Definition of modulation index for shaped FSK

I have a pretty elementary question.

For an unshaped, 2-FSK signal, the modulation index h is defined
as the ratio of the difference between the two tone frequencies
to the symbol rate.

For shaped 2-FSK, how is h usually defined?  I can think
of a few possibilities:

(1) Base it on the peak frequency excursion exhibited by
one isolated symbol, ignoring ISI from other nearby symbols.

(2) Base it on the peak frequency excustion exhibited by
any pattern of symbols (which will be a slighty higher 
excursion than in (1), due to ISI).

(3) Base it on some sort of average.  For example, if one
looks at the pulse shape at baseband, and the area under this 
curve is the same as that of the rectangular pulse associated
with modulation index h, then we deem the shaped FSK to
also have modulation index h.

References I have looked up thus far mostly support definition (1).

Any opinions?

Steve

Steve Pope wrote:
> I have a pretty elementary question.
> 
> For an unshaped, 2-FSK signal, the modulation index h is defined
> as the ratio of the difference between the two tone frequencies
> to the symbol rate.
> 
> For shaped 2-FSK, how is h usually defined?

The difference of the frequencies, corresponding to the continuous 
infinite streams of 0s and 1ns, divided by modulation rate. I.e. DC gain 
of the pulse shaping filter.

Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

Vladimir Vassilevsky  <nospam@nowhere.com> wrote:

>Steve Pope wrote:

>> For an unshaped, 2-FSK signal, the modulation index h is defined
>> as the ratio of the difference between the two tone frequencies
>> to the symbol rate.

>> For shaped 2-FSK, how is h usually defined?

>The difference of the frequencies, corresponding to the continuous 
>infinite streams of 0s and 1ns, divided by modulation rate. I.e. DC gain 
>of the pulse shaping filter.

Works for me.  Thanks.

Steve
>Vladimir Vassilevsky
>DSP and Mixed Signal Design Consultant
>http://www.abvolt.com