so are we having a disagreement about the word "scientific"?
Scientific is basically meaning a systematic approach that uses the predetermined methods and predetermined tests.
by this definition those comparisons would be scientific.
The approach is as scientific as possible, but it cannot be considered as scientific.
When you can weigh, measure, contain and reproduce a phenomenon then you can safely say it is a scientific approach.
In our case, when you take the altered voice and study it, all you study is the altered voice. There is no scientific way to prove that this altered voice is more Michael Jackson's than someone else's. In other words, scientifically speaking, it is not Michael Jackson's voice unless you can revert the processing and make it sound as all the voice we all know.
If you use the argument that the altered voice is due to the recording environment, thus impossible to make it sound as a usual voice, then I can say, ok, take the THRILLER 25 session where some verses were apparently recorded at the Angeliskon studio and make it sound as the voice on the Cascio tracks. Same recording environment, same processing, should give same results.
I think what you have an issue is with accuracy. Correct?
Let me explain it further
DNA test is seen as one of the most reliable tests out there. Fingerprinting is more like 7 point match , 12 point match which is known to give false match results.
So they are both scientific tests with one having more accuracy then the other.
They both are tangible and can be safely compared, unlike the altered voice.
If I modify the DNA via microbioprocessing or deliberately modify fingerprints or traces of fingerprints thanks to microtechnology or nanotechnology, I can no longer claim to be as accurate as when I had the unaltered DNA and fingerprints.
So, please, let's not lose sight of the fact that the voice might have been altered (or not --which obviously would pose a bigger problem as to relating it to MJ's).
So those computerized comparisons might not be perfect and have error chance but it doesn't mean that they aren't scientific or objective.
Their objectivity depends on the encoded data. If there is no trace of JM's encoded data in the computerized software, the software couldn't possibly compare it to something that is absent from his data --JM's voice.
Comparing MJ's voice to a similar voice cannot be conclusive. The voice still sounds similar to MJ's and not the same.
I don't think this is the case. For example US used such tests to authenticate audio messages from Osama Bin Laden , Saddam Hussein before and all those tests aims to authenticate such voices and obviously it would aim to differentiate it from a sound alike and not from a distant voice. So I again expect them to be able to differentiate between the voices - even though they might not be perfect.
Well, that's what I have been saying -- the voice authentification is not pefect. You can always fool people, even the experts, and they admit it.
But have you heard Teddy and Eddie how lightly they spoke about it? "It's Michael". I mean, what?! how? when? where? why?... But no coherent answer surfaced. A pvc, a shower? I mean, even if Michael had been drunk and had sung in a huge wine keg we'd immediately gut-recognize him and even giggle about it. But here, we clearly are dealing with something unprecedent. Every Cascio song contains past MJ's songs and that a bit alarming to my ears.
Water = Heaven can wait
Monster = 2Bad
Breaking News = You rock my world
Stay = You are not alone
Keep your head up = Earth song
Soldier boy = They don't care about us
Set aside the altered voice, the rehashed and copypasted materials from past songs onto the Cascio songs is a bit too much. Enough to become suspicious.
It is however true that Another Day contains some pasted material, as well as Behind the mask. To be honest I could understand that some additional ad-libs were added to Another Day as they were lacking, but when I heard Behind the mask's crowd from Dangerous tour, I just couldn't understand why they pasted that at all, it was completely unnecessary and unrelated.
In case of the Cascio songs, I have impression that the copy pasted material is absolutely related and that it served to fabricate the "new" songs. That's how they sound - artificially fabricated. Just listen to the word "victim" in the vocals on "Soldier Boy". It sounds as if it hasn't been processed yet and that it was blatantly copy-pasted from another recording session.
Now let's talk some statistics
If we are trying to make a determination between 2 choices (Michael - not Michael) chance is 50% - meaning that if I asked my 3 year old niece to give her opinion without knowing anything she has the 50% chance to give the correct answer and 50% chance the wrong answer
False.
It depends on the choice you offer. If you present two different types of voices, for example Michael vs Sponge Bob, you will not have 50% chance of the wrong answer at all.
If you present a choice between two similar voices, then yes, there's a 50% chance that people could give you the answer if they don't know any of the voices.
so we can conclude that any test should have given more than 50% correct answer.
Absolutely not. I strongly doubt, and many will probably agree, that Sponge Bob vs MJ would give different results from MJ vs JM due to the similarity of the latter.
Nothing is perfect in such testing every test would have an error rate so
so it cannot be taken as a serious scientific proof. Water is made of hydrogen and oxygen, that's scientifically provable. But the Cascio voice being made of MJ's voice, is scientifically impossible to prove.
- 99% confidence level would be perfect
- 95 % would be the the most common "best" match
- 80 % would be okay
- 60% would be weak
- and anything around / below 50% it's better to just guess.
now we need to add probability into the mix. Let's assume a really weak test like Bumper portrays actually let's go with guess 50%.
The lacking factor here is that we are not guessing. We are putting side aside two voices: JM's and Cascio's and it sounds identical. It's not a guess, it is a comparison. Yet, impossible to find an MJ's vs Cascio comparison.
You know what that means 12 Cascio songs if they ran through the test 6 of them will come as Michael and 6 of them would come up as not Michael.
Not at all. This is where I am goig to use my Sponge Bob. If I give you 12 songs sung by Sponge Bob vs MJ, and ask you to guess, I doubt you would give a 50% chance answer, even if you don't know the voices. I doubt that your brain wouldn't from the very beginning make the difference between MJ's and Sponge Bob's voice. So, you'd guess with 100% accuracy that all 12 songs are sung by Sponge Bob. That's simply natural voice recognition.
and guess what any researcher would see that the test wasn't testing what it's supposed to test.
Well, the way you presented it in your example that I just countered, there's no need to be a scientist to see that.
so I think the tests ability to differentiate between the voices and the accuracy of the outcome will be better than you portray.
You are omitting the altered voice factor which is impossible to revert back. So how accurate can it be anyway, especially if it is compared only to Michael's voice and not JM's voice.
This is actually something I have been thinking about since reading PG's post on Maxjax about mother's not recognizing their babies cry.
Unfortunately, since PG banned me for merely posting an innocent joke on his website, I couldn't possibly comment on his mother-baby argument which is completely out of scope and has nothing to do with the issue we are dealing with. Here is what I have to say to PG argument:
-First of all Michael's voice isn't as simple as a baby's voice.
-Second, a baby's voice is easier to imitate beacuse of its simplicity.
-Third, if a voice is easier to imitate, there is a bigger chance to have more soundalikes (other babies).
-Fourth, it is not the mother that does not recognize her baby's voice, that's completely false! The issue is that a mother (or a father by the way) does recognize her baby's voice, but has a hard time telling if it is her baby's or someone else's baby's, due to the simplicity in the voice factor.
Now let's get serious and talk about a more complex voice which we are dealing with - Michael Jackson's vs Cascio's.
After Bumper's posts of "our hearing better determinant than experts/ tests" I had to go back and post this information.
I stand by that. In some cases some people can naturally better determine what they hear than what the experts hear. You can compare this actually to a visit to your doctor. You perfectly well know what seems to be the problem and the doctor tells you the opposite. You listen to the doctor's piece of advice and realize that your state gets worse. When you come back to see the doctor again, finally he tells you that your problem is exactly what you had been suspicious about in the first place.
This comes from Hastings Law Journal and it's about speaker identification. The relevant part is that it quotes a lot of research about humans ability to recognize voices. This is the summary
1. Voice recognition software is not as reliable as DNA or finger print analysis but it's improving "sophisticated mathematical modeling techniques, and the ability of acoustic engineers to decompose the human voice into a host of different components have led to enormous improvement in voice recognition technology". So like I said in the last post it's not perfect but it's improving.
2. Some people are better at identifying voices then others. Nobody knows why but musical ability and memory seems to be a factor.
---- So musicians will generally be better at identification. so musicians that worked with Michael would have a better accuracy.
3. Familiarity with the voice is important. Research shows people can identify their family members with 89% accuracy but they can only identify their friends / co workers / people they have occasional contact by only 66% accuracy.
4. Number of times a person is exposed to the voice is more important then the length of the time they are exposed to the voice. Research shows that hearing a voice for 20 seconds for 3 times is far better than hearing a voice for once for 60 seconds.
---- So both of the above shows that more exposure to the voice is the most important thing. so not necessarily the long term fans will be better judges, the people who listen/ exposed to the voice the most times will be better.
---- Although being family strengthens Jacksons family claims, how often they were exposed to Michael's voice also becomes important. For example in her book Latoya says between 1992 - 2009 she didn't talk to Michael for 14 years and she only talked to him 6 or so times. So in this reality most of the fans would be better in identifying Michael's voice than Latoya.
5. The ability to identify a voice also deteriorates sharply over time if the person is no longer exposed to the voice frequently. Research shows that when asked to pick up a voice among 5 voices , people has 83% accuracy the next day, 51% accuracy 3 weeks later and 13% accuracy after 5 months.
---- this with the above points shows that not only familiarity in Michael's voice but familiarity in Jason's voice is needed. If you aren't exposed to Jason's voice before and frequently or if you didn't listen to him for some time then your ability to distinguish his voice is going down sharply.
6. Now sound- alikes. Bumper is right. They are tricky. Research shows that voices of brothers are confusing. In a research done in Sweden when asked to pick Carl Bildt voice (prime minister) but actually played an impersonator voice people almost always wrongly identified the impersonator as Carl Bildt.
---- This is a two way street in our argument. It shows that Malachi can be mistaken for Michael and Micheal can be mistaken for Malachi so any one of us can be wrong in what we think we hear. It shows us that we are actually trying for the hardest differentiation in this instance.
7. Disguise of a voice (in our case processing) is also problematic. They ask people to simply whisper and recognition accuracy goes down a lot. Distortion makes it even worse in regards to accuracy.
---- so any type of processing present on Cascio and Malachi songs is making our determinations even less accurate. Raw acapella vocals are needed to reduce this disguise / distortion effect.
8. Also multiple case histories show that being confident in a voice identification doesn't necessarily mean that the person is correct. Several cases are cited where people were convicted based on mainly a very confident witness voice identification but later proven innocent with DNA evidence and such. In short there's no correlation between a person's confidence in what he heard and the accuracy of his identification.
Finally let's discuss the comparisons from a legal perspective
"it is no longer permissible for police officers to invite the victim of a crime to the police station, bring her within earshot of a defendant who is asked to say a few words, and then ask the victim if the defendant was the perpetrator of the crime"
Why?
Because it's leading and proven to cause false identification.
Similarly any comparison video or any person you asked to compare Jason Malachi vocals to Michael's wouldn't hold in court.
I'll comment later on your neither-melting-pot-nor-salad-bowl-but-Irish-stew post