Matlab and Texas Instruments DSP chips

denverdoc · « **on:** 21 Feb 2005, 09:42 pm »

Anyone out there know whether you really can use the Maple Mathlab products to program the DSP chips from TI? Specifically, I am curious about the software's ability to test various filters and then to implement these on the new TI series of tri-path DSP/Class d chips?
Thinking I like this idea much better than using having a PC full time in the loop.
Thanks in advance,
John

JoshK · « **Reply #1 on:** 21 Feb 2005, 10:46 pm »

I thought Matlab was owned by Mathsoft not Maple. I have not heard that before, but that would be cool since I know Matlab real well.

denverdoc · « **Reply #2 on:** 21 Feb 2005, 11:20 pm »

Josh,
Likely my confusion as I have never owned either product, and still looking at a software program with nuff juice to do line array dipole modeling. Anyhow, as i was reviewing Matlabs feature list, noticed an item suggesting communication with, testing , and programming of TI products. Went to TI, and under third party hardware, sure enough, reciprocal mention. Then the little light bulb went off and got curious whether the C2000 spec was robust enough to interface with the new TI chipsets. Holy s--t it said twas so, and a whole new dawn of DIY possibilities emerged. After daydreaming a minute or so, dawned om me to ask whether this match is one that truly takes place on earth or only in the minds of marketeers,
John

amitm · « **Reply #3 on:** 22 Feb 2005, 12:00 am »

Quote from: denverdoc

Josh,
Likely my confusion as I have never owned either product, and still looking at a software program with nuff juice to do line array dipole modeling. Anyhow, as i was reviewing Matlabs feature list, noticed an item suggesting communication with, testing , and programming of TI products. Went to TI, and under third party hardware, sure enough, reciprocal mention. Then the little light bulb went off and got curious whether the C2000 spec was robust enough to interface with the new TI chipsets. Holy s--t ...

Matlab does support real-time implementations -- however, I would suggest writing programs in C directly after testing the prototypes in Matlab. Typically, MATLAB translations are extremely inefficient -- that may have changed in the recent couple of years. If so, more power to you. I can tell you that the way you write programs in MATLAB make a big difference in the speed at which they run. A trivial example -- try to implement a filter using a loop, and then implement the same thing as a scalar product.

--amit

Occam · « **Reply #4 on:** 22 Feb 2005, 12:50 am »

I'm not familiar with the use of Matlab (or it clones) for your specific needs as a generative tool, though obviously it has been used extensively for filter coefficient calculations and and other quantitative tasks.
You might find this link helpful (at least it contains links to available. free, academically supported Matlab clones) -
http://www.dspguru.com/
simply enter 'matlab' in the search box.
FWIW

denverdoc · « **Reply #5 on:** 22 Feb 2005, 01:48 am »

Occam. others.

By jove, this might just then be doable! Josh, what do you think?
John