Fingerprinting the Human Voice

FINGERPRINTS: If you look at your fingertips, you will see a pattern of whorls and ridges on the surface of the skin. It has been known for a long time that the size and pattern of these whorls and ridges are pretty much unique to each individual. Since the human skin exudes sweat and oils that leave a mark on suitably dry and flat surfaces touched by our fingers, it was realized by criminal investigators how useful that might be. For example, when one grasps a gun, fingerprints are left on its surface. When "dusted" with a fine powder, the pattern left by physical fingerprints can be photographed and compared with a known or control specimen, taken for example after a suspect in a crime is arrested. Pleae note well that when people talk about fingerprint evidence being used against a suspect in a criminal case, they are talking about PHOTOGRAPHS of the ORIGINAL marks left on some piece of physical evidence, like a pistol used in the commission of a crime. Cases are often solved and closed by finding the fingerprints of a suspect on the murder weapon. However, it is not only the perpetrators of a crime that can be linked to it by fingerprint evidence but even its victims...

PARABLE OF THE TAXI DRIVER: Once upon a time there lived in the poor Archipelago an evil COUNTERFEITER (Mr. Faker), who often used a TAXI OPERATOR (Mr. Tsuper) to help him manufacture and distribute bogus P1,000 bills. One day Mr. Faker decides he needs a resupply of his favorite brain-sizzler so he sends Mr. Tsuper on an errand to a DRUG DEALER, Mr. Triplehead, to pick up some drugs. Mr. Triplehead and Mr. Tsuper do not know each other, and in fact, Mr. Tsuper does not know the real business of Mr. Tripledead. Mr. Faker tells Mr. Tsuper to pick up a sealed package from Mr. Tripledead and to give him 1000 of their fake P1,000 bills in payment (It's a million peso drug deal!) So off the Mr. Tsuper goes to Mr. Triplehead's lair. He gives the counterfeit bills to Mr. Triplehead and leaves with the package which he gives to Mr. Faker. Later on Mr. Triplehead goes on a drinking binge in a bar full of his old friends from the PNP (you see Mr. Triplehead used to be on the side of the law and was a policeman at one time, a typical happenstance). But being a drunk , a lecher and a drug addict, Mr. Triplehead gets into a fight and ends up busted when the cops find a lot of shabu on his person. They raid his house and find the counterfeit money, which because of the circumstances, contain only three sets of fingerprints, Mr. Triplehead, Mr. Tsuper, and Mr. Faker, all of whom were actually law abiding policemen at one time, with complete fingerprint records available at Western Police HQ. All are rounded up and accused of drug-dealing and counterfeiting. At first, they all stick together knowing it's their neck if the cases against them are proved. It would be reclusion perpetua or a bite from Cleopatra's asp. But something weird happens. Mr. Tsuper realizes he can actually escape the electric chair because he is guilty only of counterfeiting since he didn't even know of Mr. Triplehead's drug business. He jumps bail and disappears, aided in part by Mr. Faker who certainly doesn't want him to become a witness against him for counterfeiting. He hides out for a while, and things quiet down a bit because Mr. Faker is rich and he has real smart lawyers and PR people. But they cannot keep it there. Mr. Faker and Mr.Tsuper turns himself into the police demanding immunity in exchange for his testimony against certain competitors of Mr. Faker, namely XXX, YYY and ZZZ, against whom he claims to have evidence of being counterfeiters. He says he is getting death threats from XXX, YYY, and ZZZ. He says that the bogus bills were not at all produced by Mr. Faker, but rather by Mr. X, Y & Z. He claims he and Mr. Faker have never been in the business of counterfeiting. But Mr. Tsuper does a strange thing. He claims the police must not look at, touch, examine or use the counterfeit bills for any legal purpose because they are illegal tender and should really be treated like radioactive waste. He produces his own set of bogus bills that he claims indicate there are MANY counterfeiters running around, and all the COUNTERFEIT MONEY is all mixed up together. For awhile the Good Cops are confused by all this running around, appearing and reappearing, and the persistent PR people of Mr. Faker get their air time and printed inches. But in the end Mr. Tsuper and Mr. Faker and Mr. Triplehead all go to jail when it is discovered that the only FINGERPRINTS on the bogus bills match their own. But Mr. Tsuper is not found guilty of drug-dealing, becoming indeed the Star Witness in the prosecution and conviction of the other two.

CLIFF NOTES: I'm sorry if you don't get my meaning in the above "parable" -- but I'm not a very good fictionist. So I should perhaps state the obvious parallels with Garci, Gloria and ISAFP, respectively, Mr. Tsuper, Mr. Faker and Mr. Triplehead. "Counterfeiting" is the VOTERIGGING that has been alleged to have been conspired in by the Commissioner and the President. "Drug dealing" is the crime of ILLEGAL WIRE TAPPING, which some allege was done by the President and the Intelligence Services of the Armed Forces of the Philippines, which accusation if true, would entail a prosecution of massive proportions as it constitutes a grave violation of national security and the rights of privacy of government officials, not to mention a prostitution of the Military to perfidious purposes. The Parable has a nice ending because of the FINGERPRINTS on the bogus bills. What are they analogous too? They are analogous to the VOICEPRINTS of the people whose VOICES are in the Garci conversations, that in some manner, found their way onto analog casette tapes first, then as CDROMs and MP3s on WWW. My first awareness of the existence of the Garci Conversations came from Press Secretary Ignacio Bunye himself SIX MONTHS ago. But voiceprints are not fingerprints. And I have gone a long way around just to state the obvious problem facing theDemocratic Opposition: how can it be proved by other than testominial evidence (since Mr. Faker, Tsuper and Triplehead are in omerta) that those voices in these conversations indeed belong to the people that they undeniably sound like to most reasonable people.

VOICEPRINT IDENTIFICATION is a nice article describing the forensic science of deciding whether a given sound sample of unknown origin or provenance matches another of known origin and provenance:

Voiceprint identification can be defined as a combination of both aural (listening) and spectrographic (instrumental) comparison of one or more known voices with an unknown voice for the purpose of identification or elimination. Developed by Bell Laboratories in the late 1940s for military intelligence purposes, the modern-day forensic utilization of the technique did not start until the late 1960s following its adoption by the Michigan State Police. From 1967 until the present, more than 5,000 law enforcement related voice identification cases have been processed by certified voiceprint examiners.

Voice identification has been used in a variety of criminal cases, including murder, rape, extortion, drug smuggling, wagering-gambling investigations, political corruption, money-laundering, tax evasion, burglary, bomb threats, terrorist activities and organized crime activities. It is part of a larger forensic role known as acoustic analyses, which involves tape filtering and enhancement, tape authentication, gunshot acoustics, reconstruction of conversations and the analysis of any other questioned acoustic event.

RELIABILITY OF AUDIO FORENSICS: Some time back fingerprinting was a more advanced science than voice identification technology, but nowadays, there is every indication that audio forensics has attained a high degree of reliability:

Several studies have been published evidencing the ability to reliably identify voices under certain conditions, and a Federal Bureau of Investigation survey of its own performance in the examination of 2,000 forensic cases revealed an error rate of 0.31 percent for false identifications, and 0.53 per cent for false eliminations. (Koenig, B.E., 1986, Spectrographic Voice Identification: a forensic survey, Journal of the Acoustical Society of America, 79:2088-2090.)

And above authoritative study was almost twenty years ago. An important point to note as well, even a single phrase or less, properly compared with a known specimen can contain enough data to establish a positive identification.

RIDGES AND WHORLS OF A VOICE: Like one's face or fingerprints, the voice of a person is unique to him or her, because one's voice is produced by the unique instrument that is one's throat, voice box, tongue, lips, teeth, chest and head. It is as if each of us has the same kind of musical instrument but they are all as different from one another as our bodies are from one another. The sound of each such "voice instrument" can be rigorously and mathematically decomposed in such a way as to develop a VOICEPRINT that can be compared with any purported sample of that voice, say in some stray CDROM raised up into the air by a Presidential spokesperson, to establish that they come from the same vocal instrumentalist. Just like photographs of a known criminal's fingerprint pattern can be compared to photographs purporting to contain his fingerprints, so too can any recording purported to contain a certain person's voice, be compared to a known specimen of that person's voice. The only difference between fingerprint analysis and voiceprint analysis is that in one case the comparison process is VISUAL while in the case of voiceprinting, the process of comparison is QUANTITATIVE or MATHEMATICAL. But the LOGIC of the foresnic method is identical.

VOICE DATA IS A TIME SERIES When a person utters a word, his voice box, mouth and tongue, lips and teeth produce a disturbance in the surrounding atmosphere of air which is composed of a packet of travelling waves, like ripples on the surface of a pond when you throw a stone into it, that arrive in the hearer's ear. When one makes sound recordings with a tape recorder or digital cellphone interceptor, what one is doing is recording in time the instantaneous amplitude and phase of the wave packet as it arrives at the microphone or other collecting device. This amplitude and phase information, when properly digitized, is called a time series representation of the sound, which in the case of a spoken phrase like "Hello Ma'm" or "Yung dagdag yung dagdag" is localized in time to a few seconds. In the ubiquitous DVD genre of entertainment, the sound track consisting of music and dialogue is usually digitized at 48,000 samples per second. So the "time-domain data" for the digital sound recording of a single phrase like "I am the victim of wiretapping" might contain 2 or 3 seconds of data and result in maybe 100,000 digital samples. The beauty of digital voice recordings, such as any modern WIRE TAPPER produces, is that a digital copy of one is IDENTICAL in quality to the original.

MATHEMATICAL PHOTOGRAPH OF A VOICE: Of course, if one were to try and match time domain data one would be met with an intractable mess. Instead what experts do is they convert time domain data sets from the two sets of spoken words being compared, into their FREQUENCY DOMAIN representation. The so-called Fourier Transform of a time domain waveform is a completely equivalent representation of the data, except that instead of time along the x-axis, it contains the "frequency components" of the original wave packet signal. This new, transformed representation is an equivalent of the PHOTGRAPHS taken by police of dusted fingerprint markings on physical objects. A whole FORENSIC science has been built around the need to authenticate voice samples against "control samples" with a known origin and provenance. Let me give an example to clarify some of this jargon. Think of a pianist playing a piano concerto. The "time series" data is the music we hear as we sit listening to the performance. But the "voiceprint" of the concerto is the SHEET MUSIC from which he is reading, which graphically puts down which keys to strike when. Another example is a cake -- thats the "time series data" where the recipe and list of ingredients make up the "elements" of the cake -- that's the "voiceprint". Mathematicians would shoot these analogies full of holes, but there haven't been too many Mathematicians visiting Philippine Commentary lately, so too bad.

AUDIO FORENSICS: I think once it can be proven beyond FORENSIC doubt that that is the voice of Garcillano in the Garci tapes, he can be used as either a WITNESS or a COMPLAINANT in a criminal case against still unknown WIRETAPPERS. What people should not lose sight of is that Garci IS probably a victim of wiretapping, even if you believe he is also the Master Perpetrator of massive VOTERIGGING operation during the 2004 elections. To try and prove cheating happened based on the contents of the tapes themselves seems futile without first establishing that it IS Garci on those tapes.

PROSTITUTING THE ARMED FORCES: There is one sense in which the Wiretapping part of Gloriagate, was as great if not a greater crime than any of the cheating that might have gone on in the 2004 national elections. Now, as I've said before, there are probably MANY parties who bribed officials and cheated in those elections, but there is only ONE party that was responsible for recording the Garci Conversations. Such recordings may have been done during the course of a wiretapping operation on Garci or they could've been spliced together in a studio. No one seriously believes in the latter possibility. Athough it is not mathematically impossible, the length (over 3 hours), naturalness, absence of obvious splicing artifacts and the rather ambiguous content seem to argue against and Erap-Lacson studio production of the Garci recordings. That leaves the most popular theory -- that the President ordered the Intelligence Service of the AFP to surveil Garci and record his conversations, because as a notorious election operation in Mindanao, they simply did not trust him farther than the nearest cell site. This aspect of the case has NATIONAL SECURITY implications and the illegal use of the Military for such crass political purposes would constitute to my mind a higher crime, a crime perhaps of treason, for so prostituting the armed forces.

GLORIAGATE is still upon us, with no end in sight. I am suspicious that the Return of Garci is a trap to bury the second impeachment attempt in June 2006. There are still people who can afford their Principles and want to uphold them. But many others are either too busy making ends meet, or are mature enough to use "indifference as a tool to force" institutions to clean up their own messes without invoking the people's intervention. I personally do not share the nostalgia for People Power that so many are naively indulging in. I think People Power is finally dead in the hearts of the people, because they knew even before all the smarty pantses did, that Edsa II was a coup d'etat against the President they voted for in 1998, by people who made themselves out as Civil Society battling Evil Society. But once all the dust had settled, the people saw how democracy had been turned into a sham where unelected Justices who forgot about their oath to impartial service conspired with a power-grabbing Vice President and a mutinous Chief of Staff, to overthrow the people's democratically elected leader -- nasty, womanizing dipsomaniac that Erap was. So that not even the revelations of Gloriagate, which most of the people believe, according to the scientific surveys, could rouse them to another bit of Mob Rule in which they would not benefit, again. The message of their indifference to the wailing and chest-beating of the Opposition is that they will no longer join the intifida of one group of Civil Society against an indistinguishable other. What some ignore is that the heart and soul of that Civil Society, which provided the mob rule setting for Hilario Davide's putsch against Erap, -- that Civil Society has even joined with parts of the old Erap Evil Society in wanting GMA out. Don't believe me? Just ask Cory and all the people in this full page ad published in the major dailies last month. I think this time the overthrow of a President for alleged crimes of voterigging and wiretapping, must be done by the Rule of Law, by evidence carefully gathered and presented, by discovering the truth, the whole truth and nothing but the truth -- not prejudice or suspicion or innuendo. It must be done by the book, for it is not Gloria or Garci or Cory or Davide or Drilon or anybody else that matters in the final analysis. What really matters right now is that Democracy be restored, or that indifference of the people will surely lead to a sudden victory by the Nihilists. All patriotic Filipinos must now realize, it will have to be done the hard way, not the quick putschists' way of Davide, Arroyo, Sin and Reyes. It will have to be done by Law and Logic and Science. If I can add to the latter today, I shall be happy...[But please use Google to check on anything I say that you don't believe or understand. I am tired of producing essays that look like this and don't really want skeptical LazyBonz for readers who expect me to do all the work! You can always add to the living organism of Philippine Commentary to bring corrections or other info to its attention -- in the Comment Thread or email.]

FOR THE Hillblogger: RIZALIST used to make a living as a physicist. In a previous lifetime he used ultrasonics and computerized tomography to see into people's bodies as well as to make aircraft engines and jetliners safer for people to fly. (That was a long time ago, way, way before September 11.) Also the article you mentioned, Justice Hurried Is Justice Buried (about the Edsa II Coup d'etat) will be republished in a new blog along with all the Monday Commentaries published by the Philippine Daily Inquirer under the byline of DJB.

RELATED RECENT POSTS:

The Wire Tapping. The Wire Tapping!