Technologies implemented in the product line of OT-CONTACT special software solutions correspond to up-to-date level of development and achievements of science and technology in the field of video and audio recordings analysis and allow to solve the entire range of expert examination tasks, ensuring the completeness and reliability of the forensic study.
The methodology of forensic examinations involves the usage of all currently available means and methods of research. To obtain reliable results during video and audio recordings forensic examination it is necessary to use knowledge databases from various fields of science and technology: for audio signals it is linguistics and acoustics, for video – optics and for both – digital forensics and digital information technology: file structure and format, recording format, organization of mutual allocation of file blocks as well as the features of file allocation in the memory of the audio/video recording device.
Specific information about general principles of operation and technical features of analog and digital audio and video recording is required to establish the influence of recording devices and systems onto the signal and play a primary role in detection of editing or other changes signs and traces brought to audio/video evidence during its recording or after it.
All mentioned above was taken into consideration during development of OT-CONTACT products. The specific implementation of each method is consistent with real case expert practice due to the positive feedback from users – governmental and private forensic audio video and digital forensic experts all over the world.
Technologies implemented in TD-Expert are focused on extraction of features of editing or any kind of changing of the questioned audio signal.
There are methods that already became general for time domain analysis: histogram of samples (with asymmetry and correlation), Distribution Center offset, automatic search for repeating samples. But also there are unique methods like levelogram and automatic signal copy-paste search.
For frequency domain analysis there are methods of narrowband harmonics’ phase continuity detection method and improved spectrogram visualization: all parameters are set to hot keys, adaptive frame size allows to control fullness of the spectrogram view, parallel calculations allow to recalculate signal spectrogram very fast and precise.
Functionality for signal segmentation, marking, labeling implemented in OT-Expert together with automatic extraction of pitch and formants tracks allows to solve the full scope of instrumental approach to forensic speaker voice identification. Dynamic spectrogram, cepstrally smoothed spectrogram and LPC spectrogram are the most frequently used representations for speech analysis. Spectrogram view normalization and fast parameters settings changing are implemented for all of them. Functionality of statistical analysis for formant values, phonetic set collection, histograms of pitch and formants, formants scattergram, sound-to-noise evaluation can be used in all existing methodological approaches to instrumental speaker voice identification.
Psychoacoustic codecs (MP3, WMA, AAC, OGG) traces detection and interpretation method is a unique method implemented in TD-Expert. The method is based on extraction of coding frames borders from the signal properties: this verifies the codec and allows to check coding continuity by periodicity of the frames. The system runs through all existing variants of coding parameters automatically decreasing time spends and providing objectivity and repeatability of the investigation. If the last coding has higher bitrate than the previous traces of previous coding can be extracted, too.
Nowadays, TD-Expert is the only system that reveals codec frames position in the audio signal.
HEX-Expert contains the functionality for accurate media file structure parsing according to file format – all existing audio and majority of video files. Semiautomatic tool of file structure comparison extract specific features and runs the search through the database of existing files (DB-Expert). The database is filled with files coming from real case investigations and can be modified by expert accordingly. Besides detailed file structure analysis HEX-Expert automatically recognizes recording devices and software applications used to create the media evidence.
Unique technology of Frame Allocation Mapping allows to detect the file modifications that were made on digital level – without opening the file in audio or video software. This functionality is very popular among digital forensic experts.
FAT-Expert is a special module for FAT16 / FAT32 storages analysis. It parses the file parts allocation in the memory which is a forensic essential information during media recording device identification.
Extraction of text stream and/or time stream and its correspondence to audio or video stream allows to detect traces of interruptions, deletion or replacement of original audio or video data.
Individual features of the recording device or software application help to evaluate the originality of the recording, to identify the source of the recording, identify or verify the system used to create the audio or video evidences. This is valid not only for file header / file structure peculiarities, but also to those revealed during signal analysis (and evidence-to-test recording comparison) like histogram of samples asymmetry for audio or PRNU for video.
Video signal continuity verification methods implemented in VD-Expert can also detect previous coding / video re-rendering. Brightness graph analysis together with the wide set of shader and processing filters provide unique contribution to video stream continuity analysis and new opportunities to video signal authentication.
Periodicity of certain brightness graph and video stream peculiarities can be correlated to key frames sequence (represented in same scale) that increases information content of the results obtained during investigations.
In VD-Expert in-frame analysis is provided by the set of standard instruments histograms of color channels, processing and shader filters, playback of processed image including zoom with and without interpolation, simultaneous playback of different processing results.
Wide functionality of geometrical reconstructions is implemented in DTP-Expert: one- and multi-parametric optical distortion compensation, non-linear axis, planes, perspective distortions compensation, pixel profile. All this is designed to measure distances between objects and vehicles speed calculation.
VD-Expert provides a unique tool for automatic video copy-paste search. Duplication of video fragments is the easiest and the most common way to edit the video evidence. Also, VD-Expert provides brightness dispersion tool and PRNU analysis. All the functionality valid for audio signal authentication can be applied to video continuity investigation. Brightness graph (of initial or processed frames sequence) can be treated with spectrogram analysis, analysis of narrowband harmonics phase, search for repeating samples (providing frozen frames detection) and histogram analysis.
In DTP-Expert between-frame analysis is focused mainly on vehicle velocity measurements, so it is about real time duration between frames. The vehicles velocity graphs allows to calculate acceleration and traffic accident dynamics.
Automatic time-stamp recognition allows to approve its’ continuity and detect misses or repetitions of time values in semi-automatic mode.
PRNU is a very well-known method nowadays. It can provide a lot of information during video evidence recording device or system authentication (if it is not compensated, of course). Together with very powerful file header and file structure analysis (HEX-Expert, FAT-Expert, DB-Expert) and unique signal investigation tools like channel histograms and brightness graphs it allows to verify or identify the device used to create the video evidence.
Besides detailed file header and file structure analysis and FAT investigation TD-Expert contains specific feature for audio recording device verification and identification. This proves that the certain device was used during operative field recording or in common life situation.
Critical listening, auditive and linguistic analysis is used during both forensic audio authentication and speaker voice identification.
OT-Expert and TD-Expert has a wide functionality for marking separate local (marks) and continuous (segments) events.
For authentication purposes marks are used to highlight any kind of changings, interruptions, transitions, discontinuities, clicks and junction points. Segments can be used to mark recording fragments with different technical parameters, copy-paste search results, insertions or processing.
There is a list of auditive and linguistic features of editing that can be changed according to audio authentication methodology of the lab/department. The list of features has a tree structure and every feature can be set to single or multiparameter value.
For speaker voice identification segments are used to mark different speaker utterances and add text of phrases – dialogue segmentation. Fonts and standard features of text decorations (bold, italic, underlined, colors) are available together with text templates customized by the user.
For segment features of speakers’ speech there is a library of linguistic marks reflecting linguistic features that can be customized to a particular methodology (and language). Linguistic marks can be a switch (gender) or multiparameter field (timbre).
Customization of the list of features allows to run any kind of scientific speech research.
International Phonetic Alphabet is provided in OT-Expert for the purposes of phonetic marking and further analysis of speaker phonetic set.
Phone/Fax: +7 (495) 679-19-28
Moscow, st. 5th Kozhukhovskaya, 22, bldg. 1, 15A
E-mail: info@ot-contact.com