Frequency monitor normalization on having multiple sources


I have a question concerning the normalization of the Transmission in frequency monitor on having multiple sources.In the example for the Pump probe simulation : ,the transmission is calculated for the probe signal at 1.5 Microns but why is the transmission results guaranteed to be normalized with respect to probe source not pump source ? Also I have a question concerning the spectrum of this example , when I compared the input and output spectrum of the probe I found that they are almost equal so this means that I have 100% transmission not 200% so why is there a gain of 2 when I display the transmission results for the frequency monitor ?


Hi @hoda-essam

The pump is mainly used for populating the states. What we detect is the probe light and thus we normalize the results with the probe. Another thing to note here is that nonlinear simulations are done in the no-norm state:

Can you please provide the results and calculations for the second part of your inquiry?



Hi @bkhanaliloo
Thanks for your reply.Concerning my first part of the question I have two comments.First I tried to use the nonorm option but I still got the same transmission and same spectrum as the cwnorm case so what’s the importance of having nonorm for nonlinear simulation when both states have same results ?
Also I get the role of the pump in the design but I still don’t understand why are the results only normalized with respect to the probe . Is it because I specified the range for monitors to include the probe only and not the pump ?
Concerning by second part of the question kindly find attached an image for the spectrum I get from the simulation.The image shows that the spectrum amplitude is almost equal for the input and output time monitors .Doesn’t that mean transmission 1 and not 2 ?

Appreciate your help.


Hi @hoda-essam

  1. Transmission results are internally normalized to source power and it will return identical results in both no-norm and cw-norm cases:
    However for other quantities where the quantity does not internally use normalizations, results of no-norm and cw-norm will be different. As an example, consider a case where there is a frequency generation i.e. f1=2*f2. While the injected light is centred at f1 (and zero power at f2), generated light is centred at f2 and normalizing the output field to input filed will be nonsense. This is also the reason why we normalize the output filed to probe light which is also centred at f2.

  2. I think your post misses the attachment. Can you please share the simulation file with me.


Hi @bkhanaliloo ,
Thanks for you reply .
1-I understand that transmission is normalized with respect to source power but my question is still the source power at which frequency.The case you mentioned which is f1 and f2 ,f2 is not originally existing as a source is results from the simulation but in case of having the probe and the pump I have both as sources .So how to I force the results to be normalized with respect to the probe frequency not the pump ? I know that the function sourcepower (f) is frequency dependent so does that mean that the results for the probe are normalized with respect to probe and results for pump are normalized with respect to pump or am I miss understanding that point ?

2-As for the simulation file it is same as that of the example.I just visualized the spectrum for input and output monitor and compared the results on the same visualizer window.I will try to attach the image again.





Hi @hoda-essam

  1. When you use transmission function, results are normalized to injected power as a function of wavelength/frequency. If you have more than one source, then injected source power will be the sum of injected power by all the available sources in the simulation region. In this case, since the pump and probe are at two far wavelengths, essentially there is no power coming from pump at 1.5um. As a result you can safely say that the power at 1.5um is coming just from the probe.

  2. Have you tried to measure the reflection? The input time monitor is affected by both probe and reflected light from the nonlinear material at 1.5um.


Hi @bkhanaliloo
I approached the input time monitor to the source and measured the transmission before the source and after the input time monitor and found out that the backward transmission is almost 0.15 which was the same amount deducted from the forward transmission so is this a correct approach ?
If so then reflections coming from the nonlinear material are still not strong enough to increase the input spectrum to become almost equivalent to the output one Am I right ?
Thanks alot.


Hi @hoda-essam

The time monitors record the field for the entire simulation time that includes pump, probe, and generated light by gain medium. However, transmission monitors use apodization to record the generated light from the gain medium. If you look at output time monitor, there are a few peaks:

A zoomed in plot around “gain material” section of output time monitor:

If you look at the “T” monitor, it uses apodization to record only this field:

However, when you look at the spectrum from time monitors, you are Fourier transforming over the entire time domain field. As a results, the spectrum look similar. You can repeat the results by Fourier transforming over only a portion of electric field and then compare the spectrum.

Hope this was helpful.

Pump probe simulation in 3D

Hi @bkhanaliloo
Thank you for your detailed reply I have two questions though.
1-I read about apodization but I still can;t understand why we need to use apodization in this example.Or in other words what will happen if we don’t use apodization.I disabled apodization and found that the transmission jumps to very high values but i don’t understand why .

2-I understand for the time monitor that the spectrum should look alike what I am asking about is the amplitude of the spectrum and I guess it should change from one point to another .Isn’t that correct ?

Appreciate you help and thanks alot.


Hi @hoda-essam

  1. We are interested only in collecting the generated light by the probe. Some light might be generated before the probe light arrival that we are not interested in. We are only interested in stimulated emission not the spontaneous emission. This is the reason that the results with and without apodization are different.

  2. The structure is periodic, thus it should not change along the y-direction. Along x-direction it only matters if it is before or after the gain material. In this case the results for these two monitors are similar because the spontaneous emission is way more stronger than stimulated emission and it travels along both -x and +x directions.

Hope this was helpful.