About the continuous Fourier Transform

[dah145]

Is there any reason that even though it is implemented officially in Xcas (http://www-fourier.ujf-grenoble.fr/~pari...ourier-def) it isn't in the Prime? It is odd.

And is there any way to compute the transform through its integral definition in the Prime?

Thanks.

[victorvbc]

Unfortunately there is a licensing problem with the two-sided fourier transform program that makes it difficult to add to the Prime's CAS. There was some communication with the dev to sort that out, but I don't know if it went anywhere.

You can get the one-sided version though, by using the Laplace transform function and sub (s=i*w).

[dah145]

(12-04-2020 03:19 PM)victorvbc Wrote:  Unfortunately there is a licensing problem with the two-sided fourier transform program that makes it difficult to add to the Prime's CAS. There was some communication with the dev to sort that out, but I don't know if it went anywhere.

You can get the one-sided version though, by using the Laplace transform function and sub (s=i*w).

Well now that's just downright bizarre. Fellow Electrical Engineers will agree with me that this function is commonly used amongst various applications in the field. And obtaining the one sided version through the Laplace transform simply does not render the same results that are correctly and conveniently obtained through the native fourier() xCas function.

laplace(atan(1/(2*x^2),x,s) for example simply hangs on my Prime and delivers no result.

This automatically makes the Casio Classpad superior to the Prime in this particular application as their CAS supports the Continuous Fourier Transform natively. And I don't know if Khicas supports this particular function, but that would be ironic on various levels.

This is really a shame, considering HP already moved in the right direction with the Prime for EE interests implementing the Z transform natively. Wonder what are the 'licensing problems' for implementing the CFT, and if there's any chance it will be included in a future update by simply making the already existing Xcas/giac function available in the Prime.

[victorvbc]

This is the thread I was referring to. It would be nice to get some feedback on how things are going, but I don't know whether there will be another firmware update soon.

[dah145]

(12-04-2020 04:45 PM)victorvbc Wrote:  This is the thread I was referring to. It would be nice to get some feedback on how things are going, but I don't know whether there will be another firmware update soon.

Thanks for the link, it is quite eye opening to how these things are handled from the HP side. It really looks like this calculator could be so much more but it is limited by some baffling decisions by people up there, not the devs for sure.

[rprosperi]

(12-04-2020 06:07 PM)dah145 Wrote:  

(12-04-2020 04:45 PM)victorvbc Wrote:  This is the thread I was referring to. It would be nice to get some feedback on how things are going, but I don't know whether there will be another firmware update soon.

Thanks for the link, it is quite eye opening to how these things are handled from the HP side. It really looks like this calculator could be so much more but it is limited by some baffling decisions by people up there, not the devs for sure.

I don't think it's too baffling. The original code Bernard used in Giac is likely open source, which requires derivative products that contain that code to remain open source as well. HP clearly can't do that with the Prime, but hat's off to HP and Bernard for doing the right thing and not including it, thus respecting and preserving license integrity.

[dah145]

(12-04-2020 07:54 PM)rprosperi Wrote:  

(12-04-2020 06:07 PM)dah145 Wrote:  Thanks for the link, it is quite eye opening to how these things are handled from the HP side. It really looks like this calculator could be so much more but it is limited by some baffling decisions by people up there, not the devs for sure.

I don't think it's too baffling. The original code Bernard used in Giac is likely open source, which requires derivative products that contain that code to remain open source as well. HP clearly can't do that with the Prime, but hat's off to HP and Bernard for doing the right thing and not including it, thus respecting and preserving license integrity.

Yes, I now understand why it isn't included and I agree it was the right thing to do.

I meant to say baffling in relation to the resilience of HP against opening up or of at least give the option to open up the Prime to some degree, including the ability to use open source extensions, such as this case, developed by the community for giac/Xcas.

[Aries]

(12-04-2020 07:08 AM)dah145 Wrote:  Is there any reason that even though it is implemented officially in Xcas (http://www-fourier.ujf-grenoble.fr/~pari...ourier-def) it isn't in the Prime? It is odd.

And is there any way to compute the transform through its integral definition in the Prime?

Thanks.

Hi dah145,
in a short while I'll show you how I did with the TI-Nspire CAS, maybe it could be helpful to you.
See you later, then.
Best,

Aries ;-)

[medwatt]

I am the one that created the thread that is referenced a couple of posts above this one. The first time I created the thread, Tom (who apparently is part of the hp team) responded and said it's going to be almost impossible to add the continuous FT function because of copyright reasons. After I contacted the author, who responded very positively, I contacted Tom again several times via PM, but he didn't respond to any of them, even though he used to be responsive to other questions.

I am an electrical engineering student, with a focus on mixed signal design (read microelectronics). You'd be surprised, but I rarely needed a calculator after my circuit analysis class during my undergraduate days. In fact, I'd argue that the Casio 115MS is the best calculator for circuit analysis since it allows you to type and convert complex numbers with the least amount of key strokes. Most of the time, I am deriving equations for which I need the closed form expression of some integral (such as the FT), solution, etc. Also, I hardly need to plot anything trivial. It's nice to see all those beautiful plots that the prime can produce, but are you going to really be plotting parametric equations every day ?

So, why then did I buy the Prime. I was hoping that since it seems so capable, I would be able to do some nice stuff on it and that it might provide python support down the line as I wouldn't call the Prime Language a programming language. I also saw there's a control app for the Prime (written by a 3rd party) that was really enticing. I played with it a bit but realized it was too buggy. Anyway ... my prime has been gathering dust in my drawer for almost a year now.

[dah145]

(12-09-2020 12:42 PM)Aries Wrote:  

(12-04-2020 07:08 AM)dah145 Wrote:  Is there any reason that even though it is implemented officially in Xcas (http://www-fourier.ujf-grenoble.fr/~pari...ourier-def) it isn't in the Prime? It is odd.

And is there any way to compute the transform through its integral definition in the Prime?

Thanks.

Hi dah145,
in a short while I'll show how I did with the TI-Nspire CAS, maybe it could be helpful to you.
See you later, then.
Best,

Aries ;-)

That'll be nice yeah. Thank you.

[dah145]

(12-09-2020 10:17 PM)medwatt Wrote:  I am the one that created the thread that is referenced a couple of posts above this one. The first time I created the thread, Tom (who apparently is part of the hp team) responded and said it's going to be almost impossible to add the continuous FT function because of copyright reasons. After I contacted the author, who responded very positively, I contacted Tom again several times via PM, but he didn't respond to any of them, even though he used to be responsive to other questions.

I am an electrical engineering student, with a focus on mixed signal design (read microelectronics). You'd be surprised, but I rarely needed a calculator after my circuit analysis class during my undergraduate days. In fact, I'd argue that the Casio 115MS is the best calculator for circuit analysis since it allows you to type and convert complex numbers with the least amount of key strokes. Most of the time, I am deriving equations for which I need the closed form expression of some integral (such as the FT), solution, etc. Also, I hardly need to plot anything trivial. It's nice to see all those beautiful plots that the prime can produce, but are you going to really be plotting parametric equations every day ?

So, why then did I buy the Prime. I was hoping that since it seems so capable, I would be able to do some nice stuff on it and that it might provide python support down the line as I wouldn't call the Prime Language a programming language. I also saw there's a control app for the Prime (written by a 3rd party) that was really enticing. I played with it a bit but realized it was too buggy. Anyway ... my prime has been gathering dust in my drawer for almost a year now.

I understand your sentiment regarding the Prime, but I still believe the potential for this calculator is there.

As of today the Prime can be quite useful for the EE student, and specially in comparison to the other CAS calculators, I would say it the most capable, but that's putting the bar too low, there's a lot of room for improvement. Getting a Bode plot (a simple log scale graph) shouldn't be a programming challenge for example.

It just seems there's no real focus coming from HP for whom exactly this calculator is meant to, and it the end its software feature set is generally lack lustre for the various segments of education it tries to tackle on.

Opening up the development can certainly be a way to fill the voids.

[victorvbc]

(12-10-2020 07:19 AM)dah145 Wrote:  I understand your sentiment regarding the Prime, but I still believe the potential for this calculator is there.

As of today the Prime can be quite useful for the EE student, and specially in comparison to the other CAS calculators, I would say it the most capable, but that's putting the bar too low, there's a lot of room for improvement. Getting a Bode plot (a simple log scale graph) shouldn't be a programming challenge for example.

It just seems there's no real focus coming from HP for whom exactly this calculator is meant to, and it the end its software feature set is generally lack lustre for the various segments of education it tries to tackle on.

Opening up the development can certainly be a way to fill the voids.

Tbh, I love this calculator. Definitely takes some effort for it to do specific things (i.e., the bode plot), but as a classroom calculator it's awesome, especially if you need to deal with a lot of symbolic or numerical matrices.

That being said, the stability problems and lack of arbitrary precision in some areas would restrain me from using it as a professional engineering device in my field. But for that MATLAB/Octave, Python libraries, etc are the norm anyway.

[Anders]

(12-09-2020 10:17 PM)medwatt Wrote:  I am the one that created the thread that is referenced a couple of posts above this one. The first time I created the thread, Tom (who apparently is part of the hp team) responded and said it's going to be almost impossible to add the continuous FT function because of copyright reasons. After I contacted the author, who responded very positively, I contacted Tom again several times via PM, but he didn't respond to any of them, even though he used to be responsive to other questions.

I am an electrical engineering student, with a focus on mixed signal design (read microelectronics). You'd be surprised, but I rarely needed a calculator after my circuit analysis class during my undergraduate days. In fact, I'd argue that the Casio 115MS is the best calculator for circuit analysis since it allows you to type and convert complex numbers with the least amount of key strokes. Most of the time, I am deriving equations for which I need the closed form expression of some integral (such as the FT), solution, etc. Also, I hardly need to plot anything trivial. It's nice to see all those beautiful plots that the prime can produce, but are you going to really be plotting parametric equations every day ?

So, why then did I buy the Prime. I was hoping that since it seems so capable, I would be able to do some nice stuff on it and that it might provide python support down the line as I wouldn't call the Prime Language a programming language. I also saw there's a control app for the Prime (written by a 3rd party) that was really enticing. I played with it a bit but realized it was too buggy. Anyway ... my prime has been gathering dust in my drawer for almost a year now.

First off thank you for writing the EE s/w, it's been useful.
Secondly, I agree that the biggest issue that is preventing Prime to be truly useful day to day for EE and CE students and pros is how much work it is to effectively use complex math moving between polar to a rectangular effectively, seamlessly and consistently in CAS and in Home view. It's just painful in Prime. Some has done some programs to compensate but its just not a good enough work around.
This alone makes Prime not the right calculator for EE.

And it is super frustrating because it has the potential, e.g. prime has the best implementation of Z transforms on any out of the box calculator etc.
Falling flat on the basic complex math is just sad. And it's not because we did not raise this point over the years - We did! Multiple times but to no avail.
I've pretty much given up.

[medwatt]

(12-10-2020 02:54 PM)victorvbc Wrote:  

(12-10-2020 07:19 AM)dah145 Wrote:  I understand your sentiment regarding the Prime, but I still believe the potential for this calculator is there.

As of today the Prime can be quite useful for the EE student, and specially in comparison to the other CAS calculators, I would say it the most capable, but that's putting the bar too low, there's a lot of room for improvement. Getting a Bode plot (a simple log scale graph) shouldn't be a programming challenge for example.

It just seems there's no real focus coming from HP for whom exactly this calculator is meant to, and it the end its software feature set is generally lack lustre for the various segments of education it tries to tackle on.

Opening up the development can certainly be a way to fill the voids.

Tbh, I love this calculator. Definitely takes some effort for it to do specific things (i.e., the bode plot), but as a classroom calculator it's awesome, especially if you need to deal with a lot of symbolic or numerical matrices.

That being said, the stability problems and lack of arbitrary precision in some areas would restrain me from using it as a professional engineering device in my field. But for that MATLAB/Octave, Python libraries, etc are the norm anyway.

What exactly does "classroom calculator" mean ? Does it mean you actually sit in a classroom while the teacher/lecturer is teaching and you use the calculator ? Or did you mean a portable calculator that you can use in a place like a library when you don't want to carry your laptop ? If you meant the former, then I just cannot imagine someone having time to focus on the lecture and then busy himself playing with their calculator. How does anyone have the time ? In fact, since I left school, I don't even take notes in class anymore. I just sit down, listen, jot down a few keywords, and that's it.

[victorvbc]

(12-11-2020 12:42 AM)medwatt Wrote:  What exactly does "classroom calculator" mean ? Does it mean you actually sit in a classroom while the teacher/lecturer is teaching and you use the calculator ? Or did you mean a portable calculator that you can use in a place like a library when you don't want to carry your laptop ? If you meant the former, then I just cannot imagine someone having time to focus on the lecture and then busy himself playing with their calculator. How does anyone have the time ? In fact, since I left school, I don't even take notes in class anymore. I just sit down, listen, jot down a few keywords, and that's it.

Actually both. Here in Brasil we have a different approach to teaching engineering at university. The courses are five years full time, and the classes are less like lectures. You are expected to solve problems during class in many cases. In the last two years you even need a graphing calculator to do the tests, otherwise good luck finding eigenvalues or inverting a high order system.

[Valentin Albillo]

(12-11-2020 03:37 PM)victorvbc Wrote:  In the last two years you even need a graphing calculator to do the tests, otherwise good luck finding eigenvalues or inverting a high order system.

You don't need a graphing calculator to do any of those computations, a 36-yo non-graphing HP-71B can tackle them alright. So does an HP-42S or an even older HP-41C for that matter.

V.

[victorvbc]

(12-11-2020 06:03 PM)Valentin Albillo Wrote:  You don't need a graphing calculator to do any of those computations, a 36-yo non-graphing HP-71B can tackle them alright. So does an HP-42S or an even older HP-41C for that matter.

V.

You are correct, I am sorry. It's just that most "easily available" calculators with those capabilities also happen to be in the graphing category. A lot of the cheaper scientific calculators (i.e. fx-991 series) can only do basic matrix operations up to 4x4 or less unfortunately.

[Aries]

(12-10-2020 06:51 AM)dah145 Wrote:  

(12-09-2020 12:42 PM)Aries Wrote:  Hi dah145,
in a short while I'll show how I did with the TI-Nspire CAS, maybe it could be helpful to you.
See you later, then.
Best,

Aries ;-)

That'll be nice yeah. Thank you.

Hey dah145,
I'm back, here is a simple app:





First off you define the piecewise function (in an interval equal to a period) like this (say for rect(t) function):



Then you "describe" the function as a matrix (and store it to ff, as requested by the app), like this:



Lastly, you can apply fourier(ff):



The graph:



For every row in the matrix (cycle "for"), the program runs the integration and then the results (stored in "s") are counted together.
I hope this can be a valid "starting point", eventually you could do so much more.
Happy XMas to you and everyone here,

Aries

[toshk]

Also as simple as Laplace which is Most Elegant/Powerful of these Transformations;

laplace(1*Heaviside(x-(-t/2))-1*Heaviside(x-(t/2)),x,s)

OR

laplace(0*Heaviside(x+∞)-0*Heaviside(x-(t/2))+1*Heaviside(x-(-t/2))-1*Heaviside(x-(t/2))+0*Heaviside(x-(t/2))-0*Heaviside(x-∞),x,s)

[dah145]

(12-13-2020 08:33 AM)Aries Wrote:  

(12-10-2020 06:51 AM)dah145 Wrote:  That'll be nice yeah. Thank you.

Hey dah145,
I'm back, here is a simple app:





First off you define the piecewise function (in an interval equal to a period) like this (say for rect(t) function):



Then you "describe" the function as a matrix (and store it to ff, as requested by the app), like this:



Lastly, you can apply fourier(ff):



The graph:



For every row in the matrix (cycle "for"), the program runs the integration and then the results (stored in "s") are counted together.
I hope this can be a valid "starting point", eventually you could do so much more.
Happy XMas to you and everyone here,

Aries

Thank you for your input, will definitely look into it.