Reply to thread

Message

<blockquote data-quote="Brad Weber" data-source="post: 24152" data-attributes="member: 114">Re: DSP ProgrammingIsn&#039;t the whole point of an open architecture DSP that it is open architecture and can be configured as one needs or desires for a specific application?&nbsp; There may be some general approaches and concepts that make more sense than others, for example limiters last if they are for system protection (not necessarily last if used for some other purpose), but &quot;properly programming&quot; an open architecture DSP unit seems to be more about knowing what the algorithms do, how they affect the signal and what you are trying to do for that specific application rather than following some preset guideline.&nbsp; If you had a particular application you could probably address that in specific, otherwise the point seems to be to not be locked into specific approaches but rather to have the underlying knowledge to implement the processing in a way that makes sense for any specific application.What might be nice would be some guidelines for beginners on how to put some common algorithms together to create some typical &#039;blocks&#039; such as a channel strip equivalent or a speech processor or a look-ahead limiter and some general information such as how a X band graphic EQ is often equivalent in terms of processing to X bands of parametric EQ and the practical differences between FIR and IIR filter implementation.Added: In a very simplified perspective, most open architecture DSP devices provide algorithms that emulate analog device functions, so how you configure those devices is genrally no different than it would be in the analog world.&nbsp; There are some specifics issues such as processing latency, delay compensation, FIR versus IIR, etc. that are not relevant to the analog analogs, so maybe that creates some division for adddressing the different aspects involved.</blockquote>

[QUOTE="Brad Weber, post: 24152, member: 114"] Re: DSP Programming Isn't the whole point of an open architecture DSP that it is open architecture and can be configured as one needs or desires for a specific application? There may be some general approaches and concepts that make more sense than others, for example limiters last if they are for system protection (not necessarily last if used for some other purpose), but "properly programming" an open architecture DSP unit seems to be more about knowing what the algorithms do, how they affect the signal and what you are trying to do for that specific application rather than following some preset guideline. If you had a particular application you could probably address that in specific, otherwise the point seems to be to not be locked into specific approaches but rather to have the underlying knowledge to implement the processing in a way that makes sense for any specific application. What might be nice would be some guidelines for beginners on how to put some common algorithms together to create some typical 'blocks' such as a channel strip equivalent or a speech processor or a look-ahead limiter and some general information such as how a X band graphic EQ is often equivalent in terms of processing to X bands of parametric EQ and the practical differences between FIR and IIR filter implementation. Added: In a very simplified perspective, most open architecture DSP devices provide algorithms that emulate analog device functions, so how you configure those devices is genrally no different than it would be in the analog world. There are some specifics issues such as processing latency, delay compensation, FIR versus IIR, etc. that are not relevant to the analog analogs, so maybe that creates some division for adddressing the different aspects involved. [/QUOTE]

Verification