<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">Not only floating point operations, they also have DSP and SIMD instructions. There’s a library of functions (CMSIS-DSP) which makes these operations easy to use.<div class=""><br class=""></div><div class="">Ben</div><div class=""><br class=""><div><blockquote type="cite" class=""><div class="">On 20 Feb 2018, at 14:19, Martin Klang <<a href="mailto:mars@pingdynasty.com" class="">mars@pingdynasty.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class="">
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" class="">
<div text="#000000" bgcolor="#FFFFFF" class=""><p class=""><br class="">
</p><p class="">STMs: the F4 and F7 are Cortex M4F with single cycle floating
point operations. Not sure there's much to gain from going fixed
point, assuming 32bits. With 16 bits fixed point you can
potentially get two operations per cycle. Or you could go 8-bit...<br class="">
</p><p class=""><br class="">
</p><p class="">Martin<br class="">
</p>
<br class="">
<div class="moz-cite-prefix">On 18/02/18 19:20, Bruno Afonso wrote:<br class="">
</div>
<blockquote type="cite" cite="mid:CA+z2k4tOX8S5b=MiAieNkdgzLRnUmBP0yfv9hE0k1spncHQP+g@mail.gmail.com" class="">
<div dir="ltr" class="">If you're willing to go into DSP land then you
could also look at blackfin, Monome's Aleph uses open source
tools for develop it. For STMs, the only shot would be to
attempt to use fixed point, not a lot of juice for floating
point stuff but like other's mention, it may be a lot of work..
maybe more for as a challenge? :)
<div class=""><br class="">
</div>
<div class="">b<br class="">
<div class=""><br class="">
</div>
</div>
</div>
<br class="">
<div class="gmail_quote">
<div dir="ltr" class="">On Sat, Feb 17, 2018 at 5:24 AM <<a href="mailto:rsdio@audiobanshee.com" moz-do-not-send="true" class="">rsdio@audiobanshee.com</a>>
wrote:<br class="">
</div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">I haven’t
developed for SHARC, yet, but I can say that I shipped the
Soundplane without spending much on the compiler/assembler.
That was based on the TMS320VC5506, a member of the
lowest-power C55x line. You might want to look at the C6000
series or OMAP, such as the OMAP-L1x line. I recall that the
tools run free when you have one of the evaluation boards
connected to your computer.<br class="">
<br class="">
What about some of the less “fancy" SHARC DSP options, like
the ADSP-2191M used in the DSI eVoLver? For some reason, I
assume that a 16-bit DSP would cost less. Dave (and/or his
firmware programmer) managed to coax 24-bit audio out of this
16-bit DSP by using specific features. I haven’t looked into
the tools for that chip, but maybe they’re priced lower than
the tools for the flagship chips.<br class="">
<br class="">
Brian<br class="">
<br class="">
<br class="">
On Feb 16, 2018, at 3:50 PM, Tim Ressel <<a href="mailto:timr@circuitabbey.com" target="_blank" moz-do-not-send="true" class="">timr@circuitabbey.com</a>> wrote:<br class="">
> I hear ya. My comment about the cost of tools referred to
the compilers. I'd love to use a Sharc, but it looks like at
least $1K for the compiler.<br class="">
><br class="">
> --tr<br class="">
><br class="">
> On 2/16/2018 1:48 PM, <a href="mailto:rsdio@audiobanshee.com" target="_blank" moz-do-not-send="true" class="">rsdio@audiobanshee.com</a> wrote:<br class="">
>> You really need a DSP for this, and not a
general-purpose CPU like ARM. Even though ARM has a DSP
instruction or two, it's a far cry from a total system
designed for signal processing. Families like the TMS320 have
been evolving for decades - literally 35 years - to optimize
this sort of thing. Literally every aspect of the chip has
been tweaked to optimize signal processing.<br class="">
>><br class="">
>> Texas Instruments has cheap demo platforms with audio
I/O and the tools come free with those $50 evaluation boards.
There are a lot of open-source routines, so if you need FFT in
TMS320 assembly then it's there. You can call that from C and
keep your overall design simple.<br class="">
>><br class="">
>> Texas Instruments even has chips that are
dual-processor inside - one TMS320 for DSP, plus and ARM for
the higher level stuff and maybe even some audio processing.
Look for OMAP chips. The C6000 series of DSP would be a good
choice because it supports floating point (I've worked with
the C5500 series that is fixed point, and that's a lot of
tedium but worth it if you want to run off of batteries).<br class="">
>><br class="">
>> Another good option would be SHARC.<br class="">
>><br class="">
>> If you choose a non-DSP chip, everything will be less
efficient. That either means loss of features or higher
temperatures and shorter battery life.<br class="">
>><br class="">
>> Brian Willoughby<br class="">
>> Sound Consulting<br class="">
>><br class="">
>> On Feb 16, 2018, at 11:47 AM, Tim Ressel <<a href="mailto:timr@circuitabbey.com" target="_blank" moz-do-not-send="true" class="">timr@circuitabbey.com</a>> wrote:<br class="">
>>> I wonder if it would be possible to do a parallel
processor scheme where one proc handles the early stuff and
another to handle the longer time stuff. each proc would
output via a codec and those outputs would get summed. Hmm...<br class="">
>>><br class="">
>>> On 2/16/2018 11:38 AM, Eric Brombaugh wrote:<br class="">
>>>> I suspect that STM32 doesn't have the
horsepower you'll need to do a useful IR reverb. There are
several fairly efficient FFT in the CMSIS libraries from ARM
but even using those the best you can do is about a 4096 FFT
running at less than 48kHz with long latency and large
overlaps.<br class="">
>>>><br class="">
>>>> On 02/16/2018 12:24 PM, Tim Ressel wrote:<br class="">
>>>>> Still, even with all that jigery-pokery,
we're going to need a bigger boat, er, processor. I'd like to
avoid processor choices that needs pricey tools. STM32 would
be nice. Of course some good ol' fashion assembly code, highly
optimized, would help things. Its been a while since I went
down that rabbit hole. I wonder if someone has an optimozed
FFT library for Cortex Mn�<br class="">
<br class="">
<br class="">
_______________________________________________<br class="">
Synth-diy mailing list<br class="">
<a href="mailto:Synth-diy@synth-diy.org" target="_blank" moz-do-not-send="true" class="">Synth-diy@synth-diy.org</a><br class="">
<a href="http://synth-diy.org/mailman/listinfo/synth-diy" rel="noreferrer" target="_blank" moz-do-not-send="true" class="">http://synth-diy.org/mailman/listinfo/synth-diy</a><br class="">
</blockquote>
</div>
<br class="">
<fieldset class="mimeAttachmentHeader"></fieldset>
<br class="">
<pre wrap="" class="">_______________________________________________
Synth-diy mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Synth-diy@synth-diy.org">Synth-diy@synth-diy.org</a>
<a class="moz-txt-link-freetext" href="http://synth-diy.org/mailman/listinfo/synth-diy">http://synth-diy.org/mailman/listinfo/synth-diy</a>
</pre>
</blockquote>
<br class="">
</div>
_______________________________________________<br class="">Synth-diy mailing list<br class=""><a href="mailto:Synth-diy@synth-diy.org" class="">Synth-diy@synth-diy.org</a><br class="">http://synth-diy.org/mailman/listinfo/synth-diy<br class=""></div></blockquote></div><br class=""></div></body></html>