Reply by Ronaldo Souza August 16, 20102010-08-16
> Ron N. wrote: > You can take this very short simple Tiny Basic code for > windowed-Sinc resampling, and optimize as needed. > http://www.nicholson.com/rhn/dsp.html#3 > Common optimizations include using a cache, or table lookup, > or table interpolation for the windowed-Sinc function instead > of (slowly) recalculating it for every tap for every point.
Thanks a bunch, Ron. Best Regards, Ronaldo
Reply by Ronaldo Souza August 16, 20102010-08-16
> Tim Wescott wrote: > > Applying a library like Secret Rabbit Code blindly may work, and it > may not. If it does work, you're lucky. If it doesn't work, and it > breaks in the lab, you're a lot more lucky than if it breaks all over > a customer. > So try it, test the heck out of it, and keep your fingers crossed.
I'm painfully aware of this and believe me, my fingers are crossed so tightly that I might get gangrene or something.
> There's quite a few regulars in this group that consult; if it breaks > in a way that convinces the powers that be that it needs a fix in a > hurry you've got some experts available to trade money for fixes. Or > you can ask all the background questions, and we'll help you get > through it slowly.
IMHO, consulting with somemone like Vladimir would be not only the correct choice, but also the best bang for the buck, but I'm "just" a tech guy and "have no understanding of our financial problems..." so, I'll give it a try and if (when?) it all tumbles down, I'll have to frame the problem from a $ POV, so they'll understand.
> Get Rick Lyons's book ("Understanding Digital Signal Processing"). > If you can get through it fast enough then maybe by the time the > algorithm breaks you'll know why and can fix it.
Thanks. <browsing Amazon...> Best regards, Ronaldo
Reply by Ron N. August 16, 20102010-08-16
On Aug 16, 10:58&#4294967295;am, "Ronaldo Souza" <nos...@nospam.org> wrote:
> Can anyone recommend a good resampling lib? I need to handle rather > small buffers (100s of Ks of signed 16 bits) but also really long > streams, often covering 24+ hours @250sps, so speed is important. > > BTW, tipically I'll be downsampling from 250sps to 200sps, but there > will be other scenarios within the range 100sps <= In/Out sampling freq > <= 5000sps.
You can take this very short simple Tiny Basic code for windowed-Sinc resampling, and optimize as needed. http://www.nicholson.com/rhn/dsp.html#3 Common optimizations include using a cache, or table lookup, or table interpolation for the windowed-Sinc function instead of (slowly) recalculating it for every tap for every point. YMMV. -- rhn A.T nicholson d.0.t C-o-M
Reply by Tim Wescott August 16, 20102010-08-16
On 08/16/2010 12:23 PM, Ronaldo Souza wrote:
>> Rob Gaddi wrote: >> And unfortunately, if you don't you'll wind up misapplying any >> library you can find. Theory/foundations first, code second. > > > I'm the first to say I'd have to have some time to really understand > this stuff, but the powers that be think otherwise. :(
Applying a library like Secret Rabbit Code blindly may work, and it may not. If it does work, you're lucky. If it doesn't work, and it breaks in the lab, you're a lot more lucky than if it breaks all over a customer. So try it, test the heck out of it, and keep your fingers crossed. There's quite a few regulars in this group that consult; if it breaks in a way that convinces the powers that be that it needs a fix in a hurry you've got some experts available to trade money for fixes. Or you can ask all the background questions, and we'll help you get through it slowly. Get Rick Lyons's book ("Understanding Digital Signal Processing"). If you can get through it fast enough then maybe by the time the algorithm breaks you'll know why and can fix it. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" was written for you. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by Ronaldo Souza August 16, 20102010-08-16
> Vladimir Vassilevsky wrote: > Don't waste your time; this is not a place where they serve free > lunch.
ok. Rgds, Ronaldo
Reply by Ronaldo Souza August 16, 20102010-08-16
> Rob Gaddi wrote: > And unfortunately, if you don't you'll wind up misapplying any > library you can find. Theory/foundations first, code second.
I'm the first to say I'd have to have some time to really understand this stuff, but the powers that be think otherwise. :( Regards, Ronaldo
Reply by Vladimir Vassilevsky August 16, 20102010-08-16

Ronaldo Souza wrote:

>>Vladimir Vassilevsky wrote: >> >>>Ronaldo Souza wrote: >>>Can anyone recommend a good resampling lib? >> >>Do it yourself. No big deal. A project for one day. > > > > If I knew how to do it, I wouldn't waste the group's time, would I?
Don't waste your time; this is not a place where they serve free lunch.
Reply by Rob Gaddi August 16, 20102010-08-16
On 8/16/2010 11:51 AM, Ronaldo Souza wrote:
>> Vladimir Vassilevsky wrote: >>> Ronaldo Souza wrote: >>> Can anyone recommend a good resampling lib? >> >> Do it yourself. No big deal. A project for one day. > > > If I knew how to do it, I wouldn't waste the group's time, would I? > > Regards, > Ronaldo
And unfortunately, if you don't you'll wind up misapplying any library you can find. Theory/foundations first, code second. -- Rob Gaddi, Highland Technology Email address is currently out of order
Reply by Ronaldo Souza August 16, 20102010-08-16
> Vladimir Vassilevsky wrote: > > Ronaldo Souza wrote: > > Can anyone recommend a good resampling lib? > > Do it yourself. No big deal. A project for one day.
If I knew how to do it, I wouldn't waste the group's time, would I? Regards, Ronaldo
Reply by Ronaldo Souza August 16, 20102010-08-16
> Tim Wescott wrote: > > Search on "Secret Rabbit Code".
Thanks, Tim. <googling..> :) Regards, Ronaldo