Future of DSP?

"Andrew Reilly" <andrew-newspost@areilly.bpc-users.org> wrote in message 
news:611dsoF1t893dU1@mid.individual.net...
> On Thu, 07 Feb 2008 16:22:45 -0500, Jerry Avins wrote:
>> The question seems to
>> ask what will replace DSP, not what new techniques and devices will be
>> used.
>
> Maybe it'll all be quantum computers making inscrutable guesses at
> infinite rate and magically picking the right one...  Then we can all
> retire, as you say.
>
> [I've not looked at the maths or theory of quantum computers, but the way
> that people describe their *use* makes them seem indistinguishable from
> the analog (op-amp) computers of yore: apply enough gain and feedback and
> bingo: you get the right answer once the system reaches steady state
> (unless you hit saturation or noise-floor on an internal node).  Yes, I
> was lucky enough to play with a large, pluggable analog computer at Uni:
> warped my brain ever since.  Are the quantum computers *really* doing
> anything different?]
>
> Cheers,
>
> -- 
> Andrew
Probably  ;-)

Hello All,

      It's nice discussion and great to hear from some people
      who have experienced analog computers to the days dsp.

      Yes someone rightly corrected me "alternative to what"?

      I just wanted to know if DSP will continue to have a
      promising future. If so what application areas will
      be key drivers/differentiators to the growth.

      I know lot many innovative products will come. But as
      things are becoming more and more complex day-by-day
      do engineers at individualistic levels will have a roll
      to play? 

      Also will hardware drive the growth or will it be software
      or maybe both. 

      I very well understand that this topic is very near and
      dear to many peoples heart as all our futures lie in what
      we visualize today and continously work towards it.

Thanks and Best Regards
Bharat Pathak

On Feb 8, 5:27 am, Vladimir Vassilevsky <antispam_bo...@hotmail.com>
wrote:
> bharat pathak wrote:
> > For sometime I have been wondering what will be the future of DSP
> > maybe 5 years down the line or 10 years down the line.
>
> The cost of the answer to this question is $1.000.000.000.
>
> > Based on
> > that future of many people related to dsp "dsprelated people" will
> > also take shape. Any insights?
>
> It is very difficult to predict anything, especialy the future.
>
> > Why I am asking is many alternative technologies are picking up fast.
> > From some people I heard is nano-technology will be one of them.
>
> The exaustion of the oil resources will ignite the nuclear conflict
> between the northern mexico and the pan-european caliphate. After that,
> the common trades like horse-breeding and metal working would be in the
> great demand.
>
> Vladimir Vassilevsky
> DSP and Mixed Signal Design Consultanthttp://www.abvolt.com

Yes comrad Vladimir - a wise suggestion. Our part will re-unite after
the nuclear hollocaust.

On Feb 8, 3:55 pm, "bharat pathak" <bha...@arithos.com> wrote:
> Hello All,
>
>       It's nice discussion and great to hear from some people
>       who have experienced analog computers to the days dsp.
>
>       Yes someone rightly corrected me "alternative to what"?
>
>       I just wanted to know if DSP will continue to have a
>       promising future. If so what application areas will
>       be key drivers/differentiators to the growth.
>
>       I know lot many innovative products will come. But as
>       things are becoming more and more complex day-by-day
>       do engineers at individualistic levels will have a roll
>       to play?
>
>       Also will hardware drive the growth or will it be software
>       or maybe both.
>
>       I very well understand that this topic is very near and
>       dear to many peoples heart as all our futures lie in what
>       we visualize today and continously work towards it.
>
> Thanks and Best Regards
> Bharat Pathak

In ten years not a heluva lot except things will become smaller and
faster. Ultra fast pocket pcs for instance. Quantum computers are
still a long way off. However, with the invasion of the lizzard people
and the new dawn of mankind who knows what's just around the corner.

On 7 Feb, 17:08, "bharat pathak" <bha...@arithos.com> wrote:
> For sometime I have been wondering what will be the future of DSP

Well, I won't guess at any time scales, but DSP will cease
to exist as an independent dicipline -- if it ever was one.

As hardware becomes smaller and cheaper, there will
eventually be a time when one can buy off-the shelf
DSPs for most jobs. We may not be there in the next
10 or 20 years, but at some point even the penny-
counting mass-market devices will be run by standard
CPUs -- not even DSPs. When that happens the market
for specialist DSP programmers will vanish; anyone
who is able to read the matlab user guide will do.

As data analysis tools become widely available, users
will cease think for them selves but instead rely on
canned solutions. Our friends at the matworks have
helped us a a long way down that road already; students
don't learn filter design from scratch anymore, but
rather how to get matlab to cough out some sorts of
coefficients.

That's the way things are going, and have been for
a long time already. For better and for worse.

Rune

On Thu, 07 Feb 2008 23:54:55 -0800, Rune Allnor wrote:
> On 7 Feb, 17:08, "bharat pathak" <bha...@arithos.com> wrote:
>> For sometime I have been wondering what will be the future of DSP
> 
> Well, I won't guess at any time scales, but DSP will cease to exist as
> an independent dicipline -- if it ever was one.

Well, DSP *coding* might stop being particularly interesting.  IMO it was 
never much more than an entertaining chore anyway.  Like crossword 
puzzels.  But the digital processing of signals: that's mathematics.  
That's deep thought and invention.  That's never going away.

> As hardware becomes smaller and cheaper, there will eventually be a time
> when one can buy off-the shelf DSPs for most jobs. We may not be there
> in the next 10 or 20 years, but at some point even the penny- counting
> mass-market devices will be run by standard CPUs -- not even DSPs. When
> that happens the market for specialist DSP programmers will vanish;
> anyone who is able to read the matlab user guide will do.

Matlab doesn't think for you.  It can give you the coefficients for a 
kick-arse filter implementation, even round them to machine precision if 
you like.  It can't tell you what filter you want, though, nor how to 
best solve your real problem.  That's what DSP engineers do.

> As data analysis tools become widely available, users will cease think
> for them selves but instead rely on canned solutions. Our friends at the
> matworks have helped us a a long way down that road already; students
> don't learn filter design from scratch anymore, but rather how to get
> matlab to cough out some sorts of coefficients.
> 
> That's the way things are going, and have been for a long time already.
> For better and for worse.

Well sure: solved problems will be easy to solve.  No news there.  It's 
the new problems that make life interesting.  Don't tell us that you have 
no problems!

-- 
Andrew

Rune Allnor wrote:

> On 7 Feb, 17:08, "bharat pathak" <bha...@arithos.com> wrote:
> 
>>For sometime I have been wondering what will be the future of DSP
> 
> 
> Well, I won't guess at any time scales, but DSP will cease
> to exist as an independent dicipline -- if it ever was one.
> 
> As hardware becomes smaller and cheaper, there will
> eventually be a time when one can buy off-the shelf
> DSPs for most jobs. We may not be there in the next
> 10 or 20 years, but at some point even the penny-
> counting mass-market devices will be run by standard
> CPUs -- not even DSPs. When that happens the market
> for specialist DSP programmers will vanish; anyone
> who is able to read the matlab user guide will do.
> 
> As data analysis tools become widely available, users
> will cease think for them selves but instead rely on
> canned solutions. Our friends at the matworks have
> helped us a a long way down that road already; students
> don't learn filter design from scratch anymore, but
> rather how to get matlab to cough out some sorts of
> coefficients.
> 
> That's the way things are going, and have been for
> a long time already. For better and for worse.

The proliferation increases the market rather then collapses it. MatLab 
and such are only creating the false sense of simplicity.

Isn't it sad that the immense computing power and the intellectual 
effort is used solely to provide the meaningless chit-chat over the 
wireless, streaming pornography and guiding missiles?

VLV

On 8 Feb, 12:34, Andrew Reilly <andrew-newsp...@areilly.bpc-users.org>
wrote:
> On Thu, 07 Feb 2008 23:54:55 -0800, Rune Allnor wrote:
> > On 7 Feb, 17:08, "bharat pathak" <bha...@arithos.com> wrote:
> >> For sometime I have been wondering what will be the future of DSP
>
> > Well, I won't guess at any time scales, but DSP will cease to exist as
> > an independent dicipline -- if it ever was one.
>
> Well, DSP *coding* might stop being particularly interesting. &#4294967295;IMO it was
> never much more than an entertaining chore anyway. &#4294967295;Like crossword
> puzzels. &#4294967295;But the digital processing of signals: that's mathematics. &#4294967295;
> That's deep thought and invention. &#4294967295;That's never going away.

Did you or anybody else ever get paid for solving ceosswrds...?

> > As hardware becomes smaller and cheaper, there will eventually be a time
> > when one can buy off-the shelf DSPs for most jobs. We may not be there
> > in the next 10 or 20 years, but at some point even the penny- counting
> > mass-market devices will be run by standard CPUs -- not even DSPs. When
> > that happens the market for specialist DSP programmers will vanish;
> > anyone who is able to read the matlab user guide will do.
>
> Matlab doesn't think for you. &#4294967295;It can give you the coefficients for a
> kick-arse filter implementation, even round them to machine precision if
> you like. &#4294967295;It can't tell you what filter you want, though, nor how to
> best solve your real problem. &#4294967295;That's what DSP engineers do.

Having worked with data analysis for a couple of decades,
I would agree with you. But *my* opinin  doesn't matter;
why would anybody who is held accountable for real $$$ pay
for somebody with the correct papers if one can get away with
paying $2000 for a software license that puts everything
needed at the tips of a novice's fingers?

> > As data analysis tools become widely available, users will cease think
> > for them selves but instead rely on canned solutions. Our friends at the
> > matworks have helped us a a long way down that road already; students
> > don't learn filter design from scratch anymore, but rather how to get
> > matlab to cough out some sorts of coefficients.
>
> > That's the way things are going, and have been for a long time already.
> > For better and for worse.
>
> Well sure: solved problems will be easy to solve. &#4294967295;No news there. &#4294967295;It's
> the new problems that make life interesting. &#4294967295;Don't tell us that you have
> no problems!

Nothing would be further from my mind!

Rune

On 8 Feb, 17:05, Vladimir Vassilevsky <antispam_bo...@hotmail.com>
wrote:

> Isn't it sad that the immense computing power and the intellectual
> effort is used solely to provide the meaningless chit-chat over the
> wireless, streaming pornography and guiding missiles?

Some nationalist history books claim that the Scadinavians
(Vikings, well, Swedes if one goes anal on the details,but
let's not be bogged down in *that* quagmire...;) played a crucial
parts some 1000 years ago in founding what would later become
the Russian empire. There have been little evidence to
substantiate such claims; your post might just be one...

Rune

Rune Allnor wrote:


> 
> Some nationalist history books claim that the Scadinavians
> (Vikings, well, Swedes if one goes anal on the details,but
> let's not be bogged down in *that* quagmire...;) played a crucial
> parts some 1000 years ago in founding what would later become
> the Russian empire. There have been little evidence to
> substantiate such claims; your post might just be one...

You do consume the liquid which increases the hardware abstraction 
level! BTW there is no future hence there is no future of DSP. Amen.

A legendary person called Rodrekr or Rorek or Rorik is considered to be 
a founder of the dynasty which later became royal and ruled Russia from 
10 to 16 century. Presumably he was a Danish or Swiss warlord, and the 
whole story happened in ~860. The interesting part of the legend is that 
Rorik was voluntarily invited by the natives to rule, since they didn't 
have a competent manager of their own.

VLV














> 
> Rune