[ibis-macro] Re: question about Gaussian parameter
- From: Ambrish Varma <ambrishv@xxxxxxxxxxx>
- To: "ibis-macro@xxxxxxxxxxxxx" <ibis-macro@xxxxxxxxxxxxx>
- Date: Mon, 21 Mar 2011 14:53:30 -0700
Arpad,
I am not sure if this issue is going to be resolved with the Table
clarification BIRD. This is more related to Gaussian, Dual-Dirac and DjRj
Formats. The Jitter BIRD sounds like the most likely place for this stuff.
That being said, with the Table clarification, we would not have a need for
Arrays.
Regards,
Ambrish.
[cid:image002.gif@01CBE7EF.7EE64010]
Ambrish Varma | Member of Consulting Staff
P: 978.262.6431 www.cadence.com<http://www.cadence.com>
________________________________
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Muranyi, Arpad
Sent: Monday, March 21, 2011 9:31 AM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
Ambrish,
Is this topic going to be taken care of by the BIRD
draft on the Table syntax you are working on?
Thanks,
Arpad
==========================================================
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Ambrish Varma
Sent: Thursday, March 18, 2010 1:59 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
Hi Arpad,
Sorry for the delay in response.
As per my understanding, with the new changes, (PDF 0 0.5 0.01) will, in
future, only be possible with the addition of the array sub-parameter.
I believe having Array will provide an option to the model maker to accept a
string such as:
"(mySampleAMI (txtaps .05 -.1 .7 .1 -.05) (strength 6))"
Instead of
"(mySampleAMI (txtaps (-2 .05) (-1 -.1) (0 .7) (1 .1) (2 -.05)) (strength 6))"
Not sure if that was what you were looking for.
Thanks,
Ambrish.
________________________________
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Muranyi, Arpad
Sent: Tuesday, March 16, 2010 2:25 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
Ambrish,
Thanks for the explanation. I must admit that I need to brush
up a little on my AMI spec memorization... :-)
In that case, can you or someone explain to me what the issue
is with that .ami DLL which can't use this name / value list
format? Why did it need a vector or array instead of the list?
Thanks,
Arpad
=================================================================
________________________________
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Ambrish Varma
Sent: Tuesday, March 16, 2010 1:21 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
Arpad,
"(PDF 0 0.5 0.01)" is not illegal. The rules clearly state that
| 3. A parameter value in a name/value pair can be either a single value or a
| list of values separated by whitespace.
As stated on page 186 of IBIS 5.0. There is an example that demonstrates that
as well.
Thanks,
Ambrish.
________________________________
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Muranyi, Arpad
Sent: Tuesday, March 16, 2010 2:07 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
As far as I can tell, this format "(PDF 0 0.5 0.01)" is not only
questionable in its interpretation, but doesn't it violate
the concept of name/value pair we have in the tree syntax?
If I am not mistaken about that, this seems to be related to
the problem we ran into with the tap coefficients, where one
vendor we all know uses a similar one name, multiple number
format, which might be considered a vector (or array), that
can't be done according to the current official AMI spec.
Arpad
==============================================================
________________________________
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Ambrish Varma
Sent: Tuesday, March 16, 2010 12:02 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
Radek,
There is a slight problem with Gaussian, Dual-Dirac and DJRJ (more so with
Dual-Dirac and DjRj than with Gaussian).
Dual-Dirac can be expressed in this format: (PDF (Format Dual-Dirac <mean>
<mean> <sigma>)) for ex: (PDF (Format Dual-Dirac 0 0.5 0.01))
and
DjRj in this format: (PDF (Format Dual-Dirac <meanDj> < maxDj > <sigma>)) for
ex: (PDF (Format Dual-Dirac 0 0.5 0.01))
So if you don't specify if its Dual-Dirac or DjRj, the pram input string will
look like this for both:
(PDF 0 0.5 0.01)
How would you know which type of format to apply to PDF.
Thanks,
Ambrish.
________________________________
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of radek_biernacki@xxxxxxxxxxx
Sent: Tuesday, March 16, 2010 12:47 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
Walter,
I do not think we can "remove" Gaussian from the specification. It is already
there.
As a matter of fact we should try to clarify issues like this one, wherever
there is a possibility for misinterpretation.
For Fangyi's question I think the logically correct answer should be the second
one:
(PDF 0.0 0.01)
It follows the concept of (leaf_name value) where "value" in this case
consists of two numbers.
Radek
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Walter Katz
Sent: Tuesday, March 16, 2010 6:25 AM
To: RAO,FANGYI (A-USA,ex1); ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: question about Gaussian parameter
Fangyi,
One of the reasons why I have proposed removing Gaussian as a type in the new
BIRD.
Walter
Walter Katz
303.449-2308
Mobile 720.333-1107
wkatz@xxxxxxxxxx
www.sisoft.com
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx]On Behalf Of fangyi_rao@xxxxxxxxxxx
Sent: Tuesday, March 16, 2010 1:24 AM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] question about Gaussian parameter
Hi, Experts;
If an AMI model has a parameter (PDF (Format Gaussian 0.0 0.01)), what should
the Init input parameter string be? Should it be
(PDF Gaussian 0.0 0.01)
or
(PDF 0.0 0.01)
or other format?
Thanks in advance,
Fangyi
Other related posts:
