首页 > 其他分享 >!!!!深入理解离散傅里叶变换(DFT)

!!!!深入理解离散傅里叶变换(DFT)

时间:2022-12-15 11:34:18浏览次数:67  
标签:index 13 15 14 DFT Value 离散 14j 傅里叶

前面写过一篇傅里叶变换的文章:

https://zhuanlan.zhihu.com/p/66117227

但是在工程应用中,得益于数字技术的应用,绝大多数傅里叶变换的应用都是采用离散傅里叶变换(DFT),更确切的说,是它的快速算法FFT。这篇文章再来写写有关离散傅里叶变换的关键点。

闲言少叙,直入主题。先把DFT的式子写在这里:

 

 

 

对离散傅里叶变换变换,我认为最重要的是搞清楚两点:

 

 

 

 

 

 

 

 



最后结合一个实例来加深上述对离散傅里叶变换的结果的理解。

假设原始信号:

 

,即40Hz和90Hz两个正弦信号的叠加。来看看不同采样过程得到的DFT(FFT)结果。

1、在0~0.5s内采样500点得到

,即采样频率为

Sample rate:  1000.0 Hz
Samples size:  500
Value at index 0:       (-4.032691340321753e-14+0j)
Value at index 1:       (5.47331447460053e-14-2.3037127760972e-15j)
Value at index 499:     (5.4687509286977934e-14-6.571937883535803e-17j)
Value at index 2:       (7.420972054685959e-14+2.728368124174801e-14j)
Value at index 498:     (7.420972054685959e-14-2.7283681241748016e-14j)
Value at index 3:       (8.290505399586863e-14-7.652212197228891e-14j)
Value at index 497:     (8.358757983676411e-14+7.495170765863399e-14j)
Value at index 4:       (1.1677769275717452e-13+1.3442578178785925e-13j)
Value at index 496:     (1.1677769275717452e-13-1.3442578178785925e-13j)
Value at index 17:      (-7.504531301157937e-14+1.2723135031031761e-14j)
Value at index 483:     (-7.672049076101245e-14-1.500632837960154e-14j)
Value at index 18:      (-7.891238745560246e-15+6.78774432759792e-14j)
Value at index 482:     (-7.89123874556022e-15-6.787744327597919e-14j)
Value at index 19:      (-2.5937242233361435e-14+2.4704194714533472e-14j)
Value at index 481:     (-2.4853927632423113e-14-2.2714218335194713e-14j)
Value at index 20:      (-2.5969976162903077e-13-249.99999999999997j)
Value at index 480:     (-2.7020071315753276e-13+249.99999999999994j)
Value at index 21:      (-1.0514300973234786e-13-1.0006753400743443e-13j)
Value at index 479:     (-1.0455557375562673e-13+9.966632961798892e-14j)
Value at index 22:      (4.27052642300047e-14-7.30154103729329e-14j)
Value at index 478:     (4.2705264230004636e-14+7.301541037293293e-14j)
Value at index 23:      (-7.27269879933383e-14-2.603104120782442e-15j)
Value at index 477:     (-7.12935306401559e-14-1.3354376259863072e-15j)
Value at index 42:      (-3.1818051957819556e-13+1.1687262563911411e-13j)
Value at index 458:     (-3.1818051957819556e-13-1.168726256391141e-13j)
Value at index 43:      (-5.450564689264658e-14-1.1722764422956276e-13j)
Value at index 457:     (-5.314929135111624e-14+1.1522778100403543e-13j)
Value at index 44:      (7.675296751851385e-15+5.02082399689537e-14j)
Value at index 456:     (7.675296751851395e-15-5.02082399689537e-14j)
Value at index 45:      (-1.8656042783807358e-13-125.00000000000001j)
Value at index 455:     (-1.8283507960060743e-13+125j)
Value at index 46:      (2.9414340906931243e-14-2.8748428671670603e-15j)
Value at index 454:     (2.9414340906931256e-14+2.8748428671670793e-15j)
Value at index 47:      (-1.5607886375281088e-13+1.515599634511694e-13j)
Value at index 453:     (-1.5622888838206788e-13-1.5231699611378945e-13j)
Value at index 48:      (-3.336736633940097e-13-1.8468226362166973e-13j)
Value at index 452:     (-3.336736633940097e-13+1.846822636216697e-13j)

根据

,40Hz对应的点应该是40x500/1000 = 20,90Hz对应的点是90x500/1000 = 45。仔细看上面的结果,峰值就出现在20和45这两个X[k]上,与上面的结果是相符的。

2、在0~0.5s内采样200点。即采样频率为

Sample rate:  400.0 Hz
Samples size:  200
Value at index 0:       (-9.962313679719496e-15+0j)
Value at index 1:       (1.5375738523043498e-14-1.7763568394002505e-14j)
Value at index 199:     (1.5364526984141232e-14+1.8053874443897092e-14j)
Value at index 2:       (2.4360703442280214e-14+2.0532966203745806e-14j)
Value at index 198:     (2.4360703442280205e-14-2.053296620374579e-14j)
Value at index 3:       (2.0288475407009816e-14-2.8199664825478976e-14j)
Value at index 197:     (2.0807016603803593e-14+2.5781191134887997e-14j)
Value at index 4:       (2.3558454288221e-14+5.100011152372421e-14j)
Value at index 196:     (2.4588961133289312e-14-5.06542319400188e-14j)
Value at index 17:      (-4.948158073846615e-14-2.3807485819200347e-15j)
Value at index 183:     (-4.947317469249361e-14+2.3489297141087214e-15j)
Value at index 18:      (7.562557031226013e-15+3.176911129338048e-14j)
Value at index 182:     (7.562557031225999e-15-3.176911129338048e-14j)
Value at index 19:      (-1.6223568883664007e-14+1.01840359973437e-14j)
Value at index 181:     (-1.5700099313549882e-14-1.1498463655640398e-14j)
Value at index 20:      (-1.1820905858067227e-13-99.99999999999997j)
Value at index 180:     (-1.0938604556396042e-13+99.99999999999997j)
Value at index 21:      (-1.1620694421953064e-13-4.751106283246057e-14j)
Value at index 179:     (-1.150784893098272e-13+4.793037226530092e-14j)
Value at index 22:      (1.6099799689737958e-14-1.3271426343635642e-13j)
Value at index 178:     (1.6099799689737958e-14+1.327142634363564e-13j)
Value at index 23:      (3.73441643855716e-14+5.0081735226909336e-15j)
Value at index 177:     (3.687933703385216e-14-4.471625968067038e-15j)
Value at index 42:      (-1.1557943305798278e-13+5.097782982384095e-14j)
Value at index 158:     (-1.1557943305798278e-13-5.097782982384095e-14j)
Value at index 43:      (-1.8613645616263873e-15-3.443876516418567e-14j)
Value at index 157:     (-2.457778623388092e-15+3.637881588990269e-14j)
Value at index 44:      (-6.6130108234333566e-15-3.4073835405877573e-15j)
Value at index 156:     (-5.653045149440138e-15+8.341216716228753e-16j)
Value at index 45:      (-6.686373773508564e-14-50j)
Value at index 155:     (-6.667221489155492e-14+50j)
Value at index 46:      (2.1890070028903165e-14+2.1251334251785863e-15j)
Value at index 154:     (2.1890070028903178e-14-2.1251334251785784e-15j)
Value at index 47:      (-6.374431304487239e-14+6.92901767459968e-14j)
Value at index 153:     (-6.42527926289295e-14-7.005184499916765e-14j)
Value at index 48:      (-1.2160331765013047e-13-6.524217072453415e-14j)
Value at index 152:     (-1.2160331765013047e-13+6.524217072453417e-14j)

同样根据

,40Hz对应的点应该是40x200/400 = 20,90Hz对应的点是90x200/400 = 45。与上面的结果是相符的。

3、在0~0.2s内采样200点。即采样频率还是

Sample rate:  1000.0 Hz
Samples size:  200
Value at index 0:       (-4.3026244412314335e-14+0j)
Value at index 1:       (1.173774221692321e-14-2.2886577640004618e-14j)
Value at index 199:     (1.1737742216923204e-14+2.2886577640004624e-14j)
Value at index 2:       (5.218512159071522e-14-1.0547118733938987e-15j)
Value at index 198:     (5.304654765694656e-14+3.6369492074111756e-16j)
Value at index 3:       (-1.717798331659063e-14+4.185135291056711e-14j)
Value at index 197:     (-1.7177983316590625e-14-4.185135291056711e-14j)
Value at index 4:       (6.752817999010086e-15-7.897865029439045e-15j)
Value at index 196:     (6.752817999010086e-15+7.897865029439045e-15j)
Value at index 5:       (5.46586736889134e-16+3.113498085620893e-14j)
Value at index 195:     (5.46586736889132e-16-3.113498085620893e-14j)
Value at index 6:       (8.664379025409079e-15+2.886579864025407e-14j)
Value at index 194:     (8.122427615151029e-15-2.8218391464038713e-14j)
Value at index 7:       (-3.1583155825707856e-14+3.1187715085989398e-15j)
Value at index 193:     (-3.158315582570786e-14-3.1187715085989386e-15j)
Value at index 8:       (-7.430325360934045e-14-100.00000000000001j)
Value at index 192:     (-7.110724204215482e-14+100.00000000000001j)
Value at index 9:       (-2.6730839326555293e-14-2.0006065544472292e-14j)
Value at index 191:     (-2.6730839326555293e-14+2.0006065544472292e-14j)
Value at index 10:      (-7.856970190857711e-15-4.343435848423759e-14j)
Value at index 190:     (-6.518294109116255e-15+4.7989777969450694e-14j)
Value at index 11:      (9.453254387832055e-15-6.769310470798414e-14j)
Value at index 189:     (9.453254387832046e-15+6.769310470798417e-14j)
Value at index 12:      (5.275563422093338e-14-2.9657914357409356e-14j)
Value at index 188:     (5.275563422093339e-14+2.9657914357409356e-14j)
Value at index 13:      (6.50781429135821e-15-3.123232760292878e-14j)
Value at index 187:     (6.507814291358213e-15+3.1232327602928776e-14j)
Value at index 14:      (8.196767977858568e-14-1.2315620899056779e-14j)
Value at index 186:     (8.147505297942633e-14+1.2222322554793413e-14j)
Value at index 15:      (5.5154619065607806e-14+4.022254205534508e-14j)
Value at index 185:     (5.5154619065607806e-14-4.022254205534508e-14j)
Value at index 16:      (2.4752813406175686e-14+4.0258097738867185e-14j)
Value at index 184:     (2.475281340617569e-14-4.025809773886718e-14j)
Value at index 17:      (-1.0398979154998123e-14+4.1019175485025494e-14j)
Value at index 183:     (-1.039897915499813e-14-4.101917548502549e-14j)
Value at index 18:      (-4.4486770345657527e-14-50.00000000000001j)
Value at index 182:     (-4.328485467007317e-14+50j)
Value at index 19:      (-3.0147506558109117e-14-2.4004792372941166e-14j)
Value at index 181:     (-3.014750655810912e-14+2.4004792372941175e-14j)
Value at index 20:      (4.3617316218763913e-14-4.293074786291573e-14j)
Value at index 180:     (4.3617316218763913e-14+4.2930747862915725e-14j)

同样根据

,40Hz对应的点应该是40x200/1000 = 8,90Hz对应的点是90x200/1000 = 18。与上面的结果是相符的。

文章链接    https://zhuanlan.zhihu.com/p/71582795

标签:index,13,15,14,DFT,Value,离散,14j,傅里叶
From: https://www.cnblogs.com/fpga-notes/p/16984572.html

相关文章

  • 离散数学--关键leeds笔记
    TypesofProofdirectproofProveα→βby:assumeαistrueusethistoderiveβ.indirectproofaproofthatisnotdirect;includes:proofbycontraposition......
  • 离散数学: 集合关系
    课时4:集合与关系(上)_哔哩哔哩_bilibili       一些栗子:    下面那个直接打开就行了    栗子: ......
  • 【图像处理笔记】傅里叶变换
    【图像处理笔记】总目录0引言在之前的博客图像增强,傅里叶变换(OpenCV)中都有用到过傅里叶变换,但一直都不是特别理解,现系统地学习一下。先来看一个视频傅里叶级数与傅立叶......
  • 只做了傅里叶变换是0.69
    importmatplotlib.pyplotaspltimportnumpyasnpimportpandasaspdimportseabornassnsdf=pd.read_csv('train.csv')df=df.drop(['ID'],axis=1)df=df.to_......
  • 快速傅里叶变换的 numpy 实现
    目录理论实现numpy实现的要点reshape()和swapaxes()类型转换、一维化和坐标转换数组的复制理论现有长度为\(n=2^s\)序列和相应的多项式为\[\begin{align*}\pmba&......
  • 李永乐 傅里叶变换
    傅里叶变换,表示能将满足一定条件的某个函数表示成三角函数(正弦和/或余弦函数)或者它们的积分的线性组合。B站首发!草履虫都能看懂的【傅里叶变换】讲解,清华大学李永乐老师......
  • 离散对数&BSGS学习笔记
    离散对数定义求\(k\)使得\(a^k\equivn\pmodp\),称\(n\)在模\(p\)意义下以\(a\)为底的对数是\(k\)。如何求离散对数BSGS(BabyStep,GiantStep)大步小步算......
  • 离散化
    离散化vector<int>alls;//存储所有待离散化的值sort(alls.begin(),alls.end());//将所有值排序alls.erase(unique(alls.begin(),alls.end()),alls.end());//......
  • 时序数据EMD,VMD,小波,傅里叶分解
    EMD分解:https://www.bilibili.com/video/BV1iu411f7m7/?spm_id_from=333.337.search-card.all.click&vd_source=bdbb5be671dc7bcfadbe64bf3d0d2f95VMD分解:https://www.bi......
  • 离散数学左孝凌-图论1
    图的基本概念路与回路......