On 27 Mar, 15:00, Tim Wescott <t...@seemywebsite.com> wrote:> On Wed, 25 Mar 2009 15:53:09 -0700, Rune Allnor wrote: > > On 25 Mar, 22:51, "noodle22" <jw970...@yahoo.com> wrote: > >> I know this is a subjective question but I've always found DSP > >> interesting and have considered doing a career in it (it certainly > >> beats business apps). > > > First of all, make up your mind about what you want: An 'interesting' > > job or making a living. > > >> Is the field growing, shrinking, or not changing? �Is there much demand > >> and are there many jobs available? > > >> What type of work do DSP engineers do? > > > I'd say DSP is a transitionary discpline, like radio telegraphists used > > to be at sea in the first half of the 20th century: The operators were > > there and necessary when the field was new and worked with morse code. > > Once voiced radio was possible, the specialized skills of the radio > > operator were no longer necessary, as anyone could use a radio set. > > > DSP is headed in the same direction. > > (snip) > > I don't see it heading in that direction at all,It is. DSP is about designing - not building - the radio systems. Sure, there will be a one or two decades with trial'n error, but then the indsutry will gravitate towards one or two 'standard' solutions. 100 years ago there was lots of debate about power supplies; AC vs DC, voltages and frequencies. Now the two standard systems are 60 Hz / 110 V and 50 Hz / 220 V. It doesn't matter if somebody comes up with a design and can prove that it would work optimally with 109.78 Hz / 19.8 V. Unless it can work with one of the two standard solutions, no one will be interested. In 50 years there will be standard solutions for just about anything; the main trends will emerge in the next couple of decades. Once that has happened, there will be no more need for DSP designs. In the mean time, the main effort is about politics.> at least not if you can > do embedded software or FPGA work. �Canned DSP routines for FPGAs and > DSPs are crap,Sure. But that doesn't stop DSP courses gravitating into matlab signal processing toolbox training sessions...> unless you happen to need one for one of the big > applications. �As long as there are people out there making products that > do new things, there will be demand for other people who can make those > products work.I don't see that people do *new* thing; they keep doing the same old stuff but economy and engineering let them do them in other contexts. Once upon a time you had to be involved in classified stuff to get access to a GPS position sensor. Now you get GPS everywhere, in phones, cameras, anything with an IC. The people who design such IC don't need to know squat abou the GPS; they just need to be able to read the Interface Specification for the module. With time the robust, useful stuff will survive while all the hyped-up fads will fall through. Once people take a couple of steps back from the maths and wiring diagrams, and see the big picture, they will see what methods work almost everywhere, and which fall through. Once people obtain that awareness, there will be little incentive for checking new techniques. Using the old ones in new contexts - yes - but not developing new stuff. Rune
Is DSP a good field to have a career in?
Started by ●March 25, 2009
Reply by ●March 27, 20092009-03-27
Reply by ●March 27, 20092009-03-27
Rune Allnor wrote:> On 27 Mar, 15:00, Tim Wescott <t...@seemywebsite.com> wrote: >> On Wed, 25 Mar 2009 15:53:09 -0700, Rune Allnor wrote: >>> On 25 Mar, 22:51, "noodle22" <jw970...@yahoo.com> wrote: >>>> I know this is a subjective question but I've always found DSP >>>> interesting and have considered doing a career in it (it certainly >>>> beats business apps). >>> First of all, make up your mind about what you want: An 'interesting' >>> job or making a living. >>>> Is the field growing, shrinking, or not changing? Is there much demand >>>> and are there many jobs available? >>>> What type of work do DSP engineers do? >>> I'd say DSP is a transitionary discpline, like radio telegraphists used >>> to be at sea in the first half of the 20th century: The operators were >>> there and necessary when the field was new and worked with morse code. >>> Once voiced radio was possible, the specialized skills of the radio >>> operator were no longer necessary, as anyone could use a radio set. >>> DSP is headed in the same direction. >> (snip) >> >> I don't see it heading in that direction at all, > > It is. DSP is about designing - not building - the radio systems. > Sure, there will be a one or two decades with trial'n error, > but then the indsutry will gravitate towards one or two 'standard' > solutions. > > 100 years ago there was lots of debate about power supplies; > AC vs DC, voltages and frequencies. Now the two standard systems > are 60 Hz / 110 V and 50 Hz / 220 V. > > It doesn't matter if somebody comes up with a design and can > prove that it would work optimally with 109.78 Hz / 19.8 V. > Unless it can work with one of the two standard solutions, > no one will be interested. > > In 50 years there will be standard solutions for just about > anything; the main trends will emerge in the next couple > of decades. Once that has happened, there will be no more > need for DSP designs. In the mean time, the main effort > is about politics. > >> at least not if you can >> do embedded software or FPGA work. Canned DSP routines for FPGAs and >> DSPs are crap, > > Sure. But that doesn't stop DSP courses gravitating > into matlab signal processing toolbox training sessions... > >> unless you happen to need one for one of the big >> applications. As long as there are people out there making products that >> do new things, there will be demand for other people who can make those >> products work. > > I don't see that people do *new* thing; they keep doing > the same old stuff but economy and engineering let them do > them in other contexts. Once upon a time you had to be > involved in classified stuff to get access to a GPS > position sensor. Now you get GPS everywhere, in phones, > cameras, anything with an IC. The people who design > such IC don't need to know squat abou the GPS; they > just need to be able to read the Interface Specification > for the module. > > With time the robust, useful stuff will survive while > all the hyped-up fads will fall through. Once people > take a couple of steps back from the maths and wiring > diagrams, and see the big picture, they will see what > methods work almost everywhere, and which fall through. > > Once people obtain that awareness, there will be little > incentive for checking new techniques. Using the old > ones in new contexts - yes - but not developing new > stuff.Somewhere around 1900, The British Patent Office considered shutting down on the grounds that nothing new was left to be invented. Are you repeating that error? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●March 27, 20092009-03-27
Jerry Avins wrote:> Somewhere around 1900, The British Patent Office considered shutting > down on the grounds that nothing new was left to be invented. Are you > repeating that error?FWIW, I'm pretty sure this is an urban legend. http://answers.google.com/answers/threadview/id/361397.html
Reply by ●March 27, 20092009-03-27
Jeff Cunningham wrote:> Jerry Avins wrote: >> Somewhere around 1900, The British Patent Office considered shutting >> down on the grounds that nothing new was left to be invented. Are you >> repeating that error? > > FWIW, I'm pretty sure this is an urban legend.My faith in bureaucracy is somewhat restored!> http://answers.google.com/answers/threadview/id/361397.htmlJerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●March 27, 20092009-03-27
Jeff Cunningham wrote:> Jerry Avins wrote: >> Somewhere around 1900, The British Patent Office considered shutting >> down on the grounds that nothing new was left to be invented. Are you >> repeating that error? > > FWIW, I'm pretty sure this is an urban legend. > http://answers.google.com/answers/threadview/id/361397.htmlWhat is the verdict on :Everything that can be invented has been invented." -- Charles H. Duell, Commissioner, U.S. Office of Patents, 1899. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●March 27, 20092009-03-27
On 27 Mar, 20:46, Jerry Avins <j...@ieee.org> wrote:> Rune Allnor wrote: > > On 27 Mar, 15:00, Tim Wescott <t...@seemywebsite.com> wrote: > >> On Wed, 25 Mar 2009 15:53:09 -0700, Rune Allnor wrote: > >>> On 25 Mar, 22:51, "noodle22" <jw970...@yahoo.com> wrote: > >>>> I know this is a subjective question but I've always found DSP > >>>> interesting and have considered doing a career in it (it certainly > >>>> beats business apps). > >>> First of all, make up your mind about what you want: An 'interesting' > >>> job or making a living. > >>>> Is the field growing, shrinking, or not changing? �Is there much demand > >>>> and are there many jobs available? > >>>> What type of work do DSP engineers do? > >>> I'd say DSP is a transitionary discpline, like radio telegraphists used > >>> to be at sea in the first half of the 20th century: The operators were > >>> there and necessary when the field was new and worked with morse code. > >>> Once voiced radio was possible, the specialized skills of the radio > >>> operator were no longer necessary, as anyone could use a radio set. > >>> DSP is headed in the same direction. > >> (snip) > > >> I don't see it heading in that direction at all, > > > It is. DSP is about designing - not building - the radio systems. > > Sure, there will be a one or two decades with trial'n error, > > but then the indsutry will gravitate towards one or two 'standard' > > solutions. > > > 100 years ago there was lots of debate about power supplies; > > AC vs DC, voltages and frequencies. Now the two standard systems > > are 60 Hz / 110 V and 50 Hz / 220 V. > > > It doesn't matter if somebody comes up with a design and can > > prove that it would work optimally with 109.78 Hz / 19.8 V. > > Unless it can work with one of the two standard solutions, > > no one will be interested. > > > In 50 years there will be standard solutions for just about > > anything; the main trends will emerge in the next couple > > of decades. Once that has happened, there will be no more > > need for DSP designs. In the mean time, the main effort > > is about politics. > > >> at least not if you can > >> do embedded software or FPGA work. �Canned DSP routines for FPGAs and > >> DSPs are crap, > > > Sure. But that doesn't stop DSP courses gravitating > > into matlab signal processing toolbox training sessions... > > >> unless you happen to need one for one of the big > >> applications. �As long as there are people out there making products that > >> do new things, there will be demand for other people who can make those > >> products work. > > > I don't see that people do *new* thing; they keep doing > > the same old stuff but economy and engineering let them do > > them in other contexts. Once upon a time you had to be > > involved in classified stuff to get access to a GPS > > position sensor. Now you get GPS everywhere, in phones, > > cameras, anything with an IC. The people who design > > such IC don't need to know squat abou the GPS; they > > just need to be able to read the Interface Specification > > for the module. > > > With time the robust, useful stuff will survive while > > all the hyped-up fads will fall through. Once people > > take a couple of steps back from the maths and wiring > > diagrams, and see the big picture, they will see what > > methods work almost everywhere, and which fall through. > > > Once people obtain that awareness, there will be little > > incentive for checking new techniques. Using the old > > ones in new contexts - yes - but not developing new > > stuff. > > Somewhere around 1900, The British Patent Office considered shutting > down on the grounds that nothing new was left to be invented. Are you > repeating that error?Nope. The wheel is probably one of the most important inventions in human history, and was invented by one or more persons whose intellects were comparable by Newton or Einstein. But we don't study wheels these days. We use them. It's the same with DSP. It's one tool among many. It takes a bit of training to master - like a power drill - but that's about it. Once you know the dos and don'ts, there is little more to it. Rune
Reply by ●March 27, 20092009-03-27
On Mar 27, 4:51�pm, Jerry Avins <j...@ieee.org> wrote:> Jeff Cunningham wrote: > > Jerry Avins wrote: > >> Somewhere around 1900, The British Patent Office considered shutting > >> down on the grounds that nothing new was left to be invented. Are you > >> repeating that error? > > > FWIW, I'm pretty sure this is an urban legend. > >http://answers.google.com/answers/threadview/id/361397.html > > What is the verdict on :Everything that can be invented has been > invented." -- Charles H. Duell, Commissioner, U.S. Office of Patents, > 1899. > > Jerry > -- > Engineering is the art of making what you want from things you can get. > �����������������������������������������������������������������������All that this story verifies is that political hacks were appointed to high position back then just as they are today
Reply by ●March 27, 20092009-03-27
Thanks everyone for your responses. I have learned quite a bit about the
field from you and will need to think on it some more
@Dave You gave some encouraging advice. Another degree probably would not
be a bad idea and I think I would be up for doing it provided I found a
field that I truly liked. The Mentor comment was also good. I have a
mentor now who provides quite a bit of insight to me for project management
(beyond the PMP stuff). It is really useful and I hope I will always be
lucky enough to have a good mentor when needed. Defense makes sense. Not
sure how keen I would be to work for defense but definitely worth
considering.
@Rune You have a much darker view of this type of work and I will consider
you comments and research a bit more. I don't want to be in a dying field.
I had assumed that I would make money comparable to any other
engineer if I was in a field that had a lot of dsp work but you feel
otherwise so that is going to be a key consideration. It is true that a
lot of DSP functionaltiy now comes in software packages that are fairly
usable as is for many applications.
@Tim I have a friend who specialized in FPGAs. I had never really
considered them and their relation to dsp but will ask him more about it.
thanks.
In general, the whole concept of doing a job and then having dsp as part
of the job, rather than doing a job focused on dsp has been clarified for
me. I think the first alternative sounds better anyway. Probably more fun
to do dsp just some of the time.
Thanks
Reply by ●March 27, 20092009-03-27
>I don't want to be in a dying field.I heard that currently finance industry is suffering a lot, that does not mean MBA in Finance is bad idea or becoming CA is bad idea.> I had assumed that I would make money comparable to anyother>engineer if I was in a field that had a lot of dsp workAs Rune said earlier decide what u want 'interesting job' or 'good living'. What I belive is if u r good in what u do, money will follow. Money is always a byproduct. Chintan
Reply by ●March 27, 20092009-03-27
On Mar 28, 4:42�am, Rune Allnor <all...@tele.ntnu.no> wrote:> On 27 Mar, 15:00, Tim Wescott <t...@seemywebsite.com> wrote: > > > > > On Wed, 25 Mar 2009 15:53:09 -0700, Rune Allnor wrote: > > > On 25 Mar, 22:51, "noodle22" <jw970...@yahoo.com> wrote: > > >> I know this is a subjective question but I've always found DSP > > >> interesting and have considered doing a career in it (it certainly > > >> beats business apps). > > > > First of all, make up your mind about what you want: An 'interesting' > > > job or making a living. > > > >> Is the field growing, shrinking, or not changing? �Is there much demand > > >> and are there many jobs available? > > > >> What type of work do DSP engineers do? > > > > I'd say DSP is a transitionary discpline, like radio telegraphists used > > > to be at sea in the first half of the 20th century: The operators were > > > there and necessary when the field was new and worked with morse code. > > > Once voiced radio was possible, the specialized skills of the radio > > > operator were no longer necessary, as anyone could use a radio set. > > > > DSP is headed in the same direction. > > > (snip) > > > I don't see it heading in that direction at all, > > It is. DSP is about designing - not building - the radio systems. > Sure, there will be a one or two decades with trial'n error, > but then the indsutry will gravitate towards one or two 'standard' > solutions. > > 100 years ago there was lots of debate about power supplies; > AC vs DC, voltages and frequencies. Now the two standard systems > are 60 Hz / 110 V and 50 Hz / 220 V. >Funny enough I read that the Chinese are going back to dc transmission of power for some applications. Hardy
