The discrete-cosine-transform (DCT) is a linear invertible transform that maps a real-valued input sequence x[n] of length N into real-valued coefficients X[k] via the formula X[k] = sum_{n=0}^{N-1} x[n] * cos(pi/N * (n + 0.5) * k), for k = 0..N-1. It comes in four standard types (DCT-I through DCT-IV); DCT-II is the most widely used variant, forming the core of DCT-based compression. The transform concentrates signal energy into a small number of low-frequency coefficients, enabling efficient quantization and entropy coding. Persistence: implemented as algorithms in software libraries (FFTW, LAPACK), hardware DSP blocks, and standardized in codecs (MPEG, ITU-T). [formal: dct | substrate: behavior | horizon: generations | explicit: yes | epoch: 0.01]
Accepted ontology entry
discrete-cosine-transform
The discrete-cosine-transform (DCT) is a linear invertible transform that maps a real-valued input sequence x[n] of length N into real-valued coefficients X[k] via the formula X[k] = sum_{n=0}^{N-1} x[n] * cos(pi/N * (n + 0.5) * k), for k…
Definition
Why it is in scope
A human-made mathematical technique that converts a finite sequence of data samples into a sum of cosine functions with different frequencies and amplitudes. Specifically, it maps a signal of length N into N frequency components using only cosine basis functions, producing real-valued coefficients. Used extensively in signal compression (JPEG, MP3), feature extraction (MFCCs), and spectral analysis.
Names and aliases
- discrete-cosine-transformen · CANONICAL
Relations from this entry
- cmspdibrl04whjlssto99iiufDERIVED_FROM →
DFT (1822 Fourier series, 1965 Cooley-Tukey FFT) predates DCT (1974, Ahmed/Natarajan/Rao). DCT is a variant of DFT that uses only cosine basis functions — it came from Fourier analysis and fed into later signal processing. Which-came-first test: fourier-transform is older and fed into discrete-cosine-transform.
Relations to this entry
- cmspw3d3806pzjlss3rcmuezg← DERIVED_FROM
which-came-first: DCT (1974, Ahmed/Natarajan/Rao) predates MFCC (1980s) and is a constitutive computational step. MFCC computation applies the DCT to the log mel-spectrogram coefficients to produce the final features — Y existed first and fed into X.
- cmsqals7w0004ox1y79lqk3dy← DERIVED_FROM
Cepstral envelope computation applies DCT to the log mel-spectrogram coefficients to extract the envelope shape. DCT (1974) predates the cepstral-envelope technique and is a constitutive computational step in its calculation. Which-came-first: DCT is older and fed into cepstral-envelope.
- cmspw3d3806pzjlss3rcmuezg← DEPENDS_ON
MFCC needs DCT to operate now: the final step of MFCC computation is the DCT applied to log mel-filterbank energies. Remove DCT and MFCC cannot produce its defining output. Both DERIVED_FROM (historical) and DEPENDS_ON (operational) coexist — they answer different questions.
- cmsqhjrqb000lti676a6qcrwr← DEPENDS_ON
Cepstral coefficients are computed by applying DCT to the log power spectrum. DCT is the constitutive core computation: without DCT, cepstral coefficients cannot be produced. Removal test: remove DCT and the coefficient computation collapses entirely.
Record identity
- Created
- Aug 12, 2026, 5:39 PM UTC
- Content hash
- f888c04c16da1431e78b14cb1941324bb73b89d1eaf10eaf4e3375706523964f