🚀 go-pugleaf

Sounds great. I would also like to encourage discussion on the topic of
using DSP techniques to improve the quality of measurements/inputs made with
exisiting adc hardware such as dithering and averaging that improve the
available dynamic range to much better than the bit level of the adc. After
all, any real analog input to a DSP system must pass through an adc.

Roger-

I noticed that several people have asked about DSP textbooks.
"Discrete-time Signal Processing" by Oppenheim and Schafer is a good
one.
--
+---------------------------------------------------------------------------+
| Email:        pshen@atrp.mit.edu                  |       Paul Shen       |
| Address:      MIT, 36-665, Cambridge, Ma 02139    |  Tel. (617) 864-3210  | 
+---------------------------------------------------------------------------+

"Discrete-time Signal Processing" by Oppenheim and Schafer 
is about the best I know of as well (I work with Dr. Scahfer here at Ga. Tech
and I think I may try to convince him to read this news group)

P. Allen Jensen
Georgia Tech, School of Electrical Engineering, Atlanta, GA  30332
USENET: ...!{allegra,hplabs,ihnp4,ulysses}!gatech!eedsp!jensen
INTERNET: jensen@eedsp.gatech.edu

In article <459@eedsp.gatech.edu> jensen@bessel.eedsp.gatech.edu (P. Allen Jensen) writes:
>"Discrete-time Signal Processing" by Oppenheim and Schafer 
>is about the best I know of as well (I work with Dr. Scahfer here at Ga. Tech
>and I think I may try to convince him to read this news group)
>
>P. Allen Jensen

Please try to get him to interract with the net on this subject.  I have a copy
of another of their books "Digital Signal Processing" that I have used for 
many years; it's my favorite.  Dr. Schafer is certainly one of the most 
knowledgeable and well-known experts on DSP in the world.

This could become a real learning experience group.

In a slightly different vein... does anyone have an idea for a way to share
equations with a (hopefully large) group of net users?  One particular 
platform's graphics or typesetting would likely be out.  I'm thinking of
postscript, but I have no idea of the spread of users that could use it.

This concern has arisen before; DSP needs sigmas, limits, etc to even begin
to say anything.  Ideas here?  Dr. Schafer?

Jim ... jisom@sun.com

Time flies like an arrow.
Fruit flies like a banana.

In article <668@suntops.Tops.Sun.COM> jisom@santa_fe.tops.sun.com (Jim Isom)
writes: 
>In a slightly different vein... does anyone have an idea for a way to share
>equations with a (hopefully large) group of net users?  One particular 
>platform's graphics or typesetting would likely be out.  I'm thinking of
>postscript, but I have no idea of the spread of users that could use it.

Is LaTeX sufficiently common to use as a medium?  It has just about everything
you might want, and it is high-level enough that you can still read the source
(well, most of the time, anyway :-)

\sum_{i=0}^p a_i y_{n-i} = \sum_{l=0}^q b_i \xi_{n-i}

SR

In article <668@suntops.Tops.Sun.COM> jisom@santa_fe.tops.sun.com (Jim Isom) writes:
>
>(much stuff deleted...)
>In a slightly different vein... does anyone have an idea for a way to share
>equations with a (hopefully large) group of net users?  One particular 
>platform's graphics or typesetting would likely be out.  I'm thinking of
>postscript, but I have no idea of the spread of users that could use it.
>
>This concern has arisen before; DSP needs sigmas, limits, etc to even begin
>to say anything.  Ideas here?  Dr. Schafer?

I'm not Dr. Schafer, but I can certainly toss in my two cents' worth...

I, for one, would not be able to make much use of PostScript equations.
Though I have access to PostScript printers, it would be awkward to have
to print out selected bits of an article before I could follow the discussion.
Also, the size of an article with PostScript equations concerns me.

I might feel differently if I had a NeXT, or something else that spoke
PostScript and displayed it on the screen.

May I suggest TeX? TeX math mode, while still awkward, is somewhat more
transparent than PostScript, and just about everybody should be able to
use it in one way or another.

Comments?

...laura

In article <668@suntops.Tops.Sun.COM> jisom@santa_fe.tops.sun.com (Jim Isom) writes:
>
>This could become a real learning experience group.
>
 
One would hope...

>
>In a slightly different vein... does anyone have an idea for a way to share
>equations with a (hopefully large) group of net users?  One particular 
>platform's graphics or typesetting would likely be out.  I'm thinking of
>postscript, but I have no idea of the spread of users that could use it.
>

Personally, I would advocate TeX/LaTeX for communicating equations.
It has the advantage of running on a vast number of horribly baroque
machines, having a wide array of output drivers and, most importantly,
it is relatively readable in unprocessed form, unlike something like
postscript or eqn.

Oh, yea...and it's free...;')

	...ken seefried iii
	   ken@gatech.edu

Programming, v., an activity not unlike beating ones head, repeatedly,
against a wall, but with less opportunity for reward...

In article <19237@gatech.edu> ken@gatech.UUCP (Ken Seefried III) writes:
>Personally, I would advocate TeX/LaTeX for communicating equations.
>It has the advantage of running on a vast number of horribly baroque
>machines, having a wide array of output drivers and, most importantly,
>it is relatively readable in unprocessed form, unlike something like
>postscript or eqn.

Uh, agreed about postscript, but somebody has misinformed you about eqn.
TeX's math notation is a LESS-readable clone of eqn (more or less).
See related posting.
-- 
"Where is D.D. Harriman now,   |     Henry Spencer at U of Toronto Zoology
when we really *need* him?"    | uunet!attcan!utzoo!henry henry@zoo.toronto.edu

In article <3085@quanta.eng.ohio-state.edu> rob@baloo.eng.ohio-state.edu (Rob Carriere) writes:
>Is LaTeX sufficiently common to use as a medium?  It has just about everything
>you might want, and it is high-level enough that you can still read the source
>(well, most of the time, anyway :-)
>
>\sum_{i=0}^p a_i y_{n-i} = \sum_{l=0}^q b_i \xi_{n-i}

I think it's going to come down to a war between TeX and eqn.  Each is more
common in some environments.  Expressive power is similar; the eqn version
of the above is (assuming I've read the TeX correctly):

sum from i=0 to p a sub i y sub {n-i} = sum from l=0 to q b sub i xi sub {n-i}

Sticking strictly to readability, and disregarding the who's-got-which-
software issue, I think eqn comes out ahead -- fewer silly backslashes and
fewer artifacts of the odd character set Knuth used.
-- 
"Where is D.D. Harriman now,   |     Henry Spencer at U of Toronto Zoology
when we really *need* him?"    | uunet!attcan!utzoo!henry henry@zoo.toronto.edu

In article <3085@quanta.eng.ohio-state.edu> rob@baloo.eng.ohio-state.edu (Rob Carriere) writes:
>In article <668@suntops.Tops.Sun.COM> jisom@santa_fe.tops.sun.com (Jim Isom)
>writes: 
>>platform's graphics or typesetting would likely be out.  I'm thinking of
>>postscript, but I have no idea of the spread of users that could use it.

>Is LaTeX sufficiently common to use as a medium?  It has just about everything

I think it's not. Encapsulated PostScript, as can be fed to any decent
layout program, is, though. I think :')

h+@nada.kth.se

-- 
Anarchy is better that no government at all.

In article <19237@gatech.edu> ken@gatech.UUCP (Ken Seefried III) writes:
>>In a slightly different vein... does anyone have an idea for a way to share
>>equations with a (hopefully large) group of net users?  One particular 
>>platform's graphics or typesetting would likely be out.  I'm thinking of
>>postscript, but I have no idea of the spread of users that could use it.
>
>Personally, I would advocate TeX/LaTeX for communicating equations.

I think Mathematica notebooks are a better medium for distributing signal
processing ideas.  While it is true that Mathematica's notation is a bit 
baroque the advantage of having runnable code with graphics and text is
pretty important.  A well written notebook can be read like a normal paper
but since all the equations are "live" readers can modify them and use the
equations in their own investigations.  Perhaps, someday the notation might
be prettier but I think live, interactive documents are the way to go.

It is important to note, also, that Mathematica notebooks are straight ASCII
text.  This makes them very digestible by bulletin board software.  In addition
a free Mathematica notebook reader is available for the Mac (so you can see
the graphics) (and I guess the NeXT too :-).  On other machines (PCs, Suns, 
etc) the notebooks can be read and used for new work but all the benefits of
a notebook aren't available, yet.

An example of this type of document is a report titled "Lyon's Cochlear Model,"
which I wrote to describe some of the research we are doing at Apple.  I used
Mathematica to write and distribute this report because I wanted readers to
be able to explore the design as I had when writing the report.  The appendix 
to the report includes a large number of functions for designing and plotting 
continuous and discrete second order sections.  This report is available
for free from me (malcolm@apple.com) or from the Apple Corporate Library if
you send a paper mail address.  (The notebook has so many graphics that it is 
1.5Mbytes long so it is a bit large for mass distribution over the networks.  
The technical report has been published as a typeset paper with an attached 
Macintosh floppy that contains the electronic version of the notebook.)

I describe other benefits of Mathematica notebooks for signal processing in
an upcoming article in the January 1990 issue of IEEE Signal Processing 
Magazine (the tutorial and general interest magazine).  I see the main 
benefits of a Mathematica-style notebook for research and publishing as:
	1)	An interactive research and learning environment.
	2)	Hypermedia for learning tailored to the reader's needs.
	3)	A literate programming environment.

TeX and even Troff are wonderful pieces of software but they only solve part 
of the problem.  Computers give us the ability to interact with our 
environment so why should we be stuck with static papers?

						Malcolm Slaney
						Speech and Hearing Project
						Apple Computer
						malcolm@apple.com

To all DSP  new comers! "Introduction to DSP" By: J. Proakis and D Manolakis is one 
of the best NEW reference books on the subject. Chapter 11 (Last one) has Spectral
Analysis (Power Spectrum Estimation). It is written that all newcomers can follow
at thier own paste, also  examples with FORTRAN subroutines are provided.


************ Those who can, do. Those who can't, simulate *************
    .    .          
   /\\  //\  Rafael E. Diaz  UUCP: { ..!cs.utexas.edu!oakhill!irit!diaz}
  /  \\//  \              	   {    oakhill!irit!diaz@cs.utexas.edu} 
 /    `'    \OTOROLA  Inc. DSP Operations, Austin TX     (512) 891  2977

In article <34990@apple.Apple.COM> malcolm@Apple.COM (Malcolm Slaney) writes:
>
>I think Mathematica notebooks are a better medium for distributing signal
>processing ideas.  While it is true that Mathematica's notation is a bit 
>baroque the advantage of having runnable code with graphics and text is
>pretty important.  
>

Yes...but the disadvantage of having to pay a not-so-small fortune to
get a copy of Mathematics is pretty important, too...

Not to mention the fact that Mathematica is availible on a relatively
small subset of computers (roughly 25 makes, no?).

Then, of course, is the fact that most readers probably have neither
Mathematica nor reader software.

>
>It is important to note, also, that Mathematica notebooks are straight ASCII
>text.  This makes them very digestible by bulletin board software.  In addition
>a free Mathematica notebook reader is available for the Mac (so you can see
>the graphics) (and I guess the NeXT too :-).  On other machines (PCs, Suns, 
>etc) the notebooks can be read and used for new work but all the benefits of
>a notebook aren't available, yet.
>

Ummm...last time I checked, TeX was straight ascii also.  Plus, I can run
TeX and get output from IBM VM/CMS, CDC Cyber NOS/VE, DEC TOPS-10/20, Amiga,
Atari-ST...on and on.

>
>TeX and even Troff are wonderful pieces of software but they only solve part 
>of the problem.  Computers give us the ability to interact with our 
>environment so why should we be stuck with static papers?
>

Sure they only solve some of the problem, the part of the problem that we
are discussing: how to communicate effectively on charater terminals
mathematical equations.

Now if all of us...most of us...well, even a few of us had
Mathematica, we'd be all set.  However, the vast majority don't (and
I'm not whining cause I don't have it...we have NeXT machines all
over campus).  Without Mathematica the notebooks aren't that useful
to the majority of readers.

With TeX, however, a vastly greater percentage of comp.dsp's readership
will be able read the equation on the screen, and print off a reasonably
good copy of it.  Sure, we won't be able to run it, or generate graphics
from it, etc., but it will still be useful to us.

Now mind you...I'm not trying to rain on your parade.  Mathematica is
wonderful stuff.  However, in TeX we have a common denominator that
will benefit the maximum number of users...

	...ken 

|>Is LaTeX sufficiently common to use as a medium?  It has just about everything
|>you might want, and it is high-level enough that you can still read the source
|>(well, most of the time, anyway :-)
|>
|>\sum_{i=0}^p a_i y_{n-i} = \sum_{l=0}^q b_i \xi_{n-i}
|
|I think it's going to come down to a war between TeX and eqn.  Each is more
|common in some environments.  Expressive power is similar; the eqn version
|of the above is (assuming I've read the TeX correctly):
|
|sum from i=0 to p a sub i y sub {n-i} = sum from l=0 to q b sub i xi sub {n-i}
								   ^^
Probably should be \(*x (Greek Xi).

|Sticking strictly to readability, and disregarding the who's-got-which-
|software issue, I think eqn comes out ahead -- fewer silly backslashes and
|fewer artifacts of the odd character set Knuth used.

If one doesn't like _ and ^ one can always do this

\def\sub{_}
\def\sup{^}
$\sum\sub{i=0}\sup p a\sub i y\sub{n-i} = \sum\sub{l=0}\sup q b\sub i \xi\sub{n-i}$

which might look better with full braces

$\sum\sub{i=0}\sup{p} a\sub{i} y\sub{n-i} = \sum\sub{l=0}\sup{q} b\sub{i} \xi\sub{n-i}$

Looks even better than eqn to me since you can tell which scripts
belongs to which symbols. What's the big deal?

In article <1989Sep24.032613.11841@utzoo.uucp> henry@utzoo.uucp (Henry Spencer) writes:

)I think it's going to come down to a war between TeX and eqn.  Each is more
)common in some environments.  Expressive power is similar; the eqn version
)of the above is (assuming I've read the TeX correctly):

I think it's going to take a little closer look at "some environments."

I think it's probably far and away easier to find TeX to run on non-unix
machines, including PC, Mac, Amiga, ST, and so on.  Also, I'd expect
it to be available on most unix machines which:
    1) are associated with technical departments
    2) hooked up to a laser printer or equiv.

I suggest we determine which is more common in the *subscribers* 
environments, and that this means more than whether there are backslashes
in there.

I expect that people might frequently print the equations out, esp.
from the more useful postings.

	jimm

-- 
Jim Mackraz, I and I Computing	   	"... the signs are very ominous,
{cbmvax,well,oliveb}!amiga!jimm          and a chill wind blows."
							- Justice Blackmun
Opinions are my own.  Comments are not to be taken as Commodore official policy.

In article <34990@apple.Apple.COM> malcolm@Apple.COM (Malcolm Slaney) writes:
>In article <19237@gatech.edu> ken@gatech.UUCP (Ken Seefried III) writes:
>> does anyone have an idea for a way to share equations [...]
>>Personally, I would advocate TeX/LaTeX for communicating equations.
>
>I think Mathematica notebooks are a better medium for distributing signal
>processing ideas.

Unfortunately, while both TeX/LaTeX and Mathematica would be nice,
this won't work for Usenet articles. Only a small percentage of the
readers of your articles have either of these packages. Also, even if they
did, it would be very inconvenient to shift articles from a normal
news reader (like 'rn' or 'vnews') into these packages for proper display.
Furthermore, of the people who have access to such packages, they don't
always have access the instant they're reading them. For instance, right
now I'm dialed up to work from home. And lastly, this subject has been
debated frequently in groups like sci.math and sci.electronics without
ever reaching a consensus; there's no particular reason to think that
things will be different in sci.dsp.

Some people just go ahead and use LaTeX in their articles anyway, which 
I think is a real shame, because it's much harder to read than plaintext
ASCII formulae.

>An example of this type of document is a report titled "Lyon's Cochlear Model,"
>which I wrote to describe some of the research we are doing at Apple.  I used

A very interesting paper it is, too, and thanks again.

I think it's great that you're championing use of Mathematica. It's just
premature to ask for use of it as a standard distribution format.
	Doug
-- 
Doug Merritt		{pyramid,apple}!xdos!doug
Member, Crusaders for a Better Tomorrow		Professional Wildeyed Visionary

In article <4593@amiga.UUCP> jimm@batgirl.UUCP (Jim Mackraz) writes:
>I suggest we determine which is more common in the *subscribers* 
>environments, and that this means more than whether there are backslashes
>in there.
>
>I expect that people might frequently print the equations out, esp.
>from the more useful postings.

But the *most* common use is still simply reading articles online. Even
the articles I save (several in this group so far), I'm not going to
print out. This is 1989...I work online! And despite working in a
relatively well equipped office with lots of modern workstations,
I don't have access to any software package which can put pretty printed
equations on my workstation displays.

(Also while it's gratifying to envision that people are going to immortalize
our postings in beautifully formatted laser printed postings to frame and
hang in their offices and admire endlessly. But that seductive vision is
mostly a mirage. It simply won't happen often enough to justify using
that as a basis for considering a standard equation format.)

There's only *one* common denominator, and that's plain text. Arguing
over, say, eqn vs. LaTeX in terms of readability makes sense. Arguing
over which is more commonly available does not. Neither one is "commonly"
available (as a percentage of readership).

If you really want to push this issue, your best bet is to find a way
to get an equation prettyprinter (for both display and printer) into
people's hands in a form they can use with news. It should initially
support plain ascii pretty printing (like Macsyma does for formulae),
plus optionally support for Suns, Apollos, NeXt (or more to the point,
X windows). Then work with the guys who distribute news software, and
get it incorporated into the next release of news. And hopefully into
the next release of 'less', also (my favorite display pager).

Eventually this would result in a reasonable percentage of the Unix
systems on the net having the ability to read news with equations
automatically pretty printed to the display. Note that even then, there'd
be gripes from all the Usenet folks on non-Unix systems (e.g. IBM
mainframes, DEC 10/20's, Mac's, Amiga's, PC's, etc), of which there are
a fairly large number. Porting the equation-reading news software to
some of these boxes would be a good idea, and it'd help, but there'll
always (well, for a long time) be those people who *still* don't have
it available.  So it'd *still* be necessary for the equations to be
*easily* readable without preprocessing.

This may sound pretty pessimistic. Don't get me wrong, I would very
much like to see this ability commonly available, and I would like to
see the above plan put into effect. I just think that there are severe
pragmatic limitations on what you can expect.

By the way, this is an issue very similar to that of internationalization.
It's currently a big deal with companies like e.g. Sun and AT&T to prepare
hardware and software systems that can deal effectively with non-English
languages (e.g. Japanese). Lots of work is being done in that area, and
sometime in the next few years you can expect to see a lot of systems
that can deal with Kanji, latin character diacriticals, etc. It would make
sense to integrate the issue of equation pretty printing; there are many
similar issues and many similar hw/sw requirements. Perhaps those of you
with an interest in display of equations could contact groups working on
internationalization and talk to them about meeting common goals.

Until then, it's easy to see there won't be a standard. Some people will
continue posting LaTex, others eqn, others plain text, etc. I'm tempted
to make a pitch for *preparing* for the eventual standard by making sure
that whatever we use is easily machine recognizable, like e.g. LaTeX.
But it is significantly less readable before preprocessing, and besides
which people who don't have access to it aren't likely to know it well
enough to use it in their postings anyway. Oh, well.
	Doug
-- 
Doug Merritt		{pyramid,apple}!xdos!doug
Member, Crusaders for a Better Tomorrow		Professional Wildeyed Visionary

In article <1989Sep25.033923.8607@cs.rochester.edu> ken@cs.rochester.edu writes:
>|sum from i=0 to p a sub i y sub {n-i} = sum from l=0 to q b sub i xi sub {n-i}
>								   ^^
>Probably should be \(*x (Greek Xi).

No, eqn knows the names of the Greek characters.
-- 
"Where is D.D. Harriman now,   |     Henry Spencer at U of Toronto Zoology
when we really *need* him?"    | uunet!attcan!utzoo!henry henry@zoo.toronto.edu

In article <4593@amiga.UUCP> jimm@batgirl.UUCP (Jim Mackraz) writes:
>I think it's probably far and away easier to find TeX to run on non-unix
>machines, including PC, Mac, Amiga, ST, and so on.  Also, I'd expect
>it to be available on most unix machines...
>I suggest we determine which is more common in the *subscribers* 
>environments...

Actually, I suggest we drop this entire line of debate, because it misses
the most important point.  Most people do not want to run the news articles
through a text formatter to read them.  Even those who might consider this
have a good chance of not having the relevant formatter.  Few people
are going to want to run out and get a big, slow formatter program just
so they can read postings (and they're *all* big and slow).  There is a big
difference between being able to get the software and actually having it
on hand; most non-Unix machines in practice have *neither* eqn nor TeX.
For that matter, there are a good many Unix machines -- especially System
V boxes -- that have neither.

The issue should be readability to the unaided human eye, not whose text
formatter is found on more machines.
-- 
"Where is D.D. Harriman now,   |     Henry Spencer at U of Toronto Zoology
when we really *need* him?"    | uunet!attcan!utzoo!henry henry@zoo.toronto.edu

..> Communicating equations in newsgroups:
..> both TeX and Mathematica format have been suggested.

>[Doug Merritt]:
>Unfortunately, while both TeX/LaTeX and Mathematica would be nice,
>this won't work for Usenet articles. Only a small percentage of the
>readers of your articles have either of these packages. Also, even if they
>did, it would be very inconvenient to shift articles from a normal
>news reader (like 'rn' or 'vnews') into these packages for proper display.
>Furthermore, of the people who have access to such packages, they don't
>always have access the instant they're reading them.

Some considerations:

    Mathematica is proprietary. 

    TeX is, while not quite public domain (copyright owned by
    American Mathematical Society), at least is free.

    Mathematica can produce TeX output.
    Can it accept TeX input?

What would be nice is to have Mathematica and TeX interconvertibility,
perhaps with a sink to ASCII representation.

---

Why does every new newsgroup I'm interested in have these discussions?
There was a similar discussion in sci.electronics, for exchange of
circuit diagrams.
--
Andy "Krazy" Glew,  Motorola MCD,    	    	    aglew@urbana.mcd.mot.com
1101 E. University, Urbana, IL 61801, USA.          {uunet!,}uiucuxc!udc!aglew
   
My opinions are my own; I indicate my company only so that the reader
may account for any possible bias I may have towards our products.

Related to, but distinct from the TeX vs. EQN war: does anyone know of
a way to get TeX runnable on an Epson RX-80 with graphics mode?
Ideally, ASCII only - ie. is there an nroff-like version of TeX?

--
Andy "Krazy" Glew,  Motorola MCD,    	    	    aglew@urbana.mcd.mot.com
1101 E. University, Urbana, IL 61801, USA.          {uunet!,}uiucuxc!udc!aglew
   
My opinions are my own; I indicate my company only so that the reader
may account for any possible bias I may have towards our products.