FPGA Central

I'm trying to think of examples where using an anti-alias filter is a
bad idea, or must at least be approached with extreme caution. I
already know about control systems and video applications, and I believe
that this is a big issue with EKG machines.

Does anyone else have any examples that they can share?

If someone has a sampled EKG output in some readable file format that
they'd be willing to share I would be grateful -- I can whomp something
up out of nothing, but then some doctor will read my article and know
it's BS.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Tim Wescott wrote:

> I'm trying to think of examples where using an anti-alias filter is a
> bad idea, or must at least be approached with extreme caution.

Digital Oscilloscopes and misc. sensor data acquisition equipment, the
control systems with the ADC in the loop. Anywhere you may be interested
in the instantaneous value rather then in the averaged waveform.

Vladimir Vassilevsky

DSP and Mixed Signal Design Consultant

http://www.abvolt.com

Vladimir Vassilevsky wrote:

>
>
> Tim Wescott wrote:
>
>> I'm trying to think of examples where using an anti-alias filter is a
>> bad idea, or must at least be approached with extreme caution.
>
>
> Digital Oscilloscopes

Check.

> the control systems with the ADC in the loop.

Check. Actually treating this separately because measurement equipment
can have _delay_ but no time distortion, while a control loop can't even
have delay.

> and misc. sensor data acquisition equipment, Anywhere you may be interested
> in the instantaneous value rather then in the averaged waveform.
>
Yes, but do you have some specific examples that will make good charts
for the non expert in the field of the measurement?

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Tim Wescott wrote:
> I'm trying to think of examples where using an anti-alias filter is a
> bad idea, or must at least be approached with extreme caution. I
> already know about control systems and video applications, and I believe
> that this is a big issue with EKG machines.
>
> Does anyone else have any examples that they can share?
>

I believe that this would be a consideration any time transient or
impulsive phenomena are being digitized. However, Nyquist still holds,
and if the anti-aliasing filter is interfering with your recording,
it's an indication that you aren't sampling fast enough and the filter
bandwidth needs to be wider.

Some people try to use overly-sharp (eg: Butterworth or Chebyshev)
analog AA filters to push the cutoff closer to Nyquist. These filters
have a tendency to ring, even when implemented at low orders--not good
for sampling transients.

Why would EKG recorders be especially problematic? If they are, then
maybe EEG and polygraph recorders too?

One trick that I used in process control instrumentation was to
oversample by a factor of several hundred, do the main anti-aliasing
digitally and then decimate to a more appropriate sample rate. This
means that the analog anti-aliasing filter can be very cheap and when
the sampling can be synchronized to a multiple of the power line
frequency, you can get a double whammy.

Tim Wescott <tim@seemywebsite.com> wrote in news:1q-
dndrgmcy9n4XYnZ2dnUVZ_vWdnZ2d@web-ster.com:

> while a control loop can't even
> have delay.
>

That seems somewhat overly generalized, no?

--
Scott
Reverse name to reply

John Hadstate wrote:

> Tim Wescott wrote:
>
>>I'm trying to think of examples where using an anti-alias filter is a
>>bad idea, or must at least be approached with extreme caution. I
>>already know about control systems and video applications, and I believe
>>that this is a big issue with EKG machines.
>>
>>Does anyone else have any examples that they can share?
>>
>
>
> I believe that this would be a consideration any time transient or
> impulsive phenomena are being digitized. However, Nyquist still holds,
> and if the anti-aliasing filter is interfering with your recording,
> it's an indication that you aren't sampling fast enough and the filter
> bandwidth needs to be wider.

Nyquist holds, but it can be difficult to go from a real-world (i.e.
inherently infinite bandwidth) signal to a bandlimited one without
ringing -- I'm discussing this, in the context of how you need to be
aware of the issue when establishing your sampling rate.
>
> Some people try to use overly-sharp (eg: Butterworth or Chebyshev)
> analog AA filters to push the cutoff closer to Nyquist. These filters
> have a tendency to ring, even when implemented at low orders--not good
> for sampling transients.

No kidding!
>
> Why would EKG recorders be especially problematic? If they are, then
> maybe EEG and polygraph recorders too?

AFAIK they aren't especially problematic, it's just one that I'm
familiar with being an issue -- goodness knows why, since I've never
worked with medical electronics. I just recall discussions about the
impact of sampling on EKG waveforms. I do know that if you're going to
have problems with ringing or rounding that big old spike in the middle
of an EKG waveform is going to make it happen -- particularly since
there are interesting features of the waveform that happen just before
and after the spike, and which would be obscured by ringing.
>
> One trick that I used in process control instrumentation was to
> oversample by a factor of several hundred, do the main anti-aliasing
> digitally and then decimate to a more appropriate sample rate. This
> means that the analog anti-aliasing filter can be very cheap and when
> the sampling can be synchronized to a multiple of the power line
> frequency, you can get a double whammy.
>
I do that too. Doing the anti-aliasing as an average of all the samples
leading up to the current decimated sample passes DC unmolested, but
puts a notch at all the harmonics of your decimated sampling frequency
-- so any noise that _would_ alias into your system's passband gets
notched, at the least cost in phase shift in the anti-alias filter.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Tim Wescott skrev:
> > Why would EKG recorders be especially problematic? If they are, then
> > maybe EEG and polygraph recorders too?
>
> AFAIK they aren't especially problematic, it's just one that I'm
> familiar with being an issue -- goodness knows why,

I once attended a conference on DSP where somebody presented
some scheme for processing EKG (that's the heart monitor, right?)
signals.

According to that presentation, ringing and transients are severe
issues since physicians inspect the transient characteristic of the
heartbeat in order to come up with a diagnostics. Any transient
that is not "natural" (i.e. does not conform with text-book
descriptions of normal heartbeat behaviour) is, by necessity,
interpreted as an indicator of some non-healthy condition.
It would be very, very dangerous to meddle with these
systems such that the sensor system somehow alters
the transient shape of the signal.

Rune

Scott Seidman wrote:

> Tim Wescott <tim@seemywebsite.com> wrote in news:1q-
> dndrgmcy9n4XYnZ2dnUVZ_vWdnZ2d@web-ster.com:
>
>
>>while a control loop can't even
>>have delay.
>>
>
>
> That seems somewhat overly generalized, no?
>
Yes.

I can either say "that's very clever of you to have caught that", or I
can fess up to leaving an "almost" out of there somewhere.

Certainly the performance of many control loops out there are limited by
delay, and the performance of almost all "high performance" control
loops ends up getting limited by the phase shift in the plant, sensors
or controller.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Rune Allnor wrote:

> Tim Wescott skrev:
>
>>>Why would EKG recorders be especially problematic? If they are, then
>>>maybe EEG and polygraph recorders too?
>>
>>AFAIK they aren't especially problematic, it's just one that I'm
>>familiar with being an issue -- goodness knows why,
>
>
> I once attended a conference on DSP where somebody presented
> some scheme for processing EKG (that's the heart monitor, right?)
> signals.
>
> According to that presentation, ringing and transients are severe
> issues since physicians inspect the transient characteristic of the
> heartbeat in order to come up with a diagnostics. Any transient
> that is not "natural" (i.e. does not conform with text-book
> descriptions of normal heartbeat behaviour) is, by necessity,
> interpreted as an indicator of some non-healthy condition.
> It would be very, very dangerous to meddle with these
> systems such that the sensor system somehow alters
> the transient shape of the signal.
>
> Rune
>
That is my understanding, too. Come to think of it, I'm not sure if it
would be as much of a case with EEG and polygraph machines, because they
don't have the same sort of sharp transient embedded in bumps the way
EKGs do.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

"Tim Wescott" <tim@seemywebsite.com> wrote in message
news92dnYQDa9DIjYXYnZ2dnUVZ_sGdnZ2d@web-ster.com...
>
> Nyquist holds, but it can be difficult to go from a real-world (i.e.
> inherently infinite bandwidth) signal to a bandlimited one without
> ringing -- I'm discussing this, in the context of how you need to be aware
> of the issue when establishing your sampling rate.

Tim,

There are classes of windows or filters that have no ringing in one domain
and have minimum transition width in the other. That's as good as it gets!
See:
Temes, Barcilon and Marshall "The Optimization of Bandlimited Sytems" Proc
IEEE Vol 61 No. 2 Feb 1973 pp 196-234

I believe one can find approximations to the optimum that are close
enough....

Fred

>>>>> "Tim" == Tim Wescott <tim@seemywebsite.com> writes:

Tim> I'm trying to think of examples where using an anti-alias filter is a
Tim> bad idea, or must at least be approached with extreme caution. I
Tim> already know about control systems and video applications, and I
Tim> believe that this is a big issue with EKG machines.

Isn't there always some kind of anti-alias filter, whether you want it
or not? Or are you saying that aliasing is better than whatever
artifacts an anti-aliasing filter would produce?

Just curious,

Ray

Raymond Toy wrote:

>>>>>>"Tim" == Tim Wescott <tim@seemywebsite.com> writes:
>
>
> Tim> I'm trying to think of examples where using an anti-alias filter is a
> Tim> bad idea, or must at least be approached with extreme caution. I
> Tim> already know about control systems and video applications, and I
> Tim> believe that this is a big issue with EKG machines.
>
> Isn't there always some kind of anti-alias filter, whether you want it
> or not? Or are you saying that aliasing is better than whatever
> artifacts an anti-aliasing filter would produce?
>
> Just curious,
>
> Ray

There's always some sort of filtering going on, because no system has
infinite bandwidth. I _am_ saying, however, that in many instances
(such as EKG strips, video, or control systems) aliasing is better than
the artifacts that you'd get from a filter that you could honestly call
"anti-aliasing".

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Tim Wescott wrote:

> There's always some sort of filtering going on, because no system has
> infinite bandwidth. I _am_ saying, however, that in many instances
> (such as EKG strips, video, or control systems) aliasing is better than
> the artifacts that you'd get from a filter that you could honestly call
> "anti-aliasing".
>

i'm interested how any aliasing artifacts could ever be acceptable. The
signal is corrupted, you're saying this corruption is quantifiable? I
would think the anti-aliasing filter signal corruption is more
quantifiable.

"Tim Wescott" <tim@seemywebsite.com> wrote in message
news:GMCdnRPEzZNTtoXYnZ2dnUVZ_o2dnZ2d@web-ster.com...
> Raymond Toy wrote:
>
>>>>>>>"Tim" == Tim Wescott <tim@seemywebsite.com> writes:
>>
>>
>> Tim> I'm trying to think of examples where using an anti-alias filter
>> is a
>> Tim> bad idea, or must at least be approached with extreme caution.
>> I
>> Tim> already know about control systems and video applications, and I
>> Tim> believe that this is a big issue with EKG machines.
>>
>> Isn't there always some kind of anti-alias filter, whether you want it
>> or not? Or are you saying that aliasing is better than whatever
>> artifacts an anti-aliasing filter would produce?
>>
>> Just curious,
>>
>> Ray
>
> There's always some sort of filtering going on, because no system has
> infinite bandwidth. I _am_ saying, however, that in many instances (such
> as EKG strips, video, or control systems) aliasing is better than the
> artifacts that you'd get from a filter that you could honestly call
> "anti-aliasing".

Tim,

As you know, it's all about what you mean by "honestly".
No real filter is perfect for theoretical anti-aliasing purposes.
Many real signals don't threaten to cause aliases that are troublesome.
Of course, there is always a pathological case that one can theorize
about....

Fred

>>>>> "Tim" == Tim Wescott <tim@seemywebsite.com> writes:

Tim> Raymond Toy wrote:

>>>>>>> "Tim" == Tim Wescott <tim@seemywebsite.com> writes:
Tim> I'm trying to think of examples where using an anti-alias
>> filter is a
Tim> bad idea, or must at least be approached with extreme caution. I
Tim> already know about control systems and video applications, and I
Tim> believe that this is a big issue with EKG machines.
>> Isn't there always some kind of anti-alias filter, whether you want
>> it
>> or not? Or are you saying that aliasing is better than whatever
>> artifacts an anti-aliasing filter would produce?
>> Just curious,
>> Ray

Tim> There's always some sort of filtering going on, because no system has
Tim> infinite bandwidth. I _am_ saying, however, that in many instances
Tim> (such as EKG strips, video, or control systems) aliasing is better
Tim> than the artifacts that you'd get from a filter that you could
Tim> honestly call "anti-aliasing".

I think I understand what you're saying. Basically, you have some
kind of a priori knowledge that the signal is bandlimited and you're
sampling at a high enough frequency that aliasing isn't there. No
problem with that.

But that does bring up the question of how you got the a priori
knowledge.

But unless you do the experiment, it's kind of hard to tell if
anti-aliasing filter effects are worse than aliasing effects. Maybe
those EKG strips are mostly aliasing artifacts. :-)

Ray

Raymond Toy <raymond.toy@ericsson.com> writes:

>>>>>> "Tim" == Tim Wescott <tim@seemywebsite.com> writes:
>
> Tim> I'm trying to think of examples where using an anti-alias filter is a
> Tim> bad idea, or must at least be approached with extreme caution. I
> Tim> already know about control systems and video applications, and I
> Tim> believe that this is a big issue with EKG machines.
>
> Isn't there always some kind of anti-alias filter, whether you want it
> or not?

Yes, but you can oversample and therefore greatly reduce the "nastiness"
of the antialias filtering. Especially for such low bandwidth signals.
--
% Randy Yates % "...the answer lies within your soul
%% Fuquay-Varina, NC % 'cause no one knows which side
%%% 919-577-9882 % the coin will fall."
%%%% <yates@ieee.org> % 'Big Wheels', *Out of the Blue*, ELO
http://home.earthlink.net/~yatescr

steve wrote:
> Tim Wescott wrote:
>
>
>>There's always some sort of filtering going on, because no system has
>>infinite bandwidth. I _am_ saying, however, that in many instances
>>(such as EKG strips, video, or control systems) aliasing is better than
>>the artifacts that you'd get from a filter that you could honestly call
>>"anti-aliasing".
>>
>
> i'm interested how any aliasing artifacts could ever be acceptable. The
> signal is corrupted, you're saying this corruption is quantifiable? I
> would think the anti-aliasing filter signal corruption is more
> quantifiable.
>
Quantifiable in both cases -- you take the original signal and the
corrupted one, you calculate the 'badness' of the corruption by whatever
criteria is important to your problem at hand, and you compare the two
numbers (hopefully you can reduce the corruption down to a number).

In the case of the EKG what's important (AFAIK) is the size and timing
of the roundish humps that occur before and after the great big spike.
If your anti-aliasing filters are contributing roundish humps of their
own, then that's a bad thing because it knocks the pattern recognition
element (the doctor's brain) for a loop. In this particular case it is
very important not to have artifacts from the spike either before or
after the spike -- so just sampling the thing and showing what you
sample may be the best.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Raymond Toy wrote:

>>>>>>"Tim" == Tim Wescott <tim@seemywebsite.com> writes:
>
>
> Tim> Raymond Toy wrote:
>
> >>>>>>> "Tim" == Tim Wescott <tim@seemywebsite.com> writes:
> Tim> I'm trying to think of examples where using an anti-alias
> >> filter is a
> Tim> bad idea, or must at least be approached with extreme caution. I
> Tim> already know about control systems and video applications, and I
> Tim> believe that this is a big issue with EKG machines.
> >> Isn't there always some kind of anti-alias filter, whether you want
> >> it
> >> or not? Or are you saying that aliasing is better than whatever
> >> artifacts an anti-aliasing filter would produce?
> >> Just curious,
> >> Ray
>
> Tim> There's always some sort of filtering going on, because no system has
> Tim> infinite bandwidth. I _am_ saying, however, that in many instances
> Tim> (such as EKG strips, video, or control systems) aliasing is better
> Tim> than the artifacts that you'd get from a filter that you could
> Tim> honestly call "anti-aliasing".
>
> I think I understand what you're saying. Basically, you have some
> kind of a priori knowledge that the signal is bandlimited and you're
> sampling at a high enough frequency that aliasing isn't there. No
> problem with that.
>
> But that does bring up the question of how you got the a priori
> knowledge.
>
> But unless you do the experiment, it's kind of hard to tell if
> anti-aliasing filter effects are worse than aliasing effects. Maybe
> those EKG strips are mostly aliasing artifacts. :-)
>
> Ray

I would contend that you need a priori knowledge about what you're going
to process, no matter what the problem space is. The less a priori
knowledge you have about your signal, the more capable -- and big, and
power hungry, and expensive, your signal processing equipment is going
to have to be.

So you always start with some a priori knowledge or at least
assumptions. If you're doing basic research you'll be inclined to get
the biggest bestest fastest equipment, but eventually you'll know enough
that you'll be wanting to build light, inexpensive stuff (itty bitty
battery powered EKG recorders, for example).

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Tim Wescott <tim@seemywebsite.com> wrote in
news:fpidnRdl2OFeuYXYnZ2dnUVZ_rudnZ2d@web-ster.com:

> I can either say "that's very clever of you to have caught that", or I
> can fess up to leaving an "almost" out of there somewhere.
>
> Certainly the performance of many control loops out there are limited by
> delay, and the performance of almost all "high performance" control
> loops ends up getting limited by the phase shift in the plant, sensors
> or controller.
>
> --
>


Most likely the latter. I can remember spending considerable time in class
figuring out just how must delay a feedback system can tolerate before
going unstable, but "almost" all of them could tolerate "some", if I recall
correctly, based upon just when the phase would kick a nice negative
feedback into the much less pleasant positive variety. Of course, delay
generally hurts performance, and often simple stability is much less than
what one needs.

While this example is extremely nonlinear, one of my favorite lab exercises
in my biomedical engineering course was simply using a pot to zero a chart
recorder (yeah, messy ink, green paper, clogged pens.... that's the stuff!)
line, but the pot was wired through a delay line. We changed the delay
until we started to oscillate. Because our brains tend to "sample" in
strange ways, and we produce semi-ballistic responses, people did
considerably better than a run of the mill feedback system would do in the
same situation. Wonderfully laid out lab, though.


--
Scott
Reverse name to reply

Tim Wescott <tim@seemywebsite.com> wrote in
news:fpidnRZl2OGEuIXYnZ2dnUVZ_rudnZ2d@web-ster.com:

> That is my understanding, too. Come to think of it, I'm not sure if it
> would be as much of a case with EEG and polygraph machines, because they
> don't have the same sort of sharp transient embedded in bumps the way
> EKGs do.
>

EEG is much more random than ECG, and you can't really learn as much even
from a perfectly recorded signal. From ECG, things like the width of the
spike (the QRS complex) are critical for understanding of the heart
physiology, as is the time delays between one part of the wave and what
follows.

The useful ECG passband only goes to 250 Hz or so. There are some really
good noise limiting techniques that are often used, like 90dB or more
common-mode rejection amplifiers, and active ground ("driven right leg")
designs. It shouldn't be that tough to get useful analog schemes, as it's
been done for decades on chart recorders (and maybe even kymographs!!). If
I needed to do it accurately and digitally, I'd probably go with a sample
rate of about 4 khz, a nice gentle anti-aliasing filter at about 400 hz,
and maybe some nonlinear smoothing (median filtering and Savitsky-Golay
would both be tested)


--
Scott
Reverse name to reply

Tim Wescott wrote:

> Quantifiable in both cases -- you take the original signal and the
> corrupted one, you calculate the 'badness' of the corruption by whatever
> criteria is important to your problem at hand, and you compare the two
> numbers (hopefully you can reduce the corruption down to a number).
>
no I meant analytically, in the general case, not by experiment

> In the case of the EKG what's important (AFAIK) is the size and timing
> of the roundish humps that occur before and after the great big spike.
> If your anti-aliasing filters are contributing roundish humps of their
> own, then that's a bad thing because it knocks the pattern recognition
> element (the doctor's brain) for a loop. In this particular case it is
> very important not to have artifacts from the spike either before or
> after the spike --

I don't know, this doesn't make any sense to me, the artifacts
generated by aliasing doesn't guarantee specific signal corruption
won't happen, which you seem to imply

Scott Seidman wrote:


> Most likely the latter. I can remember spending considerable time in class
> figuring out just how must delay a feedback system can tolerate before
> going unstable, but "almost" all of them could tolerate "some", if I recall
> correctly, based upon just when the phase would kick a nice negative
> feedback into the much less pleasant positive variety. Of course, delay
> generally hurts performance, and often simple stability is much less than
> what one needs.

The feedback loop can't provide any correction at the frequencies higher
then 1/(6*Tdelay).
The conservative rule for the stability would be having a unity loop
gain at 1/(8*Tdelay).

(This assumes that the delay is the dominating source of the phase shift
in the loop, the response is monotonic and the bandwidth of the loop is
from 0 to Fmax).

>
> While this example is extremely nonlinear, one of my favorite lab exercises
> in my biomedical engineering course was simply using a pot to zero a chart
> recorder (yeah, messy ink, green paper, clogged pens.... that's the stuff!)
> line, but the pot was wired through a delay line. We changed the delay
> until we started to oscillate. Because our brains tend to "sample" in
> strange ways, and we produce semi-ballistic responses, people did
> considerably better than a run of the mill feedback system would do in the
> same situation. Wonderfully laid out lab, though.
>



Vladimir Vassilevsky

DSP and Mixed Signal Design Consultant

http://www.abvolt.com

Scott Seidman wrote:
> Tim Wescott <tim@seemywebsite.com> wrote in
> news:fpidnRZl2OGEuIXYnZ2dnUVZ_rudnZ2d@web-ster.com:
>
>
>>That is my understanding, too. Come to think of it, I'm not sure if it
>>would be as much of a case with EEG and polygraph machines, because they
>>don't have the same sort of sharp transient embedded in bumps the way
>>EKGs do.
>>
>
>
> EEG is much more random than ECG, and you can't really learn as much even
> from a perfectly recorded signal. From ECG, things like the width of the
> spike (the QRS complex) are critical for understanding of the heart
> physiology, as is the time delays between one part of the wave and what
> follows.
>
> The useful ECG passband only goes to 250 Hz or so. There are some really
> good noise limiting techniques that are often used, like 90dB or more
> common-mode rejection amplifiers, and active ground ("driven right leg")
> designs. It shouldn't be that tough to get useful analog schemes, as it's
> been done for decades on chart recorders (and maybe even kymographs!!). If
> I needed to do it accurately and digitally, I'd probably go with a sample
> rate of about 4 khz, a nice gentle anti-aliasing filter at about 400 hz,
> and maybe some nonlinear smoothing (median filtering and Savitsky-Golay
> would both be tested)
>
>
This pretty much confirms my understanding. I've spend the last two
hours of my day finding a pretty picture of an EKG waveform on the web,
synthesizing a fake one in SciLab, and filtering it a couple of
different ways for this article I'm writing. I'm laughing right now
because right before I read your post I found out that if I filter it at
about 400Hz it looks pretty good...

In light of what I've seen and what you've just said your nonlinear
smoothing certainly sounds interesting. I would add to that that should
I be tasked with storing an EKG recording in some storage-sensitive way
I'd probably look at some adaptive compression scheme. I'm really not
sure if I'd consider lossy compression without lots of research (and
indemnification!) behind me, but one could probably do some sort of
lossless compression that would vary the compression rate depending on
the signal (and consequently possibly run out of storage -- dangit, more
indemnification, please!).

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

steve wrote:

> Tim Wescott wrote:
>
>
>>Quantifiable in both cases -- you take the original signal and the
>>corrupted one, you calculate the 'badness' of the corruption by whatever
>>criteria is important to your problem at hand, and you compare the two
>>numbers (hopefully you can reduce the corruption down to a number).
>>
>
> no I meant analytically, in the general case, not by experiment
>
One ought to be able to do it either analytically or by experiment.
Usually when I do this kind of stuff I do both -- analytically to make
sure I understand the problem, then experimentally to verify that my
analysis was correct and later to verify that my implementation was also
correct.
>
>>In the case of the EKG what's important (AFAIK) is the size and timing
>>of the roundish humps that occur before and after the great big spike.
>>If your anti-aliasing filters are contributing roundish humps of their
>>own, then that's a bad thing because it knocks the pattern recognition
>>element (the doctor's brain) for a loop. In this particular case it is
>>very important not to have artifacts from the spike either before or
>>after the spike --
>
>
> I don't know, this doesn't make any sense to me, the artifacts
> generated by aliasing doesn't guarantee specific signal corruption
> won't happen, which you seem to imply
>
I didn't mean to. Artifact == corruption, in this context.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

Tim Wescott <tim@seemywebsite.com> wrote in
news:4OmdnSi4SpDrxIXYnZ2dnUVZ_tednZ2d@web-ster.com:

> This pretty much confirms my understanding. I've spend the last two
> hours of my day finding a pretty picture of an EKG waveform on the web,
> synthesizing a fake one in SciLab, and filtering it a couple of
> different ways for this article I'm writing. I'm laughing right now
> because right before I read your post I found out that if I filter it at
> about 400Hz it looks pretty good...


I "think" I have an ECG simulator, if I can find it. It was given to me, I
have no idea if it works, and I don't know what kind of batteries it takes.
If I can dig it up, you can give it a go. Alternatively, I might be able
to record some for you, AC coupled at a frequency slightly higher than
what's wise, and no promises re aliasing. I'd use an AD instruments
powerlab, but I need to dig up an accessory pack before I could record
this. It wouldn't be standard clinical EKG lead locations, but leads I,
II, and III on the Einthoven triangle. It might take some time.

If you absolutely need the waveform from long-term acquisition (like a
Holter Monitor), I wouldn't go for lossy compression, but I'd try to dump
the data to a large SD card, or disk. If you just need heart rate, I'd do
some sort of QRS detection using a window discriminator or fancier
technique, and just save the time stamps.


--
Scott
Reverse name to reply