Normally I don't read newsletters, let alone make a big deal of of them. But this one piqued my interest from reading the subject: Do You Have The Ears Of An Audio Producer?
Really fun short test to see how good you are at spotting a problem:
Very interesting but IMHO they hurt the idea by telling you what to listen for.
For example, having Golden Ears has nothing to do with what frequencies your ears can hear but about what you can find/notice that is wrong or could be better and the ability to differentiate things others might not even realize are there. If such a person points such a thing out to you, you'll likely immediately hear it when they point it out.
__________________ Music is what feelings sound like.
Last edited by karbomusic; 02-09-2017 at 01:54 PM.
Interesting, but also puzzling. The first test I had all 6 right, but strange enough I could not at all hear any difference between the mp3 and wav in the second test.... What does that tell about my hearing? Or about my ears?
What does that tell about my hearing? Or about my ears?
That your hearing is perfectly normal. AFAIK, no one has ever successfully been able to reliably distinguish MP3 @ 320kbs vs uncompressed WAV under normal conditions. You'll find plenty who tell you they hear it because they have the most awesome monitoring system eva but AFAIK zero can actually demonstrate it beyond chance. If memory serves, to truly demonstrate you can reliably tell, you'd have to get it right something like ~99.8% of the time.
__________________ Music is what feelings sound like.
Last edited by karbomusic; 02-10-2017 at 10:34 AM.
It is indeed difficult. I thought I could hear it. I heard some difference in highs in one clip. Very slight though. I was correct, but I had all others wrong. I listened with headphones.
It is indeed difficult. I thought I could hear it. I heard some difference in highs in one clip. Very slight though. I was correct, but I had all others wrong. I listened with headphones.
I only got 2 right for the quality one, but I did hear I had made a mistake on one of them. 320kbps is very close to WAV, 128 is a bigger difference, but a number of the samples they had were from older sources which are already a little bit hazy by nature, so, to me, it was a bit of a poor test.
Or, it exposed flaws in my listening setup. I use ATH-m50s for my DAW. Maybe a quiter room with nice monitors would be better, but these headphones are pretty good. I am not sure headphones vs monitors makes such a difference for that test, but certainly quality of your listening equipment would. Maybe my interface is not up to snuff also, that could be one reason for sure.
but a number of the samples they had were from older sources which are already a little bit hazy by nature, so, to me, it was a bit of a poor test.
That shouldn't matter really because it isn't the quality of what is being converted but the conversion itself. I will say if you could find the difference it might show up in splashy cymbals etc. first but again, I *could be wrong* but I thought there was a pretty big monetary prize awaiting anyone who can reliably tell the difference.
I'm guessing that there is some psychology going on, psychoacoustics and things like that.
This is somewhat depressing as well:
In a double-blind test by professional violinists, most couldn't determine — by sound alone — which violin was an original Stradivarius and which was a modern instrument. Above, a 1729 Stradivari known as the "Solomon, Ex-Lambert."
Interesting, but also puzzling. The first test I had all 6 right, but strange enough I could not at all hear any difference between the mp3 and wav in the second test.... What does that tell about my hearing? Or about my ears?
It doesnt tell you anything about your ears. It does tell you everything about the differences between a good mp3 and a wav-file: there is none, listening-wise.
__________________
"Dear Americans... I told you so. Sincerely, your Aldous Huxley"
The best way to notice differences this small is to listen to sample A for maybe a week, then sample B replaces it without your knowledge. Most blind tests can't cover such a scenario but blind tests are usually debunking people who are already claiming they hear a 'night and day' difference by simply switching between the two over the course of a few seconds.
Always beware when you search around for stuff like this and see people who already know which sample they are listening to then use terms like more 3-D and spatial etc. That's a big fat red flag.
__________________ Music is what feelings sound like.
I know about the very hard to hear difference between wav and 320kbs but there is a noticeable difference at 128kbs. When mp3 was new, I could hear definitely the difference between CD and mp3 at 128kbs: less dynamics and high frequencies. Maybe the current Lame decoder is much better these days than when mp3 came out?
That shouldn't matter really because it isn't the quality of what is being converted but the conversion itself. I will say if you could find the difference it might show up in splashy cymbals etc. first but again, I *could be wrong* but I thought there was a pretty big monetary prize awaiting anyone who can reliably tell the difference.
Ya, the highs are really where I notice it more. I know what you mean about quality, but for me, modern music, and also higher quality, sound more crisp and snappy, and well defined, whereas older stuff, for whatever reason, already has some sort of haze, or blur introduced, whether that's from the analog console, or tape, or whatever else, I don't know, but it certainly makes it a lot more difficult for me to hear the differences. It's already pretty subtle enough.
I think most people would not even be able to discern 128. Meaning average joe on average hearing equipment. Audiophiles and engineers definitely would, but even though it's definitely audible, it's not the most obvious thing.
The test you linked, on the first test introduced a lot of clipping on the 6db up, which kind of ruined the test.
I don't find volume is a good test either, because the volume my interface is set to, will affect how I perceive changes in volume. Since that part is not controlled, the test is not very accurate. Also I don't have a very quiet room to try it in, and everyone's gear is not consistent also. That said, I know my conditions are not the greatest, but I found more precise than 0.5dB a bit hit or miss, which is not such a great thing I guess. And even then, I found it a bit tough. I find hearing whether it goes down or stays constant is more difficult than if it goes louder.
but I found more precise than 0.5dB a bit hit or miss, which is not such a great thing I guess.
When I tried a few weeks ago, I was able to nail .5 but .2 just wasn't happening. I haven't tried at home in my more proper studio yet but I don't think I'll get 'that' much better because .2 is farking small. The only thing I'll add is .2/.5 is more valuable when you are adjusting something critical in a mix vs just telling if it is louder softer by itself due to masking etc. IME anyway.
__________________ Music is what feelings sound like.
Do the MP3 test in the second link, if you haven't.
Oh, that surprised me.
I got 4/6.
I wasn't expecting to hear any difference between 320kbps mp3 and uncompressed wav,
but I did.
The ones I didn't get right was
-the old classical (that I expect has limited top)
-and the Jay Z that typical has much distortion in the mix.
Neil Yong was just luck, as I really couldn't hear any difference there.
Laptop with my chepo earbuds.
Hm.. either the test is flawed in some way or I need to adjust a few of my audio-truths.
...which is not statistically meaningful, in case you didn't know. There's a 34% chance that happened purely randomly. Even 8/10 doesn't meet the standard "statistical significance" test (5.4% chance of randomness for 8/10), although that would be enough for me to believe something was going on. 4/6 might be provocative enough to inspire further testing, but it doesn't really show anything. For fun, I just flipped a coin, and as it turns out I happened to get 4/6 as well. :-)
Combine this with the dynamic where people like yourself who test well are more likely to post in this thread, and those who don't test well are less likely to, and I'd caution anyone from drawing too much insight from these numbers. :-)
When I tried a few weeks ago, I was able to nail .5 but .2 just wasn't happening. I haven't tried at home in my more proper studio yet but I don't think I'll get 'that' much better because .2 is farking small. The only thing I'll add is .2/.5 is more valuable when you are adjusting something critical in a mix vs just telling if it is louder softer by itself due to masking etc. IME anyway.
Ok, I don't feel so bad then. even 0.5 I found was difficult, especially for going down in volume. Going up in volume was easy enough for me at that level.
hmm i'll credit my avantones on making first test far too straightforward - but have done the 'compressed music' test before and doing so again, that one is still hard!
track dependant q often, the perry one, pure guess to me, others you get a bit more smooth sizzly bits or tiny bit more depth.
a 128 can be a bit warmer than the others so can (track dependant) be deemed 'preferable' if the 320 and wav are indistinguishable.
it does mean that you could have messed up your hearing technically over the decades and still make a decent decision !
it does mean that you could have messed up your hearing technically over the decades and still make a decent decision !
Little known fact that some of us with high frequency tinnitus (I don't have much but some) can be more sensitive and aware of HF content differences because it interacts with the ringing. I don't think you need to credit your monitors though on the first tests because those aren't about hearing ability but rather listening ability aka know what to listen for and how to go about it. Those should be discernable, again if knowing how to judge it, on fairly crappy monitors.
The nice part about that first set of tests, is that if someone misses it, someone who knows could explain it to them and they would then get it and be able to recognize it.
__________________ Music is what feelings sound like.
never ever liked cutlery clattering, physically hurt!
my left now has a little constant ring, though i see it as a 'mask' - turn that freq up and it still there, though it is slightly depressing at the same time.
like the daynot so long ago when my lady said 'oh can you hear the crickets?' and i looked around dumbly.
anyway if things get really bad we always have the 'beethoven' card to play, or you get little johhny to say 'too bright' whilst you twiddle, (hey there will be an app for that soon anyway!!)
...which is not statistically meaningful, in case you didn't know. There's a 34% chance that happened purely randomly. Even 8/10 doesn't meet the standard "statistical significance" test (5.4% chance of randomness for 8/10), although that would be enough for me to believe something was going on. 4/6 might be provocative enough to inspire further testing, but it doesn't really show anything. For fun, I just flipped a coin, and as it turns out I happened to get 4/6 as well. :-)
Combine this with the dynamic where people like yourself who test well are more likely to post in this thread, and those who don't test well are less likely to, and I'd caution anyone from drawing too much insight from these numbers. :-)
The thing is:
I believed I heard a difference for the 3 first
and picked the right ones.
For the 3 last I didn't hear any difference,
and got the statistical typical 1/3.
I'll ABX any/all of the first 3,
if you'd care to prepare files for me (flac 44/16 or similar).
I'm looking forward to the "we will write a glowing, feature-length article about you"
Oops - I withdraw my stats comment -- I hadn't visited the link and assumed you meant that you got 4/6 on an ABX between two files. Not the case, sorry about that.
Quote:
I'll ABX any/all of the first 3,
if you'd care to prepare files for me (flac 44/16 or similar).
Cool... I prepared all the files, in case you or anyone wants a more blind version of the test.
PLEASE NOTE: if anyone tests these, please do not report your choices in this thread, to avoid the "gathering consensus" effect. Once all interested parties have listened and privately recorded their guesses, we can all report at the same time.
The more tests per file set, the more meaningful the results. I'd suggest 10 or more. And the stated goal here is "find the .wav file amongst the three options", so I set the web test up as a shootout, as opposed to a three-way ABX. (An ABX would be "given an unknown file of the three, identify which one it is: 128, 320, or wav".)
Quote:
I'm looking forward to the "we will write a glowing, feature-length article about you"
Okay I tried the test using the latest link.
I got 100% on all tests
Does this mean I succeeded?
There is nothing to say what to listen for on the chart?
I just listened for what I thought was the best sounding track.
Am I an idiot?
Okay I tried the test using the latest link.
I got 100% on all tests
Does this mean I succeeded?
There is nothing to say what to listen for on the chart?
I just listened for what I thought was the best sounding track.
Am I an idiot?
Perhaps what you did was use my link and just take each test just once, which is not meaningful: it will show you that you chose that file 100% of the time, which of course you did. You need to click "next shootout" and repeat 10 or 15 times per song, and after that "show results" and it will show you the percentages over all the trials. My link will not reveal to you how well you did (unless it reveals to you that you can't find the WAV because your choices are all over the place): that will be revealed after all interested parties have taken the test.
Thanks -- that web ABX tool represents the most development work I've done for the least reason; I don't think a single person has used it. :-) Glad it could be put to use here.
Ah, that link. As for what to listen to, they make it pretty clear on each test, no? E.g. "Which clip has phase issues? Phase issues can make you feel like your head is swimming ... and have serious consequences for the listener." Not sure what 'chart' you were referring to, though.
That is odd I tried the link again and I got quite the different web page
than that which I first trialed. I then tried the second link (first post of the thread) and got another web page.
This present two seem like the pages I should have got when I used the first link.
Will try again tomorrow
About the compression example, sometimes, I find that more compression works as an effect in some cases. Here is a good example, around 0:50 and onwards:
And Also The Trees - Your Guess
In many cases, it would be too much compression. But in this case, it seems to give the vocal more character and somewhat intimate, dark, raspy, silvery, metallic, retro-like sound. I am not sure if it is intentional, but I really like it as an effect in this song.
Last edited by Flaneurette; 02-13-2017 at 01:02 AM.
In a double-blind test by professional violinists, most couldn't determine — by sound alone — which violin was an original Stradivarius and which was a modern instrument...
The BBC tested that out in the fifties. Most musicians aren't very good at spotting good sound. Or problems with the sound quality.
Engineers were very good, but found a different problem each time, as a group.
The best test persons to judge the quality of sound were frequent concert visitors. AKA the audience...
Humbling, isn't it?
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
Since no one has reported taking the test, I won't bother posting the results. If anyone in the future finds this and wants to try their hand at the shootout (link above), PM me or post here and I will reveal the answers.
The BBC tested that out in the fifties. Most musicians aren't very good at spotting good sound. Or problems with the sound quality.
Engineers were very good, but found a different problem each time, as a group.
The best test persons to judge the quality of sound were frequent concert visitors. AKA the audience...
Humbling, isn't it?
I don't know man. Deceiving ourselves seems a human characteristic. The human brain, memory, perception, recollection, it's all very fallible. Science shows that we can even create memories out of thin air. I guess it is humbling to know that we're imperfect, fallible and just another animal but with a slightly better organized brain.
Since no one has reported taking the test, I won't bother posting the results. If anyone in the future finds this and wants to try their hand at the shootout (link above), PM me or post here and I will reveal the answers.
Well, proves your point.
I had a second go, and, I can't tell the difference..