Adaptive lookback indicators - page 17

 
mladen:

FilterBellow and FilterAbove - frequencies sequence that are going to be filtered (cut-off) If you set these to 0 and some big value, there will be no transform at all, since inverting initial fft will return the original values. Default values are fairly good for this purpose

[fft end pont]

Hi,

as I understand it will pass frequencies that are in between this values likle a bandpass filter and then reconstruct the signal. Is it possible to ad a amplitude (magnitude) filter for those passed frequencies before reconstruction?

 
Files:
 

This is my experiment with FFT end point version.

It has several components:

A. type- 0 it's a reconstructed signal, 1 - it is transformed signal (both use value)

B. value - used to select a value from an array. If it's a reconstructed signal use 0 to obtain end pointed FFT, if it's a transformed signal you can use any positive value to obtain that specific component from the spectrum

C. filterOP - used to specify what kind of filter you want.

0 - frequency ( one that uses FilterBellow and FilterAbove to cut of frequencies left and right), ( filters on the x axis of the transformed signal)

1- "amplitude" ( uses filterAMP to filter out any frequency that is currently between [filterAMP, - filterAMP]) ( filters on the y axis of the transformed signal)

2 - uses both options

Note: you can use reconstructed signal that filters anything except one freq. to obtain the same result like in the transformed signal

Question: I would like to try this - find the dominant cycle by Hilbert and then filter out all cycles that are further then the specified distance from that cycle. example: HT dominant cycle is 20 and distance is 10 so we filter all except the periods from [10 to 30]. Unfortunately the x axis has been normalized or something so I'm a little stuck.

Question 2: What's with the 0. and 1. array point? The 0 value is 300+, and the 1. one looks like a fast noise component. Both of them theoretically should be transformed into long period components. Does the FFT break since the x axis approaches 0 [ the end]?

Example for the 1.:

Kind regards

-- just to say I didn't find any usefulness for the filtering on the y axis of the transformed signal, it just doesn't look good if you filter all the time frequencies with the small amplitude since every signal will be cut of after it starts to fall on it's way down from the high or vice versa

Files:
 

ma pa wprsmooth

This was requested, phase accumulation on wpr smooth, the histo can be turned off or on!

 

Mrtools

Este "WPRsmooth" esta hecho sobre "Precio" ,....o sobre "Accumulacion_Distribucion" ?

 

Mrtools

This "WPRsmooth"is made on "Price",.... or "Accumulacion_Distribucion?

Thanks

 
bebeshel:
This "WPRsmooth"is made on "Price",.... or "Accumulacion_Distribucion? Thanks
bebeshel:
I mean, the indicator "Williams Percent Range "on Accumulation / Distribution "is made on the Accumulation / distribution, not on price.

or

the indicator "Williams Percent Range "on "Phase accumulation"

This is phase accumulation of wpr, wpr based on high,low and close, using jurik smoothing.Its what you asked for on the second part of your request!

 

Mrtools

May you creat the "TPL" of this image, and put in on the web.

Thanks

 

Mrtools

"pa wprsmooth_w_histo_ma.mq4 "--https://www.forex-tsd.com/advanced-elite/22848-adaptive-lookback-indicators-17.html

I can understand that this moving average on this indicator is equivalent to an exponential moving average for period 28 on the same indicator, and not on the price ?

Thanks a lot

 
bebeshel:
May you creat the "TPL" of this image, and put in on the web. Thanks

Bebeshel, you can make one for yourself 2 bollinger bands, 1st 20 period with std. deviation at 3.25 the other band period = 50 with std deviation =1.125, and Labtrend1 and the chart properties how you like, try it its easy.All the indicators are either here in advanced elite section or regular elite sections.

Reason: