|
Post by thewalrus1966 on Sept 30, 2013 16:33:52 GMT -5
Using Wasp 1.54 copyright of Mark Huckvale 2013, University college London, department of speech, hearing and phonetic sciences we are able to analyze five separate video's of McCartney from 1963 to 2006. Firstly, five sound clips were grabbed and placed on one complete sound recording from the following video's: 1. Frost show, 1965 (UK) 2. Beatles interview in 1966 in the USA 3. 1968 interview in USA of Lennon and McCartney 4. 1963 Beatles interview with Dusty Springfield (UK) 5. 2006 Parkinson show (UK) The results are as follows:- please enlarge There is absolutely no argument that these five sounds are the same person. By the way, anyone can perform these tests independently, just give me a PM and I will show you how to do it and read the data correctly. The only way that this could be a different person is if you transplanted the voice box of one person into another, you cannot hide from this technology. It is simply quite impossible. Even if the person sounded similar or nearly the same, each persons voice box produces a unique fingerprint just like your fingerprints or your DNA. For example here is Lennon and McCartney's sound fingerprint on the same program compared:- Completely different, no lateral matching of frequency ranges in the narrow or wide bandwidths, tone, compression and pitch. Whereas the McCartney clips all match in pitch, tone, frequency, bandwidth and form patterns which simply cannot be copied by another person. So I have established and put my mind at rest and I hope everyone else's that the Macca of 63 is the Macca of 2006. The pictures however are a very different kettle of fish, I have an idea what is going on but I'm working on it. To be continued.......
|
|
|
Post by thewalrus1966 on Oct 1, 2013 15:40:16 GMT -5
OK, for the benefit of those who don't understand the sound analysis results I will go into detail about what it all means but first a brief explanation of how we produce sound and how it is measured. I will make this whole exercise as simple as I can. All sound no matter where it comes from is a series of vibrations from an object that travel through the air to our ears. Those vibrations vary in many different ways such as frequency/pitch, tone, volume amplitude and compression. We measure vibration in Hertz, an Hertz is simply one complete vibration (like hitting a drum once). We measure sound frequency in the amount of Hertz per second or vibrations per second. Human beings can hear frequencies usually between 20 Hertz and 20,000 Hertz. When we talk we use our lungs to enable us to vibrate our vocal chords at certain frequencies usually between 20 and 10,000 Hertz. Each individual person however has a unique set of vocal chords apart from identical twins who are remarkably similar, those vocal chords produce a unique sound which characterizes our voice. This unique fingerprint is made up of tone, frequency/pitch, time and volume. But not only that, it is complex and in each word that we speak there is a multitude of different frequencies used and those frequencies are sometimes close to each other or far away from each other. This is compression and is a measure of how fast your vocal chords can change from one frequency to another. When we change frequency slowly those compressions are far apart and when we change frequency quickly they are closer together or more compressed. When we say a word our vocal chords as good as they are cannot hit the correct frequency right away so they have a natural progression through firstly low frequency, then mid range frequencies and finally an high frequency. During that progression there are also what we call frequency nulls. These are frequencies that our own individual vocal chords are not comfortable with and sometimes cannot produce or an accent may prevent the use of. It is the combination of low, mid and high frequencies together with the nulls that give us all a unique voice fingerprint. This picture is from a sound file that was analyzed on Wasp 1.54 of three interviews given by McCartney. They are combined into one sound file so that the fingerprints can be compared. Along the vertical scale to the left is frequency and along the horizontal scale at the bottom is time in seconds. All the spikes, lines distortions you see are all the different aspects of that sound being broken down into readable data. There are two sets of results, narrow band (bottom set) and wide band (top set), narrow band just allows us to examine the results more closely but they are both the fingerprint of McCartney in all three sound clips. This next picture is a break down of what we see, the nulls, the low, mid and high frequencies which we are going to concentrate on. Firstly we examine the nulls, all three sound clips have nulls between 500-2400 Hertz and 4600-6100 Hertz respectively. This doesn't mean there are no sounds at all in those ranges, it means that the mean average of data we see in those zones is far less than the overall picture. McCartney's use of words, sounds and expressions tends to lean away from those zones depending what he's saying. On the picture at the bottom left in the frequency column is a blue number 1. I have added this to indicate where McCartney's low frequencies are. They start at about 10hz and finish at 500 Hertz at the first null, these are what I call the ramping frequencies and can depict tone and how croaky a voice sounds . As you can see they all match all way across time scale from left to right. The blue number 2 is McCartney's mid range frequencies which make up most of the sound he makes. They start at 2400 and finish at 4600 Hertz at the bottom of the second null. The blue number 3 is McCartney's mid to higher frequencies which start at 6100 and finish at about 9900 Hertz possibly touching 10,000 Hertz at times. All these cut off points between mid, low, high and null frequencies on McCartney's fingerprint are in the same places across the mean average of the graph (timescale). Look at the top picture without zooming in or clicking on it, you will see the obvious patterns. There are many, many more of these patterns in more complex close ups under examination but I have chosen to show you the main five speech characteristics of McCartney. McCartney in 2006 is slightly different from earlier but the nulls are still in the same places and the peak frequencies are the same. What has changed are his tones and power which comes with age, the frequencies and nulls are identical. I have examined thousands of sound files of speech over the years and seen many different fingerprints, it is absolutely impossible for two different people to have the same speech fingerprint, everyone is different because they have different length vocal chords, different mechanics in the parts that make up their throat and voice box, their lungs that power the sound, their diet which allows them to develop for the better or worse. Even identical twins over time will develop different length vocal chords with different characteristics, this is because their respective ways of life will change the make up of their body. Even if someone sounds the exact same as someone else their respective fingerprints will be completely different, what sounds the same to a human ear when examined electronically is quite different.
|
|
|
Post by beacon on Oct 2, 2013 3:44:11 GMT -5
Many thanks for taking the time to explain. I think your research on this is very interesting and seems to support the theory that, in the studio at least, it was always the same McCartney. I guess, if it were possible, the thing to do would be to compare your evidence with that that Dr Henry Truby compiled back in 1969 for the famous Life magazine article. Details of which can be found here and here. Truby alleged that there were differences in McCartney's voice. Quoting from the second source above... Voiceprints have been used to detect Faul, the imposter. In 1969, Dr. Henry M. Truby of the University of Miami used samples from three Beatles songs sung by Paul McCartney (Yesterday, Penny Lane, and Hey Jude) and produced three very different sonagrams. If you could debunk that research that would be hugely significant. One question though. If your theory concerning the stroke is correct, would that not have possibly permanently affected McCartney's voice?
|
|
|
Post by thewalrus1966 on Oct 2, 2013 10:28:56 GMT -5
Many thanks for taking the time to explain. I think your research on this is very interesting and seems to support the theory that, in the studio at least, it was always the same McCartney. I guess, if it were possible, the thing to do would be to compare your evidence with that that Dr Henry Truby compiled back in 1969 for the famous Life magazine article. Details of which can be found here and here. Truby alleged that there were differences in McCartney's voice. Quoting from the second source above... Voiceprints have been used to detect Faul, the imposter. In 1969, Dr. Henry M. Truby of the University of Miami used samples from three Beatles songs sung by Paul McCartney (Yesterday, Penny Lane, and Hey Jude) and produced three very different sonagrams. If you could debunk that research that would be hugely significant. One question though. If your theory concerning the stroke is correct, would that not have possibly permanently affected McCartney's voice? Hi, I can't comment and speculate on that, all I can say is that my analysis shows the three voices to be the same person. Truby wasn't using the equipment we have available today and to compare songs against voice would be different anyway because singers often change the key they sing in and the tone depending on the song. Also, a lot of the time accent isn't present during songs so part of the structure of the voice can be missing. But it would be interesting to compare results with Truby.
|
|
|
Post by Jai Guru Deva on Oct 2, 2013 15:39:17 GMT -5
Interesting audio analysis
However, I'm not convinced all five are of the same person. When I watch and listen to the videos, I see and hear the same person Dusty Springfield interview, the Frost interview, and 1966 interview (Ator Towers Hotel, Los Angeles interview, last interview?). And I see and hear the same man in the 1968 Lennon & McCartney interview and the Parkinson's interview. But I don't see and hear the same man in the '63, '65, '66 interviews as being the same person in the '68, '06 interviews.
|
|
|
Post by cherilyn7 on Oct 2, 2013 16:58:30 GMT -5
Interesting audio analysis However, I'm not convinced all five are of the same person. When I watch and listen to the videos, I see and hear the same person Dusty Springfield interview, the Frost interview, and 1966 interview (Ator Towers Hotel, Los Angeles interview, last interview?). And I see and hear the same man in the 1968 Lennon & McCartney interview and the Parkinson's interview. But I don't see and hear the same man in the '63, '65, '66 interviews as being the same person in the '68, '06 interviews. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ This is what appears to be the case; there was a "real Paul" pre 1966 but also doubles/imposters being used as well but after 1966 that person known as "real Paul" was never seen again and another (2) individuals took over. One was having operations on his chin and eyes; the other was his understudy..... As for voice: listen to circa 1963 songs such as "A Taste of Honey" and "All My Loving" where JPM took lead vocal and you will see the voice was much mellower and richer and a lower tone than later and I think Truby's analysis proved it was a different "voiceprint".
|
|
|
Post by thewalrus1966 on Oct 3, 2013 11:33:54 GMT -5
Truby's analysis ain't worth Jack and wouldn't stand up in a court of law. The equipment I am using is state of the art hardware and voice analysis software used by departments of Phonetics in Universities and labs the world over, this hardware and software is used to prove murder cases where telephone's have been used and is 100% reliable. I'll tell you what you can do, those of you who are doubting Thomas's at least. Record McCartney talking on any interview from any year 1962 to present day. Record him using Windows sound recorder in 16 bit stereo, 44Khz which will cover all frequencies. Save the sound files to your hard drive. Contact any University Phonetics department or any sound analysis lab in the world including the FBI at Langley Virginia via email and ask them if they will partake in a little test to see if it is the same person on the recordings. Or......download the software yourself, read how to use it, talk to people on University websites and ask their advice if necessary on how to read the data. There are about six programs available that are industry standard and are widely accepted as 100% reliable by experts in the field. Jai Guru Deva and cherilynzennah7, with all due respect, it doesn't matter what you 'feel' sounds different or what you 'think' is or isn't McCartney. You are voicing an opinion without one shred of evidence that is based on speculation. Why don't you do what I have suggested above and prove it to yourselves for real.
|
|
|
Post by decitfaul on Oct 4, 2013 10:51:01 GMT -5
This is very interesting, this is the type of research that I'm open to. I do have one question.. I know that it has been stated that even if two people sound the same the program would still be able to spot the differences in their voiceprints.. how about an example of that. How about putting a known voice impersonator and comparing him with the person he is impersonating? I think that could possibly rule out if impersonators can produce similar voiceprints. Just a thought though..
|
|
|
Post by thewalrus1966 on Oct 4, 2013 11:32:36 GMT -5
This is very interesting, this is the type of research that I'm open to. I do have one question.. I know that it has been stated that even if two people sound the same the program would still be able to spot the differences in their voiceprints.. how about an example of that. How about putting a known voice impersonator and comparing him with the person he is impersonating? I think that could possibly rule out if impersonators can produce similar voiceprints. Just a thought though.. OK, I can do that no probs.
|
|
|
Post by thewalrus1966 on Oct 4, 2013 12:20:06 GMT -5
OK here we go: During the Iraq war Tony Blair was interviewed about the UK's involvement in the war. At the same time a pretty good impersonator Rory Bremner impersonated Blair from that interview and I may say was pretty convincing. Here are the results of the voice fingerprints. Please click to enlarge. Here is the Analysis: The ramping frequencies are similar but the compression is wider on Bremner and that is where any similarities stop. Blair's entire voice apart from a few small peaks is in the mid range from 900Hz to 7000Hz with a very faint null between 3200 and 3900Hz which is hardly visible. Then a Massive null between 7000 and 10600Hz which is his high frequency. Bremner on the other hand hits a similar ramping frequency then has a smooth frequency layout with no prominent mid ranges right up to 11,000Hz which is his high frequency. He has a clear null between 2900 and 3400hz which is missing from Blair's. Two people sound remarkably similar but you CANNOT hide from this technology.
|
|
|
Post by decitfaul on Oct 5, 2013 0:05:27 GMT -5
Thanks for the reply.. however does this mean everybody has a unique voiceprint no matter what? What about if its the same person but with different moods? Is the voiceprint the same if they are energetic when they talk or if they are more mellow? I'm just trying to get a consistency with the results. So it would be impossible for an impersonator to practice again and again on a particular person ( tone of voice, style of speaking, etc.) and they would still not be able to get within range of what would be considered normal of the person they are impersonating?
|
|
|
Post by thewalrus1966 on Oct 5, 2013 4:46:08 GMT -5
Thanks for the reply.. however does this mean everybody has a unique voiceprint no matter what? What about if its the same person but with different moods? Is the voiceprint the same if they are energetic when they talk or if they are more mellow? I'm just trying to get a consistency with the results. So it would be impossible for an impersonator to practice again and again on a particular person ( tone of voice, style of speaking, etc.) and they would still not be able to get within range of what would be considered normal of the person they are impersonating? Hi decitfaul, if you look over on the general section I posted my conclusions of my analysis, I strongly advise you do the analysis yourself because I haven't got the time at the moment, full instructions are on there. But I will answer your query. Everyone's voice box is incapable of producing frequencies out of its range otherwise we would all be able to sing like Celine Dion. Each person has frequency nulls or dead spots if you like in their voiceprint that no matter how hard you try, you will not be able reproduce certain frequencies. It's the physical makeup of your body that causes this including how you were brought up. If your voice box is incapable of producing perhaps 2200-2250Hz then I'm afraid it doesn't matter if you are sad, happy, energetic or otherwise you will not be able to produce that frequency even if you practice and it will show up as a null on the voiceprint, its a physical characteristic of the vocal chords that remains with you for life. Like I have said, if we didn't have those properties in our voice we would all talk the same and sound like Celine Dion when ever we felt like it. Hope this answers your query.
|
|