future of DSP in digital communication, wireless industries

Hi,

While working for audio systems builder, I've been witnessing DSP work
slowly disappearing. We are no longer developing audio algorithms as we
used to. The optimization work is disappearing too. Either we buy DSP
with software and use it as a black box or we simply use the already
optimized DSP library. The work that's left for us is the integration
and the test.

How is it for the wireless and the digital communication fields? It
looks like same thing will happen IMO. Since the modems are standard,
few people can develop them and used by the rest.

<dspinwallstreet@gmail.com> wrote in message
news:1147730637.205989.318980@v46g2000cwv.googlegroups.com...
> Hi,
>
> While working for audio systems builder, I've been witnessing DSP work
> slowly disappearing. We are no longer developing audio algorithms as we
> used to. The optimization work is disappearing too. Either we buy DSP
> with software and use it as a black box or we simply use the already
> optimized DSP library. The work that's left for us is the integration
> and the test.
>
> How is it for the wireless and the digital communication fields? It
> looks like same thing will happen IMO. Since the modems are standard,
> few people can develop them and used by the rest.
>

In every field, the trend is the same. First batch I know of are the
computer programmers
in the last few years, now audio engineers like you guys. Maybe few years
later, communication
modem designers like me willl have to look for other careers also. We see
channel coding,
encryption algorithms and modulations, various transforms (FFT, WWT)
replaced by standard
chips and IPs, it won't take long before we become system builders in my
opinion.

Hi,

Does system builder means system integrator ? Integrating bits and
pieces from other blackbox to build a complete system ? I interested in
wireless modem design using DSP. What is the prospect for this area ?
What kind of work is usually perform in this area ? Implementing
existing algorithm using DSP with reduce complexity ? Or just integrate
chips from other company......

Regards,
Arthur.

Mr. Ken wrote:
> <dspinwallstreet@gmail.com> wrote in message
> news:1147730637.205989.318980@v46g2000cwv.googlegroups.com...
> > Hi,
> >
> > While working for audio systems builder, I've been witnessing DSP work
> > slowly disappearing. We are no longer developing audio algorithms as we
> > used to. The optimization work is disappearing too. Either we buy DSP
> > with software and use it as a black box or we simply use the already
> > optimized DSP library. The work that's left for us is the integration
> > and the test.
> >
> > How is it for the wireless and the digital communication fields? It
> > looks like same thing will happen IMO. Since the modems are standard,
> > few people can develop them and used by the rest.
> >
>
> In every field, the trend is the same. First batch I know of are the
> computer programmers
> in the last few years, now audio engineers like you guys. Maybe few years
> later, communication
> modem designers like me willl have to look for other careers also. We see
> channel coding,
> encryption algorithms and modulations, various transforms (FFT, WWT)
> replaced by standard
> chips and IPs, it won't take long before we become system builders in my
> opinion.

I don't often top-post
*BUT* you ask _WRONG_ question.
I'll take role of 'elder', though not as old as Jerry
      [ Owl ducks for cover ;]

Back in 60's I bailed from an Ivy League BSEE program.
My 'plaint was "all theory -- no practice"

HOWEVER, since I had a very sound theory background, I was rarely 
unemployed. And when employed it was "enumerative" LOL!

Do you have skills to offer employer other than "placing NUT-A on 
'loose-screw Z'?

Do you get a hint in that I *NEVER* commented on _ANY_ application???

If you wonder why I'm so "cruel and 'uncaring' " ask me privately. [my 
reply-to is valid]



dspinwallstreet@gmail.com wrote:

> Hi,
> 
> While working for audio systems builder, I've been witnessing DSP work
> slowly disappearing. We are no longer developing audio algorithms as we
> used to. The optimization work is disappearing too. Either we buy DSP
> with software and use it as a black box or we simply use the already
> optimized DSP library. The work that's left for us is the integration
> and the test.
> 
> How is it for the wireless and the digital communication fields? It
> looks like same thing will happen IMO. Since the modems are standard,
> few people can develop them and used by the rest.
> 

dspinwallstreet@gmail.com wrote:
>
> While working for audio systems builder, I've been witnessing DSP work
> slowly disappearing. We are no longer developing audio algorithms as we
> used to. The optimization work is disappearing too. Either we buy DSP
> with software and use it as a black box or we simply use the already
> optimized DSP library. The work that's left for us is the integration
> and the test.

What I've been observing during the recent years though is that
sometimes (I'm tepmted to say "usually") certain
"super-highly-integrated-chips" that are supposed to dramatically speed
up the  development process, in reality turn out to be harder to
program/configure adequately (several hundreds registers, poor
documentation, unexpected glitches, poor internal visibility, etc.),
compared to, say, some in-house DSP implementation of the particular
algorithm (like, plain old fixed-point TI c54 function). And certainly
less customizable in the long run.

Sometimes it helps projects, sometimes it really hurts, but it's not
that deterministic, and of course the theoretical background/experience
helps in any case.

I wonder if these are only my feelings...

Regards,
Dmitry.

Wouldn't more reliable and better performing chips and libraries come
out?
Right. Theoretical background would help whether developing from the
scratch or using the chip/library. Does this mean that demand for DSP
engineer on this field will slowly decrease?

I think what is lost is "product differentiation." When everyone
uses the same chips and libraries, there are no longer any
optmizations and/or innovations in the product development,
so all implementations offer similar abilities/costs/sizes. Think
QualComm.

Fight the flow - think for yourself.

--Randy