI at times. But at least they are all small, a few kB.

Now it's off to the barbie, pork chops tonight. Then a neighbor called saying she has a couple of rather active hornets nests. Oh man...

--
Regards, Joerg

http://www.analogconsultants.com
Reply to
Joerg

Get some freeze spray... knocks 'em right down.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
I love to cook with wine.      Sometimes I even put it in the food.
Reply to
Jim Thompson

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.