KVR Audio

IncarnateX · Post by **IncarnateX** » Sat Aug 22, 2015 3:11 pm

Ogopogo wrote:
IncarnateX wrote:There is no contradiction here. To judge the validity of a test you need to know what it is aimed at testing and for this purpose you need a main hypothesis. To judge the validity of your main hypothesis the results have to reject the null hypothesis. The Null hypothesis is the opposite of your main hypothesis, it is the enemy you have to defeat (statistically)
Ok, but as analogue said there are only really two possible hypothesis so that's not really an issue.

So after touching on all these issues I think my original point stands, there's nothing to really disqualify this as a scientific test assuming that there are enough responses. It's certainly not the best test one could design but it's not just subjective.

Sorry but you have not got it. There is no hypothesis given and now you say there are two. Read analogues last response to me and see which hypothesis could have been put forward with this design but wasn't. And even if it was, the fixed sequences of testing would have made it less optimal. If you do not grasp it this time I must recommend that you read something about scientific tests. Start with wiki articles and move forward to the textbooks they refer to.

kvaca · Post by **kvaca** » Sat Aug 22, 2015 3:21 pm

cocoazenith wrote:
nasenbluten wrote:Test 3 gives it away. Hardware is B.
is there a test 3? i only listened to 2 tests...

me too...
but I start to think it degraded to pointless debate plagued with OT comments only...nobody is interested in sound examples now

Ogopogo · Post by **Ogopogo** » Sat Aug 22, 2015 3:41 pm

IncarnateX wrote:Sorry but you have not got it. There is no hypothesis given and now you say there are two. Read analogues last response to me and see which hypothesis could have been put forward with this design but wasn't. And even if it was, the fixed sequences of testing would have made it less optimal. If you do not grasp it this time I must recommend that you read something about scientific tests. Start with wiki articles and move forward to the textbooks they refer to.

Analogue didn't give a hypothesis at first, but then said "However in reality its fairly obvious if you had to make a hypothesis it would either be 'people on KVR can tell the difference between Diva and OB8 in specific examples' or 'people on KVR cannot tell the difference between Diva and OB8 in specific examples'." There are the two options I was talking about.

Whichever one of these you want to use doesn't really matter because as I said in another post the data still makes it clear whether people are distinguishing between the two or not. You can pick either hypothesis and write it up, the test still works.

We've established that the order wasn't optimal but we've also established that doesn't render it unscientific. not all scientific tests are perfectly optimal.

Ogopogo · Post by **Ogopogo** » Sat Aug 22, 2015 4:00 pm

The point is that whether the hypothesis happened to be stated is a technicality. 50% right means people can't tell, significantly higher means they can. Nothing subjective about that.

SWAN808 · Post by **SWAN808** » Sat Aug 22, 2015 4:08 pm

Come on guys. ..you both have a point but its done now. The forum test is not intended to be scientific but there is the option with the wav files. Yes post-hoc you could make a scientific observation about the results if you want to but it's limited.

What is your guess with the test and anything stand out? I really recommend downloading the wav of the second video for detailed listening

Ogopogo · Post by **Ogopogo** » Sat Aug 22, 2015 4:10 pm

I already guessed a is hardware but I'll try the wavs. When are you going to reveal the answers?

IncarnateX · Post by **IncarnateX** » Sat Aug 22, 2015 4:16 pm

Ogopogo wrote:
IncarnateX wrote:Sorry but you have not got it. There is no hypothesis given and now you say there are two. Read analogues last response to me and see which hypothesis could have been put forward with this design but wasn't. And even if it was, the fixed sequences of testing would have made it less optimal. If you do not grasp it this time I must recommend that you read something about scientific tests. Start with wiki articles and move forward to the textbooks they refer to.
Analogue didn't give a hypothesis at first, but then said "However in reality its fairly obvious if you had to make a hypothesis it would either be 'people on KVR can tell the difference between Diva and OB8 in specific examples' or 'people on KVR cannot tell the difference between Diva and OB8 in specific examples'." There are the two options I was talking about.

Whichever one of these you want to use doesn't really matter because as I said in another post the data still makes it clear whether people are distinguishing between the two or not. You can pick either hypothesis and write it up, the test still works.

We've established that the order wasn't optimal but we've also established that doesn't render it unscientific. not all scientific tests are perfectly optimal.

Wrong again. For those performing statistics, a design always has to be as optimal as possible. Not at least the fact that people were TOLD that the sequences were fixed and thus knew they only had to get tit right once, would have been enough for the statisticians at our university to reject it as a study. I have no idea why you think sloppy science is as good as any science but to me it sounds extremely student-like. This is obviously not your game, so stick to things you know something about be ause this is surely not one of them.

zerocrossing · Post by **zerocrossing** » Sat Aug 22, 2015 4:18 pm

andrelafosse wrote:A month ago, I had the opportunity to record with a fabulous engineer in his home studio. This is someone who's recorded and/or mixed for Nine Inch Nails, Clint Mansell, Kronos Quartet, Harold Budd, Terry Riley, etc. He had lots of boutique hardware gear - effects, preamps, tube amps, etc.

I was asking him his opinion about analog, digital, in the box vs. out of the box, etc. At one point, he kind of summed it up like this (I'm paraphrasing):

"(Pointing to his rack gear) I have a hardware LA-2A over there. It sounds really good. (Points to his computer screen, where a plug-in window was open). I have that plug-in emulation of an LA-2A in there. It sounds really good. Are there some subtle differences between them? Probably. Are there differences between two different hardware LA-2A's? Definitely. Some people get really obsessive about these distinctions. I don't have time to get wrapped up with that stuff - I have records to make."

Hi Andre! It's Mark, from the old LD list.

I totally get this, and I agree with it... mostly. You play guitar. Could you forget about practicing guitar and instead just use virtual guitar plug ins like String Studio? Some well made scripted Kontakt libraries? Just think of all the time you'd get back in your life!

But, as a fellow guitarist I already know your answer.

Sometimes you just need the real thing.

Synthetic Wav · Post by **Synthetic Wav** » Sat Aug 22, 2015 4:24 pm

can we get the answer ? what are we waiting for

SWAN808 · Post by **SWAN808** » Sat Aug 22, 2015 4:30 pm

Hi its only been up a day so ill give it a bit longer for people who might check in less frequently.. ...but in the meantime I really recommend people check out the wav in their DAW and don't rush to answer. A kneejerk response isn't always accurate IME...you have to get used to the audio before the difference becomes clearer...Some of the difference people refer to when they discuss analogue is there IMO. ...

zerocrossing · Post by **zerocrossing** » Sat Aug 22, 2015 4:33 pm

aciddose wrote:
penguinfromdeep wrote:What is your guess aciddose?
My best guess is that your methods here are so far from scientific that it makes me laugh.

What are you attempting to prove? What this "test" proves is that given two entirely random bits of data, it isn't possible to accurately identify one or the other.

It's random chance.

All you've done is used your outstanding brilliance to reproduce a coin flip in an amazingly obtuse way.

Except almost everyone had a very specific idea of which was which. That's no coin toss. (oddly the results on Gearslutz are less polarized.)

Here's the deal about "scientific." (I saw this in a movie last night (Ex Machina, it's excellent) and it made me think of this) What kind of woman/man are you into? Blonde? Brunette? Black, white, fat, skinny, tall short? You all have your "thing." Is it scientific? Maybe. Maybe you have some biological imperative based on your environment to either water down or shore up your genetic code. Maybe your kindergarten teacher set your mode for life. Who knows? It's your thing and you know it when you see it and "science" at that point becomes irrelevant. I didn't have to put the examples though some spectrum analyzer or scope. I just knew. Similarly, I "knew" the sound of the Prophet 12's sawtooth and pulse sounded odd. I did investigate with tools, but only because I had the general feeling first. The result, for me, was they don't sound very good if you have the filter opened all the way up. Now I know. Anyway, the day we pick our instruments based on science is the day art is dead. If there's one thing that should be totally "feelings" based, it's that.

Ogopogo · Post by **Ogopogo** » Sat Aug 22, 2015 5:04 pm

IncarnateX wrote:Wrong again...

"As optimal as possible" shows that studies can't always be perfectly optimal. Data doesn't have to be perfect to be scientific data, and the definition of "scientific" is not just the standard that you or your friends would prefer. Look up "scientific" and you'll see that. I never once said that "sloppy science" is at good as any science, just that it's science, and you just admitted that. You're grasping at every straw and trying to argue from authority to keep from admitting your were wrong from the start that it's not scientific. Keep studying, and take your own advice about not talking about stuff you don't understand.

IncarnateX · Post by **IncarnateX** » Sat Aug 22, 2015 5:08 pm

analoguesamples909 wrote: Yes post-hoc you could make a scientific observation about the results if you want to but it's limited.

Apart from the issues already discussed, I am afraid that the lack of 100% control over the experimental situation would add to the troubles making this a trusthworthy scientific study, even post-hoc. In principle people can cheat here, e.g. download the files and anlyse them for amount of noise. The noise of an analog synths is often more prominent than of a VSTi. You were asked earlier if you had made up for this somehow (like sending then out of the same outboard or adding noise to Diva) and I am still curious with regard to this, not at least because it was the basis of my own choice. I simply chosed the one that sounded least noisy to me as Diva.

Teksonik · Post by **Teksonik** » Sat Aug 22, 2015 5:15 pm

Nerds...............

IncarnateX · Post by **IncarnateX** » Sat Aug 22, 2015 5:17 pm

Ogopogo wrote:
IncarnateX wrote:Wrong again...
"As optimal as possible" shows that studies can't always be perfectly optimal. Data doesn't have to be perfect to be scientific data, and the definition of "scientific" is not just the standard that you or your friends would prefer. Look up "scientific" and you'll see that. I never once said that "sloppy science" is at good as any science, just that it's science, and you just admitted that. You're grasping at every straw and trying to argue from authority to keep from admitting your were wrong from the start that it's not scientific. Keep studying, and take your own advice about not talking about stuff you don't understand.

The discussion is about whether this test has been put forward and lives up to the standard of a well conducted scientific test. The answer is no in both cases and the first is even confirmed by the author himself. And of course scientific data has to be as perfect as possible, that is what keeps science apart from pure belief. And my "friends" are Ph.Ds and employed at a university just like I am, so I think a little appeal to authority is fair here, especially in contrast to your less than undergraduate take on this.

Class is over for today, my little student. It is time for you to go home and make your homework

Diva vs Analogue - a real world test