DOCTYPE

Although the flexibility of XML is of great advantage, it is often also necessary to define exactly how an XML document can be structured. For this purpose, a data definition language is used. There are tools for validating HTML against a DOCTYPE and XML against an XML Schema. In STx, we have our own tool which performs basic validation. We have called it the DOCTYPE as well.

You can define an STx DOCTYPE using the SET xmlfile DOCTYPE command. Once an XML shell file has a DOCTYPE associated with it, operations which are invalid according to the DOCTYPE will in most cases fail. If you need to validate a complete document, you can use the command SET xmlfile VALIDATE.

In order to guarantee data integrity in the STx DataSet, a DOCTYPE has been defined in the stxconfig.xml file and is associated with the DataSet when it is loaded.

The STXDataSet DOCTYPE

STx includes a DOCTYPE for the segment metadata and project XML structure. It is defined in the stxconfig.xml file and associated with the project when the project is opened in STx. Here is a summary of the defined elements and their relationship to each other.

AFile : ASet

The AFile element is associated with a sound file. It is derived from the ASet element. In addition to the ASet attributes, it has the following attributes:

File - the absolute path to the sound file.

ASet

The ASet element is short for 'audio set'. It is derived from the Set element. In addition to the Set attributes, it has the following attributes:

• SR - the sampling rate.
• CH - the number of channels.

In addition to the elements a Set may contain, an ASet may contain one or more of the following elements:

ASeg elements

An ASet may not be contained within an ASet. The ASet itself is not directly used in the DataSet. Rather, it's derived elements (e.g. AFile and ASequence) are used.

ASeg

The ASeg element is short for 'audio segment' and stores the data specifying and describing an audio segment. The ASeg element has the following attributes:

ID - a string uniquely identifying the segment within the parent element. P - an integer specifying the beginning of the segment in samples, as an offset from the beginning of the file. L - an integer specifying the length of the segments in samples. CH - an integer specifying which channel this segment addresses.

The following elements may be contained in an ASeg element:

APar elements

APar

The APar element contains parameters calculated for a specific audio segment. An APar must be the child of an ASeg element. The APar element has the following attributes:

type

- the type of method. The following types exist:

F0 - Fundamental Frequency

RMS - Signal Energy

Phase - Phase Spectrum

ASpec - Amplitude Spectrum

Fof - Formant Frequencies

Cep - Cepstrum

LAT - Log attack time

CLPC - LPC coefficient

TFA - Amplitude Spectrogram

TFP - Phase Spectrogram

method

the method. The following methods exist:

FO - Formants

RMSB - Frequency Band Energy

RMS - Signal Energy (RMS)

F0A - Fundamental Frequency (SIFT)

F0B - Fundamental Frequency (autocorr.)

F0H - Fundamental Frequency (harmonic grid)

ALPHA - LPC - Error Energy

IRRSP - Irrelevance Amplitude Spectrum

IRRTH - Irrelevance Threshhold

FFT - FFT Amplitude Spectrum

FFTP - FFT Phase Spectrum

CEPST - Cepstrum Smoothed Spectrum

LOFAR - Spectrum - Cepstrum

WAVE - Wavelet Amplitude Spectrum

WAVEP - Wavelet Phase Spectrum

STFT - FFT Amplitude Spectrogram

STFTP - FFT Phase Spectrogram

STWT - Wavelet Amplitude Spectrogram

STWTP - Wavelet Phase Spectrogram

LPC - LPC Smoothed Spectrum

LPAI - LPC Inverse Filter Coefficients

LPRC - LPC Reflection Coefficients

LPAR - LPC Area Coefficients

nx

the number of x axis

ny

the number of y data vectors

CH

the channel this parameter was calculated for.

lfrm

the frame length in milliseconds

sfrm

the frame shift in milliseconds

nfrm

the number of frames

date

the date the parameter was saved

time

the time the parameter was saved