SoundMap Time Frequency Analysis

Cumulative Spectral Decay (CSD) – also known as Waterfall Plots

Cumulative Spectral Decay is the traditional tool for impulse response analysis of loudspeakers. It calculates the “ringing” of the loudspeaker for each frequency using the impulse response. Data can be output in a variety of formats including the widely-used three-dimensional ‘waterfall plots’.

Short Time Fourier Transform (STFT)

The Short Time Fourier Transform is a general purpose algorithm which enables observation of the spectral changes of a signal over time. This method is ideal for the detection of manufacturing defects such as loose particles and Rub & Buzz in loudspeakers, measurement of settling time and ringing in devices including loudspeakers and telephones, and analysis of dropouts, discontinuities and instabilities in digital devices.

Wavelet

Wavelet analysis differs from CSD and STFT analysis in that it uses constant percentage bandwidth rather than constant frequency bandwidth. This offers better time resolution at high frequencies and better frequency resolution at the lower end of the spectrum. This is advantageous as it is more psychoacoustically significant and it is easy to see the entire 20Hz – 20kHz spectrum in one picture. Applications for wavelet analysis are generally the same as for STFT analysis described above; the algorithm selected depends on whether constant frequency or constant percentage bandwidth is preferred.

Wigner-Ville

Wigner-Ville is the ultimate algorithm for detailed analysis of very short events, for example fine analysis of transients or in-depth observation of rapidly evolving signals. Offering an output resolution of one spectrum per sample, it is the maximum achievable precision. It complements the more commonly used analysis methods discussed above.

Time View

• Input File Data format
  • SoundCheck Waveform
  • Wave file (*.wav)
  • TIM: MLSSA binary file format for time data
  • Text files (documented format)
• Display of time waveform + zoom
• Analysis Parameters
  • Choice of the transform: STFT, CSD, WV, Wavelet (see algorithms)
  • Frequency resolution: Hz (STFT, CSD, WV) or 1/3…1/24 octave (Wavelet)
  • Frequency resolution (in Hz)
  • Time increment (in s)
  • WV Smoothing size (in s)
  • Automatic / manual setup
• Information
  • Sampling frequency (Hz)
  • Transform Overlap (%)
  • Window/Wavelet duration (s)
  • BT product (%)
  • Analysis duration (s)
  • Analysis completeness ( “Gap Free”):
  • Number of spectra, lines/spectra, time-frequency pixels
  • Bar-graphs for windows size, smoothing size and analysis duration

Multispectrum View

• Time-frequency intensity map with color coded levels and 2 side displays
• Intensity Map (time-frequency-level)
  • Color coded levels
  • Unit = calibrated power spectral density in dB(ref. 1 U²/Hz)
  • Choice of color scales: Rainbow, Fire, Sunset, Grey Scale, etc.
  • Choice of Dynamic Range
  • 2D Max search
  • 2 cross-cursors (one absolute, one delta)
• Right Side Display= Frequency Display
  • Global Energy Spectrum (U².s/Hz)
  • Instantaneous spectrum (U²/Hz)= frequency slice at time location of cursor 1
  • Partial spectrum (U².s/Hz) between the 2 vertical cursors on the map
  • Group Delay curve(s)
  • 2 horizontal cursors coupled with the ones on the map.
• Bottom Display= Time Display
  • Time wave, analyzed segment (U)
  • Time slice (U²/Hz). The time slice is done at frequency location of cursor 1
  • Energy time curve (U²)
  • Partial Energy time curve (U²) between the 2 horizontal cursors on the map
  • Instantaneous Frequency (Hz)
  • 2 vertical cursors coupled with the cursors of the map
• Analysis Parameters and Informations + Bargraphs as in the analysis front panel
• Physical Unit U
• Energy Display: Total or Partial (defined by cursors)
• Save/recall map file
• Save jpg/bmp images

3D view (Color Waterfall)

• Log/Lin freq axis
• Color scale or Surface
• Full choice of view angle
• 3D Cursor
• Save jpg/bmp images

Algorithms

Transforms:

• STFT: Short-Term Fourier Transform with Gaussian window.
• WV: Smoothed Pseudo Wigner-Ville Transform with 2 independent Gaussian windows. Implemented in analytical form.
• CSD: Cumulative Spectral Decay with Cosine-Tapered (Tukey) Window. Low frequency limit inverse of window duration.
• Wavelet: Gaussian Analytical wavelet (Morlet). Fractional octave bandwidth and center frequencies.

Multispectrum:

• Global Energy Spectrum†: power sum of all spectrum of the multispectrum.
• Sub-Total Spectrum†: sum of all spectrum of the multispectrum, between the 2 verticals cursors.
• Energy Calculus†: sum of all time-frequency values of the whole multispectrum for the total energy, or in the box delimited by the 2 cross-cursors, for the partial energy.
• Energy-Time Curve: sum of all time slices of the multispectrum.
• Partial Energy Time Curve: sum of all time slices of the multispectrum, between the 2 horizontals cursors.
• Group Delay Curve†: centroid of each time slice over frequency.
• Instantaneous Frequency Curve: centroid of each frequency slice over time.

The functions denoted by † are only available when Analysis is complete (Gap Free).

There are 2 SoundMap versions available. Part# 1300 is the full version which includes four analysis algorithms: Cumulative Spectral Decay, Short Time Fourier Transform, Wavelet and Wigner-Ville. Part# 1301 contains the Cumulative Spectral Decay algorithm only. This makes it an ideal replacement for MLSSA’s popular waterfall feature.

Part# 1300 SoundMap Time Frequency Analysis (CSD, Short Time Fourier, Wigner-Ville, Wavelet)
Part# 1301 SoundMap CSD (Cumulative Spectral Decay) analysis only

SoundMap is available as a module for your SoundCheck system, or as a stand-alone product.

Please contact sales@listeninc.com for pricing information.