What email address or phone number would you like to use to sign in to Docs.com?
If you already have an account that you use with Office or other Microsoft services, enter it here.
Or sign in with:
Signing in allows you to download and like content, and it provides the authors analytical data about your interactions with their content.
Embed code for: FORENSIC AUDIO AUTHENTICATION & SPEAKER IDENTIFICATION
Select a size
FORENSIC AUDIO AUTHENTICATION & SPEAKER IDENTIFICATION
Meaning, Purpose & Scope:
Forensic Audio Authentication is the process to find out whether the audio recording is continuous or whether some of the audio has been deleted, added, altered or changed to the original recording
Forensic Speaker Identification is identifying an individual solely by his/her voice or speech.
Nature of Cases Received: Audio or voice is frequently encountered as evidence in cases such as trap investigation, kidnapping for ransom, threatening calls, anonymous calls, hoax calls, obscene calls, terrorist threats and so on.
Nature of Exhibits Received: CDs, DVDs, Cassettes, Micro-cassette, Mobile phones, Memory cards, Digital recorders or any digital storage media etc.
Human speech is a complex signal, an outcome of the influence of several physiological, psychological and environmental factors. The acoustic signal we produce when we speak is determined by, among others, the physiology of the vocal tract and articulators, childhood language and dialect acquisition, regional traits, and training from life experiences. These features give a distinctive ‘identity” to the speech of different individuals. In order to tell two voices apart, it is necessary to consider characteristic features based on which the voices can be discriminated. When comparing recordings of human speech, it is necessary to actually compare the two voices, minimizing the influence of the external factors that have affected the recordings. Advances in telecommunication and recording technology have meant that an increasing amount of speech is transmitted through telecommunication networks. These transmission networks as well as the recording devices have their influence on the recording of the voice.
Basic Elements of Speech Production:
Human organs that combine to produce intelligible speech are:
The lungs & windpipe
The nasal cavities
SCHEMATIC DIAGRAM SHOWING THE VARIOUS ORGANS RESPONSIBLE FOR INTELLIGIBLE SPEECH
Forensic Audio Authentication:
Physical examination of tape: Physical inspection was done to determine the condition of the tape and to inspect for any damage that would affect how it functions. It includes the amount of tape, condition of tape for any physical damage to reel/cassette or any physical splices
Auditory Analysis: The recordings are examined by critical listening for any transient sounds such as stop, start, pauses which can be the possible sites of tampering or break in the recording. The external events like fan, car starting, other speakers, any abrupt changes in recording level or type or degree of background noise, abrupt changes in voice etc. are carefully observed.
Waveform Analysis: Audio recording is examined in computer using forensic audio software’s such as GoldWave, Cool Edit Pro etc. and the waveform is viewed in terms of a graphical display. The waveform display reflects the relationship between the time and amplitude of recorded sounds and allows the identification and comparison of recorded audio present in test tapes. When a recorded event occurs such as a start, stop, stop/re-start or pause, a distinctive mark or signature is created on the tape.
Abrupt Change in wave form
Electronic Gap showing absence of waveform indicating that the audio is not continuous
Any electronic gap in the audio signal as well as edit signatures can be examined in the spectrographic analysis.
Abrupt changes in the Waveform
Edit Start Signature in between two words
If the edit signature is in between two utterances indicates the conversation is not continuous. The waveform analysis reveals the abrupt changes in the wave amplitudes, electronic gaps etc. if any which are due to editing or tampering or doctoring etc.
Forensic Speaker Identification:
Step-I) Auditory Analysis:
The questioned and standard recordings are critically listened and different characteristics such as speech rate, manner of utterance, speaking mode, stylistic features, respiratory form, pitch level, striking features of voice, fluency, understandability, loudness speech rate etc. are studied.
After listening, the phonetic & linguistics features like dialect, accent, rate of articulation, vocal quality, stress etc. are compared
Step-II) Spectrographic Analysis:
The Spectrograph instrument facilitates the comparison of two voices, one suspected and the other known or standard by means of a combination of words and sentences selected from speeches to arrive at a conclusion whether the suspect is responsible for a particular voice / speech or not.
Before spectrographic analysis to be conducted primary process like segregation of speech sample (separating the voice of an individual) followed by selection of Cue words (words containing maximum number of vowels) is done. These Cue words are then used for spectrographic analysis in which formants frequencies are generated which are said to be unique for an individual.
A voice spectrogram is a three dimensional representation of sound. Horizontal axis represents time, Vertical axis represents frequency and Darkness represents intensity. These three dimensions give information on how the frequency and intensity change with time
The formant frequencies are produced at the time of utterance are said to be unique for an individual. Different vowels have different formant frequencies. The formant frequencies of Que word of questioned and specimen voice samples are compared to find out whether the voice belongs to a particular individual or not.
Questioned Voice Sample Specimen Voice Sample
SIMILARITY OF THE ABOVE TWO SPECTOGRAMS INDICATES THAT THE VOICES WERE SPOKEN BY THE SAME INDIVIDUAL
Questioned Voice Sample Specimen Voice Sample
THE ABOVE SPECTROGRAMS OF TWO DIFFERENT SPEECH SAMPLES ARE DISSIMILAR INDICATING THE VOICES WERE SPOKEN BY TWO DIFFERENT INDIVIDUALS.
After thorough analysis the examiner integrates his findings from both the auditory and spectrographic analysis and opines that whether the voice is of the suspected person or not.
FORENSIC VIDEO AUTHENTIFICATION
Video Cassettes, Video CD’s, Mobile phones, Spy cams are being increasingly used in recent times to record certain events secretly in cases such as receiving bribes by using miniature Cameras, Digital Cameras etc., by the victim / investigative journalists. With the advent of cheap and readily available capture media, the number of digital multimedia files has increased manifold. Editing and manipulating these files can be done easily and seamlessly. In majority of the cases there will not be any other evidence except the Video Cassette/VCD. When these are submitted to the investigating agency the Investigating Officer has to first satisfy himself that the video Cassette/VCD is authenticated and actually depicts a true event / incident.
Video Authentication is the process of verifying whether the digital video is genuine (not been tampered with in anyway). This technology assumes immense importance, as without it video can never be admissible proof in the court of law. Video Authentication confirms the integrity of the video.
Nature of Problems:
Authentication of video
Whether the video recordings in DVD/VCD/CD are edited or not, tampered or not, depicts a single uninterrupted incident or not, the pictures therein have been manipulated or not, various incidents have been joined to form a single recoding or not, etc.
Identification of the persons in the video
In addition to the problem of authentication, it is required to find out whether the person in the video is the same person present in the specimen photograph or not.
Sync of Audio with video
Whether the audio in the questioned VCD is in sync with the video or whether there is any manipulation of either audio or video.
Types of editing
Real time editing
Post production editing
a) Assemble editing – laying one scene after another. Both audio & video editing is done.
b) Insert editing – replacing original video with new video leaving the audio portion intact.
1. Visual Observation: The Video Cassette/CD has to be viewed carefully observing the scenes, background, smears, vertical jumping, horizontal wavering, colour & hue saturation, any discontinuities, jumps, breaks, abrupt increase / decrease in audio background sounds, sudden strange sounds, variation in level of illumination etc.
2. Examination under Waveform Monitor: The signal from Video Cassette/CD is fed to Waveform Monitor, which provides voltage verses time display of the video signal there by evaluating the luminance portion of the signal. If there is any signal aberration, which could be due to editing, the Waveform Monitor displays this signal on the screen. Thus the luminance values of the active video signal are carefully studied for any abrupt changes to identify edit points. A typical output of waveform monitor – peaks showing brightness, minima showing darkness is shown below:
The Luminance values of the active video signal are carefully studied for any abrupt changes for identification of edit points.
3. Examination under Vector Scope: The signal output from the disputed Video Cassette/CD is fed to the Vector Scope, which demodulates and displays the (R-Y) colour difference component on the vertical axis and (B-Y) colour difference component on the horizontal axis. By measuring chrominance information and matching burst of multiple signals, Vector Scope facilities detection of edit points.
4. Frame-by-Frame Study: The Video recording is converted into digital video format as to the hard disc of a computer. Frame-by-Frame study of the digital video format can then be studied by using appropriate software such as Dazzle, Virtual Dub etc., to observe any edit points.
Frame-by-Frame Study of a Video Clip
5. Observation of Lip Sync: The audio & video is carefully listened & watched from the beginning to the end. The lip movement of the persons in the video with respect to the audio is carefully observed to ascertain whether they are perfectly matching or not. The time delay of audio to video if any can also be measured to establish editing.
Observation of Lip movements and audio
6. Study of time Code: Study of time code (an electric signal which tracks individual frames for the purpose of editing) for its continuity frame by frame will also yield edit points.
Simulated Studies: Extensive simulated video recording, editing has to be carried out if the original recording devices such as video camera or camcorder are recovered for the study of any color aberrations (or) any other peculiar characteristics.
* Dr. TSN Murthy, Director ,Technical Division,Truth Labs, Hyderabad
Former Joint Director, APFSL, Hyderabad, Former Director, Arunanchal FSL, Former Consultant Forensic Expert to the Govt. of Goa.
Change in waveform
ot been tampered with in anyway). This technology assumes immense importance, as without it video can never be admissible proof in the court of law. Video Authentication confirms the integrity of the video.