SigmaDSP(AD) miniDSP(TI)

I am not familiar with the TI part, but I can provide some basic information about SigmaDSP and its strengths and weaknesses.

Small DSP core optimized for audio post-processing

Not suitable for frequency-domain processing and audio decompression or decoding algorithms

24-bit audio I/O and 28-bit fixed-point audio path, with 56-bit accumulators, ideal for 24-bit audio processing

Flexible programming model using SigmaStudio graphical tool and a library of hundreds of preset algorithms

Not programmable in C or assembly code, but algorithms can be created from low-level blocks in the library

Hello Brett,

Now with the release of SS4SHARC, frequency-domain processing and audio decompression or decoding algorithms should also be possible making this platform more powerful and complete. I have tons of questions at this time about the SS4SHARC, its use and capabilities. My first attempt will be to come up with a minimalist eval board concept to get 'hands-on' on this. Very excited!!

KKSL.

Hi KKSL,

That's great. By the way, there will be a new forum showing up on EngineerZone very soon that will be devoted to Sharc development in SigmaStudio. I will still only be supporting SigmaDSP, so any Sharc-related questions will be answered by the Sharc engineers.

Brett,

In addition to SigmaStudio / CrossCore for SHARC, I've noticed some blocks bordering on the frequency domain (pitch shift, Hilbert transform) finding their way into the newer SigmaDSPs. Is there something fundamentally different about the 175 MIPS core that enables these blocks, or is it merely the extra horsepower? I get the impression that at AD a wonderful convergence is proceeding toward the next level of easy (and fun!) DSPs and IDEs. Up to very recently, DSP code development has been where microprocessor programming was back in the '70s -- an exclusive club of digital gurus with pricy tools. AD has done more than anyone to change that. Keep up the good work!

Bob

Hi Bob,

Those blocks that are bordering on frequency domain processing are still done in the time domain, just using some workarounds that we can achieve thanks to the extra horsepower. For example, we have a nice noise removal algorithm that is fundamentally based on frequency-domain processing concepts, but since the SigmaDSPs it is implemented on cannot perform FFT or IFFT operations, we had to do some tricks to implement it as a time-domain function (for example, limiting it to a 12 kHz sample rate, effectively making it a voice-only algorithm).

True frequency-domain processing might find its way into SigmaDSP cores in the future.

I think the new SigmaStudio for Sharc library will really open things up for more target applications and easier development for the higher end Sharc processors.

So Shark now supports SigmaStudio?? Great!! Guyz could you then just enlighten me about a couple of things:

1. Where to download SS for shark??

2. Is there any board with 24/192+ ADC/DAC for shark?

http://www.analog.com/en/dsp-software/sh_sigst_00/sw.html

http://www.analog.com/en/press-release/10-24-2012-ADI-Releases-Graphical-Audio-Develop/press.html

This should help you to navigate to the relevant links.

Thanks!!

'The SigmaStudio for SHARC tool is powerful enough to satisfy the demands of experienced DSP engineers, while allowing audio developers with little or no DSP code writing experience to easily implement a SHARC audio processor in their design.'

Will this still remain true?

KKSL

I'm not an expert on the SHARC side of SigmaStudio, but from what I understand, you will still be able to graphically create a signal flow, and there will be some basic framework that is pre-made and should allow you to easily compile and run the software on the SHARC. So, the short answer is "yes."

This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

Thank you,
EZ Admin