The Audiophile Challenge

[cruising for burgers]

http://www.machinadynamica.com/machina51.htm

 

MACHINA DYNAMICA

 

NEW!! Dark Matter Optical Coating for CDs

 

What You Can't See

 

 

Dark Matter Optical Coating is the first audio product to address the (generally unrecognized) problem in CD playback produced by invisible, infrared scattered laser light inside the transport compartment. Dark Matter is a permanent optical coating that is applied to the LABEL SIDE of CDs using a spray bottle. The liquid evaporates quickly leaving a thin, clear optical coating. Dark Matter absorbs background scattered infrared light that could otherwise make its way into the photodetector as noise. Since the laser scatters light all around inside the CD transport compartment, the CD label is an effective location for Dark Matter.
Dark Matter Optical Coating is a transparent, emerald green liquid that comes in a 3 oz spray bottle, sufficient for treating 125 CDs. Place a Bounty towel under the CD in case of overspray. Hold the bottle 3 inches from the CD, spray Dark Matter onto the LABEL SIDE using 4 pumps of the spray bottle. Allow Dark Matter to dry for 5 minutes or until completely dry. Drying time depends on temperature and humidity of the ambiant air. Dark Matter dries clear, forming a long-lasting thin film. Dark Matter can be applied to the top surface of the CD tray for additional benefit. Dark Matter is designed ONLY for CDs and will not work for DVDs, SACDs or Blu Ray discs.

Edited March 17, 2018 by THIS IS MICHAEL JACKSON

[cruising for burgers]

If you really can't tell when it switches, congratulations! You're free! You can encode all your albums in 128kbps and have peace of mind. But if you notice the switches... well, maybe there's a good reason for FLAC encoding after all...

this is not a fair comparison, compare it at least with a 320kbps mp3... if you can tell the diference, then yes you can deduce that there's a good reason to use flacs...

[KovalainenFanBoy]

[TubularCorporation]

I don't consider myself an audiophile (because it's kind of synonymous with "hearing with your wallet") but the difference between 320kbps mp3 and CD audio is night and day. 

 

Doesn't stop me fro listening to ow bitrate Youtube uploads on laptop speakers sometimes, though.  Just because you can hear a difference doesn't mena you should get insane about it.

[cruising for burgers]

i just did the OP test and while i usually can tell the diference between 128 and 192 (mainly cymbals and hi-hats), in this example the little differences i can spot are mostly in the white/pink background noise/tape hiss, i cannot spot any diferences in the transients and in the instruments timbres whatsoever, there's only 3 instruments playing and i repeated the test 9 times, 3 times for each instrument, nothing... even the cymbal where i though i would notice the most evident diference i could not... D- for me... :(

 

btw, i dunno if i read somewhere or some teacher commented that 320kbps mp3 are actually worse than 192... it is quite possible since the mp3 conversion process is not as simple and linear as a wav conversion...

Edited March 17, 2018 by THIS IS MICHAEL JACKSON

[eugene]

I don't consider myself an audiophile (because it's kind of synonymous with "hearing with your wallet") but the difference between 320kbps mp3 and CD audio is night and day. 

the only people who claim that the difference between modern high bitrate mp3s and lossless is night and day are the people who never did a proper test.

 

here, do this one, 10 trials for each track please: http://abx.digitalfeed.net/lame.320.html

 

i'm pretty sure you'll give up within a minute or two after realizing that those are pretty much impossible to tell apart. i've been super into this topic in recent years and i've yet to see a person who passes a proper abx test when comparing anything above 192kbps mp3 vs flac.

also mp3 is not the cutting edge of lossy compression anymore, no one will be able to tell apart a 96kbps opus encode vs flac either.

Edited March 18, 2018 by eugene

[TubularCorporation]

EDIT 2: I forgot that I actually had the "loudness maximization" on in Windows sound manager (it's useful if you're watching stuff on Youtube with just laptop speakers, which I was doing last night before bed) for the whole test and also the driver for the on board sound on my laptop ALWAYS applies EQ to the output that makes it sound better through the internal speakers - but it still applies it even if you use headphones, there's no way to defeat it - so my results were completely invalid. I'll do it again tonight or tomorrow with a proper audio interface and more time to spend.

 

EDIT:  I'll just leave these here, too:

 

https://phys.org/news/2013-02-human-fourier-uncertainty-principle.html

https://tapeop.com/blog/2012/04/11/problem-bing-and-why-neil-young-right-about-sound-/

Edited March 18, 2018 by RSP

[eugene]

 

I was right in the low to mid 60% range.

 

that's really not far from guessing. for proper statistical significance you need to identify at least 9/10.

 

regarding taxing the algorithm, choose a track that you think is taxing and i'll find it in flac, i'll cut an mp3 sample from it for you to try to abx.

there are so called killer samples that do make mp3 a bit more obvious, but they are very rare to come by in actual music.

[eugene]

regarding the links you added after edit.  the first one - i don't care. the second one is just bollocks. there's no better way to reliably test whether you can differentiate between two samples. abx tests completely control for all other possibly intervening factors. no one prevents you from doing abx tests over long periods of time either. the author's assumption is also idiotic, if you can't tell the difference in a proper test then somehow you will be able to tell the difference with a much less focused, casual listening? nonsense.

Edited March 18, 2018 by eugene

[auxien]

I've never cared much about these, and I really don't think most anyone should. If you have a higher quality file or source, you should listen to that if you can. Compromise when you have to, if it makes sense. Are you gonna listen to it on your phone's speakers, or through your okay headphones through your phone or iPod? Sure, cut the quality down, probably not a big deal. If 320kbps mp3s give you 10 extra hours of music on your iPod and you REALLY want that extra variety, you likely won't notice unless you're bringing it to listen on a decent stereo anyway, so go for it. I've always preferred to have higher quality when possible, at the sacrifice of extra storage. Would I notice on 50% of the music I listen to? Nope. Maybe I'd notice on some of it though, and I don't need to bring EVERYTHING with me everywhere I go, so I opt for higher quality when I can. I'm definitely not an audiophile, though if I had money to blow, I probably would splurge in some areas (not $10,000 speaker cables though).

 

Basically the argument comes down to space available to store or perhaps bandwidth to download or stream (streaming's a bit of a tangent though). Storage space is cheap as fuck if you're stationary so why not listen to FLACs? Or even better, straight from a CD. Highest quality available, go with that if you can. If you can't, as in you're mobile or whatever, compromise, okay. You probably won't notice, and you're probably an idiot/asshole if 'the shitty quality mp3 is ruining your experience' and that's why we laugh at memes about people like that. I mean yeah, in 1999 the mp3 quality you downloaded was pretty shit, but things aren't like that now. Is Spotify streaming at lower quality going to sound worse than a WAV or from a CD? Well fuck, duh. What's the discussion?

 

That said, I nailed that damned The Killers track. The James Blake one is harder, and if I even finish the test I know I'm going to do worse on it. My ears are fatigued anyway from working on a track all morning. Having to go off of tracks you've not necessarily ever heard is fair and all, but imo is part of the problem with these discussions. There's some music I have that I know sounds better from the CD than from the downloaded 320 mp3. A lot of music I have? Couldn't tell the difference with a gun to my head. Then there's other issues of shit being mastered differently for iTunes vs CD/etc., it's just a big messy waste of time (I've wasted like 20 minutes at least now so that's enough).

Edited March 18, 2018 by auxien

[Ivan Ooze]

i always buy 320kps mp3 now, it's all about the tools you use to play it man

[cruising for burgers]

yeah most people who defend the "fact" that vinyl sounds better than cd don't even realize that the masters/mixes are completely different depending on the format...

[Candiru]

I put a piece of string through a speaker, tie it to a tin can in the next room and listen to it there. They call me Professor Audiophile because I fux wit sounz

[TubularCorporation]

I'm still down to do a proper A/B test but it's going to have to be on downloadable material, because of the limitations of the onboard audio on my internet-connected laptop (which doesn't have a Firewire interface, so I can't use my interface with it).

 

If someone wants to upload five suitable CD quality (to avoid the extra layer of quality loss from wordlength reduction and resampling between a 24/48 or better file and an mp3), preferrably stuff that taxes the mp3 algorithm, dynamic, modern classical recordings tend to be good for this) I'd be happy to do 20 passes of each with the Foobar2000 ABX Comparator plugin and post the results.

 

 

EDIT: in the mean time, here's the results of taking a 16/44.1 mix of one of my old tracks, converting it to a 320kbps mp3 with LAME at its highest quality encode setting, then converting it back to 16/44.1 uncompressed and nulling it with the original file.  It only nulls to -23db, and is enough to easily hear through even the crappy laptop speakers I'm playing it on, t the point where I can hear the drum machine in particular IN THE NULL FILE.

 

https://www.dropbox.com/s/iw19jqzb1cbwb8a/CDDA%20nulled%20with%20CDDA-mp3-CDDA.wav?dl=1

 

 

And again, I definitely don't consider myself an audiophile, I think the whole "audiophile community" is mostly full of marks and con men (although the people who design and build their own DIY turntables are cool), and I really don't sweat this kind of stuff much in practice.  But the mp3 vs CD audio comparison is not subtle to my ear, certainly nowhere near placebo/audiophile subtle, and that quick null test certainly supports my experience.

Edited March 31, 2018 by RSP

[eugene]

than null test doesn't really prove anything, of course there is stuff that gets lost in the compression, but it doesn't mean that you'll hear that difference in a blind comparison. the people who program lossy codecs are smart, their whole goal is too fool the human hearing by masking stuff and implementing other psychoacoustical tricks.

 

 

But the mp3 vs CD audio comparison is not subtle to my ear,

but you haven't done any proper ABX yet, so what exactly are you basing this on exactly?

 

choose ANY track(s) you want and i'll chop you a 320mp3 and flac samples from them.

[cruising for burgers]

“High Definition Vinyl” Is Happening, Possibly as Early as Next Year
With a new $4.8 million investment, an Austrian startup says it could have “HD vinyl” in stores by 2019
 
https://hdvinyl.org/

• It offers better sound quality, higher frequency response, 30% more playing time, and 30% more amplitude than current vinyl records.

• It also eliminates the toxic chemicals currently used in the vinyl mastering process.

• We can also correct the tangential/radial error (a lacquer is traditionally cut at a tangential angle, but most turntables read in a radial angle, so the needle is constantly tilting).

Edited April 13, 2018 by THIS IS MICHAEL JACKSON

[auxien]

^ but will it still make the audio sound demonstrably shittier in the way that vinylheads love, or will it just be shittier in a different way that they'll claim isn't as pure?

 

 

;D

[cruising for burgers]

it will probably sound cold as ice... :^)

[usagi]

some of the posts itt makes me wanna transcode my entire collection to 64kbps and fart it through a laptop speaker.

[Candiru]

When I buy cantaloupes, I shake them and listen for liquid noises to see if they are ripe. This sound is in 128kabips and if I use apple earbuds this is throughly satisfying to my well seasoned ears. No need to buy any of that fancypants high falutin' Skull Candy malarky. 

[Lane Visitor]

i`m so much of an audiophile that i wear earplugs all day and listen to the music in my MIND

Flol

 

I'm so much of an audiophile, I had the top ear doctor in Germany perform surgery on my ears while I was coming out of the womb to disable my hearing permanently so that the only memories of sound I have through out life are the pure sound of my mother's heart beat and internal processes.

[manmower]

I aborted eugene's test before even responding to the first trial, having immediately realized I'd be guessing.

 

Most of my listening time is on the go with LAME V4 encodes of my lossless collection. Never found a track (including Autechre, who feature(d) among the LAME killer samples) where I could tell a difference, and I mean ABX testing at home.

 

I don't know whether my hearing is damaged (plenty of reasons why it could be) or in the normal range for someone my age, but I haven't really noticed other people hearing more than me in everyday life. Blissfully tin-eared is my guess.

[TubularCorporation]

than null test doesn't really prove anything, of course there is stuff that gets lost in the compression, but it doesn't mean that you'll hear that difference in a blind comparison. the people who program lossy codecs are smart, their whole goal is too fool the human hearing by masking stuff and implementing other psychoacoustical tricks.

 

 

But the mp3 vs CD audio comparison is not subtle to my ear,

but you haven't done any proper ABX yet, so what exactly are you basing this on exactly?

 

choose ANY track(s) you want and i'll chop you a 320mp3 and flac samples from them.

 

I've done plenty of ABX tests over the years on this and other things, just not the one on that specific page you linked.

 

I don't have anything to prove to myself here, and again I'm not an "audiophile" at all. If you choose and supply some stuff (if I'm going to do this I want to do it on your terms not mine, and that includes selecting and encoding, both to preserve the integrity of the whole thing and because I really don't feel like expending any effort on this).

 

I'd rather do a forced AB test than an ABX since it's generally more accurate for these kind of things, but either way. I'll probably treat an ABX as forced AB anyhow.

 

 

Regarding the psychoacoustics of lossy compression, we're talking about 25 year old technology based on 60+ year old research and aimed at getting acceptable sound quality at file sizes that are friendly to dial up networks, and it did that.  Even expecting it to be inaudible doesn't make sense. And again, the linear mathematical models it was based on are old and flawed. The time and frequency domain resolution of human hearing is far more sensitive than the mathematical models and lab research from the 1940s and 50s that people are still basing their assumptions about human hearing on could capture.

 

Of course things that are "inaudible" affect what we can hear, that's basic acoustics. Whether it matters is more a matter of personal taste.  For me it doesn't most of the time. Here's an example - not quite analogous to the mp3 question but relevant:

 

 

Did you hear that?  The sound sources Oliveros used were all well above 20kHz.  I don't know the exact frequencies involved but I know a lot of it was created by tuning sine oscillators so that they beat against the bias tones in the tape machines she was using, which would put things in the 120kHz to 140kHz range.  I of IV and her other pieces from that period are a great example of the fact that, yes, even sounds that in this case were 5x - 6x higher than the theoretical limits of human hearing produce plenty of content in the audible range when they interact. It's not rocket science.  Whether it's musically relevant is, again, a matter of taste and for me personally it's not a huge deal.

 

 

Anyhow, post some clips here and I'll have time later in the week to AB them and post results.  Maybe I won't hear a difference after all! 

[eugene]

as i said, choose ANY track(s) you want and i'll cut you a lossless and lossy samples out of them for you to ABX.

 

 

Regarding the psychoacoustics of lossy compression, we're talking about 25 year old technology based on 60+ year old research and aimed at getting acceptable sound quality at file sizes that are friendly to dial up networks, and it did that.  Even expecting it to be inaudible doesn't make sense. And again, the linear mathematical models it was based on are old and flawed. The time and frequency domain resolution of human hearing is far more sensitive than the mathematical models and lab research from the 1940s and 50s that people are still basing their assumptions about human hearing on could capture.

 

then logically it would be easy to ABX, yet there are almost no examples of people successfully abxing high bitrate mp3s and other more modern codecs. so maybe the codec programmers know what they're doing after all?

 

regarding the olveros example - imd distortion and aliasing artifacts (though less relevant here) do indeed exist, so what exactly is your point?

[mcbpete]

Apparently in that Olveros piece it's 11x >20khz signals beating against a 1 Hz signal ( NOISEGATE 11 - http://www.stalk.net/paradigm/pd04.html ). Beat frequency is equal to the difference in frequency of two waves so the higher pitched oscilators are all being pulled down significantly by the 1Hz one rather than them all adding up to make frequencies in the 100+ Khz range.