Financial Digital Signal Processing

 
Peace be upon you, everyone.

 If anyone aspiring to further discuss, to reset into motion this DSP dialogue that seemed to decay in function of time, and to commit oneself to try to grasp that ‘maximisably unattainable’ that nonlinearly shrouds one's cognitive and intellectual humane scope, I am all agog to set things up. We may set into motion the ongoingness of this thread by a paradox – sociocultural in nature – which had as ever haunted my perception proper since I had had begun to read out innumerable threads, that is, the rate of community collaboration openeness evolves inversely proportional to technological public availability (disponibilité), I mean, back when the access and the systematic resorting to the exploiting of publicly-available AI assistance was almost null, the spirit of public collaboration and collective endeavour strided unquenchably to its peak, and now that, all publicly-available AI assistance is being equally distributed to everyone, the bound of individuality recoiled onto itself and the spirit of public collaboration and collective endeavour collapsed almost to its trough, and, in regard to which, in my personal opinion, the opposite was supposed to happen – but again, I venture to say that this supposedness signifies nothing but a personal estimate, stemming from a sense of curiosity, unknowability, and igorance-cognizance that thrives oneself to further advance and progress.

 Therefore, who finds himself entitled enough to willingly undertake the elaboration of this interlink of transgenerational dialogue ? For Francis Bacon, “𝘬𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦 𝘵𝘩𝘢𝘵 𝘵𝘦𝘯𝘥𝘦𝘵𝘩 𝘣𝘶𝘵 𝘵𝘰 𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘪𝘰𝘯, 𝘪𝘴 𝘣𝘶𝘵 𝘢𝘴 𝘢 𝘤𝘰𝘶𝘳𝘵𝘦𝘴𝘢𝘯, 𝘸𝘩𝘪𝘤𝘩 𝘪𝘴 𝘧𝘰𝘳 𝘱𝘭𝘦𝘢𝘴𝘶𝘳𝘦, 𝘢𝘯𝘥 𝘯𝘰𝘵 𝘧𝘰𝘳 𝘧𝘳𝘶𝘪𝘵 𝘰𝘳 𝘨𝘦𝘯𝘦𝘳𝘢𝘵𝘪𝘰𝘯.” Its concern is not “𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘪𝘰𝘯, 𝘸𝘩𝘪𝘤𝘩 𝘮𝘦𝘯 𝘤𝘢𝘭𝘭 𝘵𝘳𝘶𝘵𝘩,” but “𝘰𝘱𝘦𝘳𝘢𝘵𝘪𝘰𝘯,” the effective procedure. The “𝘵𝘳𝘶𝘦 𝘦𝘯𝘥, 𝘴𝘤𝘰𝘱𝘦 𝘰𝘳 𝘰𝘧𝘧𝘪𝘤𝘦 𝘰𝘧 𝘬𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦” does not consist in “𝘢𝘯𝘺 𝘱𝘭𝘢𝘶𝘴𝘪𝘣𝘭𝘦, 𝘥𝘦𝘭𝘦𝘤𝘵𝘢𝘣𝘭𝘦, 𝘳𝘦𝘷𝘦𝘳𝘦𝘯𝘥 𝘰𝘳 𝘢𝘥𝘮𝘪𝘳𝘦𝘥 𝘥𝘪𝘴𝘤𝘰𝘶𝘳𝘴𝘦, 𝘰𝘳 𝘢𝘯𝘺 𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘰𝘳𝘺 𝘢𝘳𝘨𝘶𝘮𝘦𝘯𝘵𝘴, 𝘣𝘶𝘵 𝘪𝘯 𝘦𝘧𝘧𝘦𝘤𝘵𝘪𝘯𝘨 𝘢𝘯𝘥 𝘸𝘰𝘳𝘬𝘪𝘯𝘨, 𝘢𝘯𝘥 𝘪𝘯 𝘥𝘪𝘴𝘤𝘰𝘷𝘦𝘳𝘺 𝘰𝘧 𝘱𝘢𝘳𝘵𝘪𝘤𝘶𝘭𝘢𝘳𝘴 𝘯𝘰𝘵 𝘳𝘦𝘷𝘦𝘢𝘭𝘦𝘥 𝘣𝘦𝘧𝘰𝘳𝘦, 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘣𝘦𝘵𝘵𝘦𝘳 𝘦𝘯𝘥𝘰𝘸𝘮𝘦𝘯𝘵 𝘢𝘯𝘥 𝘩𝘦𝘭𝘱 𝘰𝘧 𝘮𝘢𝘯’𝘴 𝘭𝘪𝘧𝘦.” (Bacon, “Valerius Terminus: Of the Interpretation of Nature.” 𝘔𝘪𝘴𝘤𝘦𝘭𝘭𝘢𝘯𝘦𝘰𝘶𝘴 𝘛𝘳𝘢𝘤𝘵𝘴 𝘶𝘱𝘰𝘯 𝘏𝘶𝘮𝘢𝘯 𝘒𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦, 𝘞𝘰𝘳𝘬𝘴, Vol. I, p. 281.)

 The scope of my quest mainly envisages the quantitatively digital signal processing domain (DSP), in view of sharing articles, documents, statistical rapports, and mql4-based algorithms in order for the community to engage into a collective giving-and-taking of mutual collaborative benefice.
 
Chiheb Azouz:
Peace be upon you, everyone.

 If anyone aspiring to further discuss, to reset into motion this DSP dialogue that seemed to decay in function of time, and to commit oneself to try to grasp that ‘maximisably unattainable’ that nonlinearly shrouds one's cognitive and intellectual humane scope, I am all agog to set things up. We may set into motion the ongoingness of this thread by a paradox – sociocultural in nature – which had as ever haunted my perception proper since I had had begun to read out innumerable threads, that is, the rate of community collaboration openeness evolves inversely proportional to technological public availability (disponibilité), I mean, back when the access and the systematic resorting to the exploiting of publicly-available AI assistance was almost null, the spirit of public collaboration and collective endeavour strided unquenchably to its peak, and now that, all publicly-available AI assistance is being equally distributed to everyone, the bound of individuality recoiled onto itself and the spirit of public collaboration and collective endeavour collapsed almost to its trough, and, in regard to which, in my personal opinion, the opposite was supposed to happen – but again, I venture to say that this supposedness signifies nothing but a personal estimate, stemming from a sense of curiosity, unknowability, and igorance-cognizance that thrives oneself to further advance and progress.

 Therefore, who finds himself entitled enough to willingly undertake the elaboration of this interlink of transgenerational dialogue ? For Francis Bacon, “𝘬𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦 𝘵𝘩𝘢𝘵 𝘵𝘦𝘯𝘥𝘦𝘵𝘩 𝘣𝘶𝘵 𝘵𝘰 𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘪𝘰𝘯, 𝘪𝘴 𝘣𝘶𝘵 𝘢𝘴 𝘢 𝘤𝘰𝘶𝘳𝘵𝘦𝘴𝘢𝘯, 𝘸𝘩𝘪𝘤𝘩 𝘪𝘴 𝘧𝘰𝘳 𝘱𝘭𝘦𝘢𝘴𝘶𝘳𝘦, 𝘢𝘯𝘥 𝘯𝘰𝘵 𝘧𝘰𝘳 𝘧𝘳𝘶𝘪𝘵 𝘰𝘳 𝘨𝘦𝘯𝘦𝘳𝘢𝘵𝘪𝘰𝘯.” Its concern is not “𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘪𝘰𝘯, 𝘸𝘩𝘪𝘤𝘩 𝘮𝘦𝘯 𝘤𝘢𝘭𝘭 𝘵𝘳𝘶𝘵𝘩,” but “𝘰𝘱𝘦𝘳𝘢𝘵𝘪𝘰𝘯,” the effective procedure. The “𝘵𝘳𝘶𝘦 𝘦𝘯𝘥, 𝘴𝘤𝘰𝘱𝘦 𝘰𝘳 𝘰𝘧𝘧𝘪𝘤𝘦 𝘰𝘧 𝘬𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦” does not consist in “𝘢𝘯𝘺 𝘱𝘭𝘢𝘶𝘴𝘪𝘣𝘭𝘦, 𝘥𝘦𝘭𝘦𝘤𝘵𝘢𝘣𝘭𝘦, 𝘳𝘦𝘷𝘦𝘳𝘦𝘯𝘥 𝘰𝘳 𝘢𝘥𝘮𝘪𝘳𝘦𝘥 𝘥𝘪𝘴𝘤𝘰𝘶𝘳𝘴𝘦, 𝘰𝘳 𝘢𝘯𝘺 𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘰𝘳𝘺 𝘢𝘳𝘨𝘶𝘮𝘦𝘯𝘵𝘴, 𝘣𝘶𝘵 𝘪𝘯 𝘦𝘧𝘧𝘦𝘤𝘵𝘪𝘯𝘨 𝘢𝘯𝘥 𝘸𝘰𝘳𝘬𝘪𝘯𝘨, 𝘢𝘯𝘥 𝘪𝘯 𝘥𝘪𝘴𝘤𝘰𝘷𝘦𝘳𝘺 𝘰𝘧 𝘱𝘢𝘳𝘵𝘪𝘤𝘶𝘭𝘢𝘳𝘴 𝘯𝘰𝘵 𝘳𝘦𝘷𝘦𝘢𝘭𝘦𝘥 𝘣𝘦𝘧𝘰𝘳𝘦, 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘣𝘦𝘵𝘵𝘦𝘳 𝘦𝘯𝘥𝘰𝘸𝘮𝘦𝘯𝘵 𝘢𝘯𝘥 𝘩𝘦𝘭𝘱 𝘰𝘧 𝘮𝘢𝘯’𝘴 𝘭𝘪𝘧𝘦.” (Bacon, “Valerius Terminus: Of the Interpretation of Nature.” 𝘔𝘪𝘴𝘤𝘦𝘭𝘭𝘢𝘯𝘦𝘰𝘶𝘴 𝘛𝘳𝘢𝘤𝘵𝘴 𝘶𝘱𝘰𝘯 𝘏𝘶𝘮𝘢𝘯 𝘒𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦, 𝘞𝘰𝘳𝘬𝘴, Vol. I, p. 281.)

 The scope of my quest mainly envisages the quantitatively digital signal processing domain (DSP), in view of sharing articles, documents, statistical rapports, and mql4-based algorithms in order for the community to engage into a collective giving-and-taking of mutual collaborative benefice.

Code Base

Institutional Fourier Transform (DFT) Dominant Cycle Language: MQL5

Amanda Vitoria De Paula Pereira, 2026.04.28 14:12

A digital signal processing (DSP) engine that applies the Discrete Fourier Transform (DFT) to market data, isolating the dominant cyclical frequency to project turning points and eliminate phase-lag.

 
Chiheb Azouz:
Peace be upon you, everyone.

 If anyone aspiring to further discuss, to reset into motion this DSP dialogue that seemed to decay in function of time, and to commit oneself to try to grasp that ‘maximisably unattainable’ that nonlinearly shrouds one's cognitive and intellectual humane scope, I am all agog to set things up. We may set into motion the ongoingness of this thread by a paradox – sociocultural in nature – which had as ever haunted my perception proper since I had had begun to read out innumerable threads, that is, the rate of community collaboration openeness evolves inversely proportional to technological public availability (disponibilité), I mean, back when the access and the systematic resorting to the exploiting of publicly-available AI assistance was almost null, the spirit of public collaboration and collective endeavour strided unquenchably to its peak, and now that, all publicly-available AI assistance is being equally distributed to everyone, the bound of individuality recoiled onto itself and the spirit of public collaboration and collective endeavour collapsed almost to its trough, and, in regard to which, in my personal opinion, the opposite was supposed to happen – but again, I venture to say that this supposedness signifies nothing but a personal estimate, stemming from a sense of curiosity, unknowability, and igorance-cognizance that thrives oneself to further advance and progress.

 Therefore, who finds himself entitled enough to willingly undertake the elaboration of this interlink of transgenerational dialogue ? For Francis Bacon, “𝘬𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦 𝘵𝘩𝘢𝘵 𝘵𝘦𝘯𝘥𝘦𝘵𝘩 𝘣𝘶𝘵 𝘵𝘰 𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘪𝘰𝘯, 𝘪𝘴 𝘣𝘶𝘵 𝘢𝘴 𝘢 𝘤𝘰𝘶𝘳𝘵𝘦𝘴𝘢𝘯, 𝘸𝘩𝘪𝘤𝘩 𝘪𝘴 𝘧𝘰𝘳 𝘱𝘭𝘦𝘢𝘴𝘶𝘳𝘦, 𝘢𝘯𝘥 𝘯𝘰𝘵 𝘧𝘰𝘳 𝘧𝘳𝘶𝘪𝘵 𝘰𝘳 𝘨𝘦𝘯𝘦𝘳𝘢𝘵𝘪𝘰𝘯.” Its concern is not “𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘪𝘰𝘯, 𝘸𝘩𝘪𝘤𝘩 𝘮𝘦𝘯 𝘤𝘢𝘭𝘭 𝘵𝘳𝘶𝘵𝘩,” but “𝘰𝘱𝘦𝘳𝘢𝘵𝘪𝘰𝘯,” the effective procedure. The “𝘵𝘳𝘶𝘦 𝘦𝘯𝘥, 𝘴𝘤𝘰𝘱𝘦 𝘰𝘳 𝘰𝘧𝘧𝘪𝘤𝘦 𝘰𝘧 𝘬𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦” does not consist in “𝘢𝘯𝘺 𝘱𝘭𝘢𝘶𝘴𝘪𝘣𝘭𝘦, 𝘥𝘦𝘭𝘦𝘤𝘵𝘢𝘣𝘭𝘦, 𝘳𝘦𝘷𝘦𝘳𝘦𝘯𝘥 𝘰𝘳 𝘢𝘥𝘮𝘪𝘳𝘦𝘥 𝘥𝘪𝘴𝘤𝘰𝘶𝘳𝘴𝘦, 𝘰𝘳 𝘢𝘯𝘺 𝘴𝘢𝘵𝘪𝘴𝘧𝘢𝘤𝘵𝘰𝘳𝘺 𝘢𝘳𝘨𝘶𝘮𝘦𝘯𝘵𝘴, 𝘣𝘶𝘵 𝘪𝘯 𝘦𝘧𝘧𝘦𝘤𝘵𝘪𝘯𝘨 𝘢𝘯𝘥 𝘸𝘰𝘳𝘬𝘪𝘯𝘨, 𝘢𝘯𝘥 𝘪𝘯 𝘥𝘪𝘴𝘤𝘰𝘷𝘦𝘳𝘺 𝘰𝘧 𝘱𝘢𝘳𝘵𝘪𝘤𝘶𝘭𝘢𝘳𝘴 𝘯𝘰𝘵 𝘳𝘦𝘷𝘦𝘢𝘭𝘦𝘥 𝘣𝘦𝘧𝘰𝘳𝘦, 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘣𝘦𝘵𝘵𝘦𝘳 𝘦𝘯𝘥𝘰𝘸𝘮𝘦𝘯𝘵 𝘢𝘯𝘥 𝘩𝘦𝘭𝘱 𝘰𝘧 𝘮𝘢𝘯’𝘴 𝘭𝘪𝘧𝘦.” (Bacon, “Valerius Terminus: Of the Interpretation of Nature.” 𝘔𝘪𝘴𝘤𝘦𝘭𝘭𝘢𝘯𝘦𝘰𝘶𝘴 𝘛𝘳𝘢𝘤𝘵𝘴 𝘶𝘱𝘰𝘯 𝘏𝘶𝘮𝘢𝘯 𝘒𝘯𝘰𝘸𝘭𝘦𝘥𝘨𝘦, 𝘞𝘰𝘳𝘬𝘴, Vol. I, p. 281.)

 The scope of my quest mainly envisages the quantitatively digital signal processing domain (DSP), in view of sharing articles, documents, statistical rapports, and mql4-based algorithms in order for the community to engage into a collective giving-and-taking of mutual collaborative benefice.
interesting ...
 
Ryan L Johnson #:

I won't verbally heed you with the oft-repeated evident observation that the Discrete Fourier Decomposition considers as ‘slower’ in terms of computational speed (O(N²)) compared to its FFT counterpart (O(NlogN)), but again, the elaboration of one's view functionally depends upon one's needs and goals, whether it be for an electrodynamic engineer or a market trader, in the way that Nietzsche proclaims that every fact is interpretably multi-perspectival (“𝘈𝘨𝘢𝘪𝘯𝘴𝘵 𝘵𝘩𝘢𝘵 𝘱𝘰𝘴𝘪𝘵𝘪𝘷𝘪𝘴𝘮 𝘸𝘩𝘪𝘤𝘩 𝘴𝘵𝘰𝘱𝘴 𝘣𝘦𝘧𝘰𝘳𝘦 𝘱𝘩𝘦𝘯𝘰𝘮𝘦𝘯𝘢, 𝘴𝘢𝘺𝘪𝘯𝘨 ‘𝘵𝘩𝘦𝘳𝘦 𝘢𝘳𝘦 𝘰𝘯𝘭𝘺 𝘧𝘢𝘤𝘵𝘴’, 𝘐 𝘴𝘩𝘰𝘶𝘭𝘥 𝘴𝘢𝘺 : 𝘯𝘰, 𝘪𝘵 𝘪𝘴 𝘱𝘳𝘦𝘤𝘪𝘴𝘦𝘭𝘺 𝘧𝘢𝘤𝘵𝘴 𝘵𝘩𝘢𝘵 𝘥𝘰 𝘯𝘰𝘵 𝘦𝘹𝘪𝘴𝘵, 𝘰𝘯𝘭𝘺 𝘪𝘯𝘵𝘦𝘳𝘱𝘳𝘦𝘵𝘢𝘵𝘪𝘰𝘯𝘴.” 𝘛𝘩𝘦 𝘗𝘰𝘳𝘵𝘢𝘣𝘭𝘦 𝘕𝘪𝘦𝘵𝘻𝘴𝘤𝘩𝘦 (1954) by Walter Kaufmann, p. 458). The DFT preserves some mathematical characteristics that the FFT overlooks, one of which is that the DFT permits to calculate the transform for any arbitrary frequency, not just predetermined fixed bins, and thus the sustaining of the 𝒏𝒐𝒏𝒖𝒏𝒊𝒇𝒐𝒓𝒎 power spectrum (peaks occuring at non-integer multiples of the FFT bin width) which the FFT outcasts due to its dyadic decomposition structure (Cooley-Tukey, etc.,) In either case, however, financial time series are statistically prone to be highly nonstationary (see for example Fama & French (1998) “𝘗𝘦𝘳𝘮𝘢𝘯𝘦𝘯𝘵 𝘢𝘯𝘥 𝘛𝘦𝘮𝘱𝘰𝘳𝘢𝘳𝘺 𝘊𝘰𝘮𝘱𝘰𝘯𝘦𝘯𝘵𝘴 𝘰𝘧 𝘚𝘵𝘰𝘤𝘬 𝘗𝘳𝘪𝘤𝘦𝘴” wherein they explore how stock prices have a nonstationary ‘permanent’ component (random walk) and a ‘temporary’ stationary component. See also Schmidt et al. (2013/2025) : “𝘕𝘰𝘯𝘴𝘵𝘢𝘵𝘪𝘰𝘯𝘢𝘳𝘪𝘵𝘺 𝘪𝘯 𝘍𝘪𝘯𝘢𝘯𝘤𝘪𝘢𝘭 𝘛𝘪𝘮𝘦 𝘚𝘦𝘳𝘪𝘦𝘴 𝘢𝘯𝘥 𝘎𝘦𝘯𝘦𝘳𝘪𝘤 𝘍𝘦𝘢𝘵𝘶𝘳𝘦𝘴.” Wherein they argue that standard statistical mechanics fails in finance because sample-averaged observables (like the mean and variance computed in the 𝒘𝒊𝒏𝒅𝒐𝒘_𝒔𝒖𝒎 function you had provided concerning this DFT source code proper) depend entirely on the time window chosen – which, naturally, changes over time as the former progressively and regressively scales up and down. As is the case for the fixed 𝒏𝒐𝒓𝒎𝒂𝒍𝒊𝒛𝒆𝒅_𝒂𝒎𝒑𝒍𝒊𝒕𝒖𝒅𝒆 output function, in regard to which volatility (variance) changes over time, or that the 𝒘𝒊𝒏𝒅𝒐𝒘_𝒎𝒆𝒂𝒏 output function shall lag significantly during a ‘stochastic jump' in data, or the DFT might output cycles that are actually mere outlier noise. I recommend as a starter to read these three documents, for I think that the Fourier Transform (or any spectral decomposition algorithm for that matter) assumes a specific role within this intricate mesh of mathematically reality, and it returns for our Intellect to assign the boundaries of that role proper.

Multiscale Fast Fourier Transform homogenization of additively manufactured fiber reinforced composites from component-wise description of morphology (2023) 

Efficient Simultaneous Multi-Scale Computation of FFTs (Dave Cohen, Camille Goudeseune, and Mark Hasegawa-Johnson, 2009)

Fractional Brownian Motions in Financial Models, Simulation and Pricing (Chun Ming Tam, 2015)
 
Chiheb Azouz #:
I won't verbally heed you with the oft-repeated evident observation that the Discrete Fourier Decomposition considers as ‘slower’ in terms of computational speed (O(N²)) compared to its FFT counterpart (O(NlogN)), but again, the elaboration of one's view functionally depends upon one's needs and goals, whether it be for an electrodynamic engineer or a market trader, in the way that Nietzsche proclaims that every fact is interpretably multi-perspectival (“𝘈𝘨𝘢𝘪𝘯𝘴𝘵 𝘵𝘩𝘢𝘵 𝘱𝘰𝘴𝘪𝘵𝘪𝘷𝘪𝘴𝘮 𝘸𝘩𝘪𝘤𝘩 𝘴𝘵𝘰𝘱𝘴 𝘣𝘦𝘧𝘰𝘳𝘦 𝘱𝘩𝘦𝘯𝘰𝘮𝘦𝘯𝘢, 𝘴𝘢𝘺𝘪𝘯𝘨 ‘𝘵𝘩𝘦𝘳𝘦 𝘢𝘳𝘦 𝘰𝘯𝘭𝘺 𝘧𝘢𝘤𝘵𝘴’, 𝘐 𝘴𝘩𝘰𝘶𝘭𝘥 𝘴𝘢𝘺 : 𝘯𝘰, 𝘪𝘵 𝘪𝘴 𝘱𝘳𝘦𝘤𝘪𝘴𝘦𝘭𝘺 𝘧𝘢𝘤𝘵𝘴 𝘵𝘩𝘢𝘵 𝘥𝘰 𝘯𝘰𝘵 𝘦𝘹𝘪𝘴𝘵, 𝘰𝘯𝘭𝘺 𝘪𝘯𝘵𝘦𝘳𝘱𝘳𝘦𝘵𝘢𝘵𝘪𝘰𝘯𝘴.” 𝘛𝘩𝘦 𝘗𝘰𝘳𝘵𝘢𝘣𝘭𝘦 𝘕𝘪𝘦𝘵𝘻𝘴𝘤𝘩𝘦 (1954) by Walter Kaufmann, p. 458). The DFT preserves some mathematical characteristics that the FFT overlooks, one of which is that the DFT permits to calculate the transform for any arbitrary frequency, not just predetermined fixed bins, and thus the sustaining of the 𝒏𝒐𝒏𝒖𝒏𝒊𝒇𝒐𝒓𝒎 power spectrum (peaks occuring at non-integer multiples of the FFT bin width) which the FFT outcasts due to its dyadic decomposition structure (Cooley-Tukey, etc.,) In either case, however, financial time series are statistically prone to be highly nonstationary (see for example Fama & French (1998) “𝘗𝘦𝘳𝘮𝘢𝘯𝘦𝘯𝘵 𝘢𝘯𝘥 𝘛𝘦𝘮𝘱𝘰𝘳𝘢𝘳𝘺 𝘊𝘰𝘮𝘱𝘰𝘯𝘦𝘯𝘵𝘴 𝘰𝘧 𝘚𝘵𝘰𝘤𝘬 𝘗𝘳𝘪𝘤𝘦𝘴” wherein they explore how stock prices have a nonstationary ‘permanent’ component (random walk) and a ‘temporary’ stationary component. See also Schmidt et al. (2013/2025) : “𝘕𝘰𝘯𝘴𝘵𝘢𝘵𝘪𝘰𝘯𝘢𝘳𝘪𝘵𝘺 𝘪𝘯 𝘍𝘪𝘯𝘢𝘯𝘤𝘪𝘢𝘭 𝘛𝘪𝘮𝘦 𝘚𝘦𝘳𝘪𝘦𝘴 𝘢𝘯𝘥 𝘎𝘦𝘯𝘦𝘳𝘪𝘤 𝘍𝘦𝘢𝘵𝘶𝘳𝘦𝘴.” Wherein they argue that standard statistical mechanics fails in finance because sample-averaged observables (like the mean and variance computed in the 𝒘𝒊𝒏𝒅𝒐𝒘_𝒔𝒖𝒎 function you had provided concerning this DFT source code proper) depend entirely on the time window chosen – which, naturally, changes over time as the former progressively and regressively scales up and down. As is the case for the fixed 𝒏𝒐𝒓𝒎𝒂𝒍𝒊𝒛𝒆𝒅_𝒂𝒎𝒑𝒍𝒊𝒕𝒖𝒅𝒆 output function, in regard to which volatility (variance) changes over time, or that the 𝒘𝒊𝒏𝒅𝒐𝒘_𝒎𝒆𝒂𝒏 output function shall lag significantly during a ‘stochastic jump' in data, or the DFT might output cycles that are actually mere outlier noise. I recommend as a starter to read these three documents, for I think that the Fourier Transform (or any spectral decomposition algorithm for that matter) assumes a specific role within this intricate mesh of mathematically reality, and it returns for our Intellect to assign the boundaries of that role proper.

Multiscale Fast Fourier Transform homogenization of additively manufactured fiber reinforced composites from component-wise description of morphology (2023) 

Efficient Simultaneous Multi-Scale Computation of FFTs (Dave Cohen, Camille Goudeseune, and Mark Hasegawa-Johnson, 2009)

Fractional Brownian Motions in Financial Models, Simulation and Pricing (Chun Ming Tam, 2015)
If I'm going to continue reading this thread, I'm going to need a different kind of mushrooms with dinner. This is like philosophy meets data science with abstract art as a side piece.😵
 
Chiheb Azouz:

create a blog and article. mq may even find your arguements worthy of being posted on other language forums of the site. This is a forum for questions regarding mt4/5 and specific issues with such. arguing for and against how data is used and if it should or shouldnt be used is both subjective and not up to you if a trader uses it or not -- nor if they should alter the original strategy used to get that data. whatever strategy works, just works, no matter your arguements for or against the theory used behind their indicators or stragegy.

This is not the medium or site for such discussions. Your statements and discussion are better suited in the other 2 sites.

I am not criticising your arguement, only saying that this is not a forum for such things. The articles site and the blog site is where the arguements and statements that you made -- should stay.

However if you have a more specific question, then, please rephrase your question in a much smaller paragraph, and without quoting some who lived in the 16th century.


Ryan L Johnson #:
If I'm going to continue reading this thread, I'm going to need a different kind of mushrooms with dinner. This is like philosophy meets data science with abstract art as a side piece.😵

"hiccup" -- i took a few shots before and after reading it.