shaffer said:

Clearly, you have no idea, whatsoever, what it takes to perform valid audio DBT; you described a SBT with zero statistical relevance, or otherwise. Science. It's is high school under another name. Yes, I advertise the AES. You obviously have no idea what it is; to you it's just some site. Unbelievable.

Click to expand...

hahaha i usually ignore you, but you are so blatantly wrong I just can't help but comment. what does double blind mean? it means that the information that may influence the behavior of the tester is not revealed to either the tester or researcher until after the exam. the purpose of blinding the researcher is so that they cannot unconsciously influence the results.
 
hence, if the researcher is a layman who does not know what lossy or loseless means & is looking at file 1 vs file 2 named boogers & yucky, they are effectively blinded. they have no idea what the right answer is or you are even experimenting for, they are just recording the results. there is no way for them to unconsciously influence the results because they do not know the right answer.
 
you don't need to follow any fancy protocol or pay money for AES if you simply understand how to conduct a basic scientific experiment. in terms of statistical relevance, you just have to have a large enough sample size and calculate the p-value.
 
but yes, please continue telling us how the only way for anyone to know anything about audio is to subscribe to AES and pay their membership fee. very helpful lol the whole purpose of science is reproducible results. you don't need a membership to AES to run your own experiment. you just have to understand how to set up a controlled experiment and remove confounding variables.
 
...and yes, all these concepts are taught at the high school level.
 
and btw, you do realize that the results of the double-blinded study in the AES paper with over 500 trials is that listeners picked the 'superior high-resolution' source less than 50% of the time when listening for differences between 96khz 24bit vs 44.1 khz 16bit?

money4me247 said:

hahaha i usually ignore you, but you are so blatantly wrong I just can't help but comment. what does double blind mean? it means that the information that may influence the behavior of the tester is not revealed to either the tester or researcher until after the exam. the purpose of blinding the researcher is so that they cannot unconsciously influence the results.

hence, if the researcher is a layman who does not know what lossy or loseless means & is looking at file 1 vs file 2 named boogers & yucky, they are effectively blinded. they have no idea what the right answer is or you are even experimenting for, they are just recording the results. there is no way for them to unconsciously influence the results because they do not know the right answer.

you don't need to follow any fancy protocol or pay money for AES if you simply understand how to conduct a basic scientific experiment. in terms of statistical relevance, you just have to have a large enough sample size and calculate the p-value.

but yes, please continue telling us how the only way for anyone to know anything about audio is to subscribe to AES and pay their membership fee. very helpful lol the whole purpose of science is reproducible results. you don't need a membership to AES to run your own experiment. you just have to understand how to set up a controlled experiment and remove confounding variables.

...and yes, all these concepts are taught at the high school level.

and btw, you do realize that the results of the double-blinded study in the AES paper with over 500 trials is that listeners picked the 'superior high-resolution' source less than 50% of the time when listening for differences between 96khz 24bit vs 44.1 khz 16bit?

Click to expand...

You've just proven my point. Congratulations. As a consolation prize, I'll try to talk to adults from now on.

money4me247 said:

step 1) rip a song lossless from a cd. call it boogers.
step 2) rip the same song lossy from the same cd. call it yucky.
 

Click to expand...

 
 
Hmm. I'm not sure I want to take part in that test.

No one has as yet addresssed my point that whether or not one can discern a difference under test, it's unlikely to be of great significance in everyday listening. The point is, do we sit listening for the minutest difference, or do we get involved in the music? And please, don't tell me such minute differences are the very point of hi fi. They're the point of constantly changing equipment in a state of permanent dissatisfaction, not hi-fi.  
 
There may be discernable differences, since discernable is such a subjective word, but to claim that the difference between good 320 MP3 and flac is going to seriously impair one's enjoyment of music...sorry, I just don't buy it.

pp312 said:

No one has as yet addresssed my point that whether or not one can discern a difference under test, it's unlikely to be of great significance in everyday listening. The point is, do we sit listening for the minutest difference, or do we get involved in the music? And please, don't tell me such minute differences are the very point of hi fi. They're the point of constantly changing equipment in a state of permanent dissatisfaction, not hi-fi.  

There may be discernable differences, since discernable is such a subjective word, but to claim that the difference between good 320 MP3 and flac is going to seriously impair one's enjoyment of music...sorry, I just don't buy it.

Click to expand...

I can enjoy music played on a small transistor radio, but that's hardly relevant to offering a service with a higher quality stream.

For me, the whole point of assembling a good system revolves around hearing more of what's on the recording. Why, because I want to hear more of the music and its associated components. Do I need to post pics of my record shelves and CD racks?

But honestly, tell me that the difference seriously impedes your enjoyment. I mean, like, one minute you're air conducting or air guitaring or whatever, and suddenly you think, 'Damn, if only I'd downloaded this in Flac instead of 320 MP3 it would sound soooo much better. I just can't enjoy it like this.' And if so, wouldn't there be just a little bit of self confirmation in it? I mean so many people with good ears can't hear any difference whatsoever. Maybe you and some others can, but to the point what it's a real factor in everyday listening?

pp312 said:

But honestly, tell me that the difference seriously impedes your enjoyment. I mean, like, one minute you're air conducting or air guitaring or whatever, and suddenly you think, 'Damn, if only I'd downloaded this in Flac instead of 320 MP3 it would sound soooo much better. I just can't enjoy it like this.' And if so, wouldn't there be just a little bit of self confirmation in it? I mean so many people with good ears can't hear any difference whatsoever. Maybe you and some others can, but to the point what it's a real factor in everyday listening?

Click to expand...

Yes. Plus, why would I download a MP3? To what end? If I want a record, I buy the real thing and listen to it.

Actually, let me correct myself. I did download 1.5GB of Aphex Twin unreleased cuts that were only available as MP3s. Do I listen to it often? No. I have plenty of his albums.

Edit: text

pp312 said:

  No one has as yet addresssed my point that whether or not one can discern a difference under test, it's unlikely to be of great significance in everyday listening. The point is, do we sit listening for the minutest difference, or do we get involved in the music? And please, don't tell me such minute differences are the very point of hi fi. They're the point of constantly changing equipment in a state of permanent dissatisfaction, not hi-fi.  
 
There may be discernable differences, since discernable is such a subjective word, but to claim that the difference between good 320 MP3 and flac is going to seriously impair one's enjoyment of music...sorry, I just don't buy it.

Click to expand...

This is actually very pertinent to the discussion, and frankly never given the weight in conversations it deserves. So let's just assume people can actually hear differences between lossy and lossless, but they can only do so under very controlled conditions, and as with the results I have found, even when people detect differences the rate is barely statistically significant, close to chance. So in terms of effect it is measureable but not practically significant. I think when most people around here talk about being able to hear differences we mean in normal listening situations. While most of us are audio nerds, the vast majority of the audio consuming population are far less likely to be so keen on absorbing every nuance of the sound. Frankly, a lot of people I know don't care at all, nadda, zilch, zippo, music is just background noise that on occasion catches their attention.
 
Even among the head-fi population you will find a significant number of members who say they don't dissect the music by paying that type of attention, they just enjoy. Even with good equipment in casual listening I still remain skeptical  that very many people can spot the differences. Maybe, if we train our ear and listen hard under ideal conditions it could be possible, but I'm still not convinced. I haven't been able to look at this magical AES trial in detail and many experiments are conducted and even published in peer-reviewed journals only to be later assessed as invalid or less conclusive.
 
If I was to put any stock in the AES experiment I would most certainly need to know how the experiment was designed. For instance, how many subjects, how many trials, what controls were put in place to still capture guessing? Many other questions. For instance, were dummy trials done meaning subjects only played one type of file format multiple times to see if they identified differences? Has the experiment been replicated elsewhere? Just telling me there is this great study means very little regardless who conducted it unless we can see the details.
 
Consider a meta analysis of studies used by pharmaceutical companies to demonstrate the effectiveness of flu vaccinations. These studies were used as evidence of efficacy and supplied to government regulators who then implemented mass immunizations for flu. Oh, this study is free and available at the Cochrane Database, please feel free to read it yourself. I was directed to this study by an expert who researched governmental spending on health care.
 
The authors of the study looked at I think around 146 studies. Guess what they found? Of all of the studies included in the meta-analysis only two or three were determined to be scientifically valid due to all kinds of experimental design flaws. These we're pretty deep pocket pharmaceutical companies pursuing hundreds of millions of dollars in product sales, I think on paper like the AES you would assume they would produce valid studies, but they didn't.
 
Now maybe the AES study is excellent and reliable, in which case it should be replicated, but until I know more about the design and methods used I can't simply take somebody's word that it proved anything. Hell we still can't prove that smoking causes cancer. It is quite likely, but there is no ethical (or even plausible) experimental design that would allow us to be 100% confident that smoking causes cancer. Provide us the details of the design and then we can have more confidence. Until then, I'm going to remain a healthy skeptic.

I have not been able to find any study that shows that the majority of people can reliably identify the differences between lossless and lossy in a controlled blinded test. If someone has a link to such a study, I would be interested in reading.
 
@Sonic Defender, I agree with most of your points until you brought up smoking and cancer. Smoking is a very serious risk factor for lung cancer. Of course, not all smokers will get cancer, but its strong association with lung cancer cannot be denied. For anyone who is struggling with nicotine addiction, try using e-cigs to cut back. You still relieve the oral fixation and hand-mouth habit, and you can gradually taper down the nicotine dose. While the safety of vaping has not been studied at all, I think it is a good way to try to gradually stop if cold turkey didn't work. It ends up being less expensive in the long run as well. Otherwise, there is patches, gum, bupropion, and varenicline (chantix). Good luck for anyone who wants to quit! It's a one day at a time thing.

On the question of smoking, it isn't just the cancer risk but the fact that it makes you damn sick. I couldn't lie on my back until I stopped smoking; I would get an enormous headache. Smoking is one of those things that's just so plainly injurious to health generally you don't need the threat of cancer to encourage you to stop.

money4me247 said:

  I have not been able to find any study that shows that the majority of people can reliably identify the differences between lossless and lossy in a controlled blinded test. If someone has a link to such a study, I would be interested in reading.
 
@Sonic Defender, I agree with most of your points until you brought up smoking and cancer. Smoking is a very serious risk factor for lung cancer. Of course, not all smokers will get cancer, but its strong association with lung cancer cannot be denied. For anyone who is struggling with nicotine addiction, try using e-cigs to cut back. You still relieve the oral fixation and hand-mouth habit, and you can gradually taper down the nicotine dose. While the safety of vaping has not been studied at all, I think it is a good way to try to gradually stop if cold turkey didn't work. It ends up being less expensive in the long run as well. Otherwise, there is patches, gum, bupropion, and varenicline (chantix). Good luck for anyone who wants to quit! It's a one day at a time thing.

Click to expand...

I totally agree, my point wasn't that there is any real doubt about the link between smoking and cancer (among many other effects on health). My point was that despite this we still can't claim to have 100% irrefutable scientific evidence of the link in humans. It is virtually impossible to get there. I say that in response to the proffered AES study that was being presented as absolute evidence. Even if that study showed a significant result, that by itself is still not conclusive. I just had a friend die of lung cancer (I quit about 25 years ago), my parents have horrible health due to smoking so yes I didn't want to seem like I am a smoking cancer link skeptic, far from it. Hope that clears this up.

sonic defender said:

  I totally agree, my point wasn't that there is any real doubt about the link between smoking and cancer (among many other effects on health). My point was that despite this we still can't claim to have 100% irrefutable scientific evidence of the link in humans. It is virtually impossible to get there. I say that in response to the proffered AES study that was being presented as absolute evidence. Even if that study showed a significant result, that by itself is still not conclusive. I just had a friend die of lung cancer (I quit about 25 years ago), my parents have horrible health due to smoking so yes I didn't want to seem like I am a smoking cancer link skeptic, far from it. Hope that clears this up.

Click to expand...

No worries. Understand your point now. Sorry to hear about your friend, but congrats on quittin!!!

money4me247 said:

  I have not been able to find any study that shows that the majority of people can reliably identify the differences between lossless and lossy in a controlled blinded test. If someone has a link to such a study, I would be interested in reading.
 
@Sonic Defender, I agree with most of your points until you brought up smoking and cancer. Smoking is a very serious risk factor for lung cancer. Of course, not all smokers will get cancer, but its strong association with lung cancer cannot be denied. For anyone who is struggling with nicotine addiction, try using e-cigs to cut back. You still relieve the oral fixation and hand-mouth habit, and you can gradually taper down the nicotine dose. While the safety of vaping has not been studied at all, I think it is a good way to try to gradually stop if cold turkey didn't work. It ends up being less expensive in the long run as well. Otherwise, there is patches, gum, bupropion, and varenicline (chantix). Good luck for anyone who wants to quit! It's a one day at a time thing.

Click to expand...

Oh yes, the e-cigs (with nicotine) are an amazing way to quit smoking. The hand to mouth habit is EXTREMELY important to consider. In fact, I quit by using really fat toothpicks that allowed me to figit with them  both hand and mouth. That is why a lot of smokers struggle quitting, they don't take into account the tactile needs and the brains love of the repetitive motions. The inhaled and exhaled vapour is also important hence why e-cigs help so much. Smoking is a complex addiction with more than simply a dependency on a substance, it is a lifestyle/activity based addiction and needs a holistic approach to treat (plus some willingness to suffer minor discomfort).

I thought I'd weigh in very briefly on this.  I was able to hear the difference and pass the test on Tidal but on stereo.  Not my headphone rig.  I have a nice pair of headphones (HD650) but I am only running them through a HiFiMAN EF2A for both the DAC and the amp.  I know it isn't the best set up but it sounds pretty good for everyday listening.  On the Tidal test, I'm sure you all had the same experience as me...I could discern a difference but I couldn't tell which one is better.  That's frustrating.   On the big system I had the same issue but I was running it out of the computer (PC) on USB  into a nuforce usb/coax converter and into a PS Audio Dac.  I thought, when I originally set it up that that sounded better than the USB straight into the DAC but after being frustrated with the test I switched to the straight USB and it is, in fact clearer.  It was NOT a big difference. 
 
On a daily basis I listen to Grooveshark and it is, of course, no where near Tidal quality and, I think, it is warmed up a bit but I think it sounds great.  The only time I notice anything negative is when I play a song I actually have on CD and then I can tell what I'm missing.
 
That is when it leads to enjoyment factor.  For just typical listening, especially when doing other stuff, I'm sure I could listen to 128kb files all day long. But when you know you are missing something it is hard to suspend that thought when you aren't distracted.  So, I want to be able to get the most out of what I got.  That motivates me.  If it doesn't motivate you, other than "having to have the best", I see no real point in paying for the hi-res files.  As soon as it gets to the point I can't hear a difference I'll give up but every time I have used or auditioned "better" equipment...it has in fact sounded better. *
 
It's sort of like driving your beater all week but having the Vette in the garage.  Do you need a Vette?  No.  Would you rather drive a Vette...yes...yes you would!  (Feel free to insert ferrari, lambo etc...I'm not selling corvettes).
 
*Not true of every file.  I took a Jake Bugg CD with me to audition a much better amp than I have on a system many tiers above mine and it made that particular CD sound like hell.  I still like Jake Bugg and it still sounds good in my system. 
 
Anyway, just thought I'd share.  I'm listening to Tidal on the trial right now and it really does sound good.  $20 a month good...probably not.
 
Ok...so, that wasn't as brief as I thought it would be. 

You need to repeat the experiment to at least I would suggest completing 30 trials. If you can identify the difference at least 80% of the time you might actually be hearing the difference. I'm not sure how the Tidal test is set up, but I would be shocked if it is at all valid. After all, they are trying to "guide" you into confirming a difference as that is how they will make money. Cynical I know.