ABX Testing: 128k AAC vs Lossless

[@cukedaddy]

I bet most of us wouldn't be able to notice if somehow all of our music files were switched without our knowledge.

[@aboyapart]

best comment! would love to do that someday

[@aboyapart]

+xhoggerx129 anyone can say "i can"...actually being able to do so is another thing. So go ahead and reply to me saying louder "I CAAANNN". It would still be insignificant to me.

[@cukedaddy]

+xhoggerx129 But... Scientific tests are reproducible.... This test you did wouldn't be

[@aboyapart]

+xhoggerx129 A better test would be to do the ABx testing between the mp3 and aac directly. Doing this, would be more accurate than deducing the distinction from indirect experience. Also doing the ABX tests 1 time for each lossy version doesnt provide reliable results, more so in your case... where you didnt even achieve at least 9/10 at any test. This is the required number for statistical significance. 8/10 is boderline random guessing. annnnd.. as I always point out: to me it doesnt matter to simply differentiate lossy vs lossless. For example saying "I was able to differentiate lossless vs lossy because I could detect a hiss in the lossless file which wasnt present with the mp3.] If every human can do 10/10 ABX test based on that hiss...that would only mean that the files are different....but who the f*ck would prefer the lossless with the hiss?! lossless would be inferior in such a case...so why bother discussing "i was able to distinguish between the two"? Discussions about lossless vs lossy are significant to me when the differentiation is based on some "superior" aspect that one file posseses over the other (in particular, we discuss it to determine if lossless is better than lossy). By "superior" I mean that it provides a more enjoyable audio experience. "hiss" wouldnt be such an aspect for the lossless over the lossy. In that case lossy would be superior both for not having the hiss and occupying less memory space. so please...dont join the club of people who think that they are special for such stupid things.

[@aboyapart]

+xhoggerx129 one (1) typical ABX test requires ten trials. so we are talking about dofferent "ten's". I know you did ten TRIALS for each lossy version. What I said is that you should do the test more than once in order to be reliable, i.e. if you can score 8/10 more than just once for mp3 , it would suggest to me that you are unto something. Othewise it could have been random guessing for that one test. Anyway, if you care to read the last part of my comment, you would understand why I dont actually care about your scores. I care about the aspects that allowed you to distinguish the files, I dont actually care about the mere ability to distinguish. You should read the coment, I explain what I mean there.

[@johnbradshaw5900]

ABX or similar double blind testing is the ultimate form of audio testing for the simple reason that it identifies whether or not there is even a discernible difference between whatever is being tested. I suspect a reason so many audiophiles hate it is precisely because many of those allegedly night and day differences disappear in a double blind test. Even where differences can be discerned, if it takes intense effort to do so and triggers stress anxiety then it indicates such differences to be irrelevant in the context of listening to music.

[@maka8551]

we hate it because the albums they choose are purposely poorly engineered lol. and no one who takes the test is even in the correct environment idiot. this is like driving a lamborghini in the fucking hood

[@oldskool4572]

The fact one would have to concentrate really hard to spot any differences (if there any) says it all really.

[@maka8551]

yeah people who respect music do, listen to your top 40

[@oldskool4572]

@@maka8551 Very gown up response. I listen to all kinds of music, especially from the 60's through to the 90's, so current moden music doesn't hold much interest to me on the whole. And I have a very good ear with some very good equipment on hand. But you carry on being childish by all means.

[@eLicky]

That was fun, I like the way you made it casual, rather than something overly serious in a room with padded walls. Good entertainment and educational.

[@lachlanlikesathing]

Haha. I think we all lack the authority to make it serious. I'm not out to prove anything definitively but if a fun video makes people more likely to try it for themselves, all the better!

[@maka8551]

they had to try their hardest to downplay lossless so ofc they didnt use an optimal room

[@Eric.Choudor]

"The press indicates that an AAC file in 128 kb/s is equivalent to an MP3 file in 256 kb/s, but the article is not accompanied by any concrete experimental results"

[@Eric.Choudor]

"According to Apple, listening experts have found it nearly impossible to distinguish AAC audio files compressed at 128 kbps (stereo) from an uncompressed audio source by ear"

[@jasonong23c]

In conclusion, upgrading the gear is more significant than upgrading the audio format for the best listening experience

[@scalpelboy1]

well regarding the comments below that address how the environment in this test is not dead silent, i think it is actually the point of this testing. You know, these kids have the best gear and on a minimally noisy environment they cant reliably tell whether theyre listening to lossy or lossless files. for me who mostly listen to music on my commuting to work, through my ipod and sub 50$ earphone, i am pretty happy with the test result. cheers

[@Evan-up5py]

Props to the kid using the Howl's Moving Castle soundtrack lol

[@ricardoanton-ifiaudio]

Lol! What you did with this video for Audiophiles is like if you make a video showing kids to Santa Claus taking his beard off

[@MA-ck4wu]

Audio files are something else haha. You mean audiophiles

[@maka8551]

they purposely chose poor engineered albums so they couldnt tell the difference lol

[@SaHaRaSquad]

@@maka8551 And here come the poor excuses.

[@Lexington365]

Great video guys! I ran my own test comparing some FLAC and some old 128kb stuff I ripped to iTunes around 8 years ago! I had one track by Tool called The Patient. The FLAC version I had was 24bit 96khz with a bit rate of 2600! I spent maybe 10-15 mins playing the two tracks through a decent DAC and Audioengine A5+ speakers and I honestly couldn't tell any difference. I did the same with Dire Straits Sultans Of Swing and again couldn't tell any difference at all.

[@i0001w]

with better speaker you will till

[@ES-qy2ju]

You'll need headphones to hear the difference much better, and this headphones should have at least a frecuency response of 4 hz -40,000 khz. This is required for Hi Res listening.

[@techgeek1874]

@@ES-qy2ju That's the biggest load of bullshit. First off i have an mdr z1r and a Sennheiser momentum headphone along with a DMP Z1 and a Zx300. There's no difference in the same master at 24 bit or 16 bit or an AAC 128/256 mbps. You can hear the difference if there's a poorly mastered track and the same track with better mastering. High res audio is snake oil bullshit.

[@PetePohjanniemi]

Muataz Alshammari no you won’t

[@e.gundogan8656]

Yeah. Hi-res (24-96) is bullshit. The guy who developed ogg vorbis said it as well. Frequencies beyond redbook-CD (44.1 khz) are absolutely unnecessary unless you're playing your music for your pet moth. The best way to have noticable improvements over the sound quality is getting better headphones/speakers and using them properly, another important thing is getting the right pressings. For example Radiohead is good both on vinyl records and on CD but oasis CD's are just terrible and their vinyl records are a lot better. You're gonna have to find the best pressing of each album. Most of the time an album's vinyl record version has better dynamic range than it's CD version though CD as a format is way more capable than vinyl record in terms of dynamic range and many other things, it sucks that many labels/artists don't use CD's potential.

[@snake8584]

Great video series. Very practical in my opinion. Great for "audiophiles" on a budget.

[@YugShende]

I mostly use FLAC for archival purpose. I believe that if I want to archive something it has to be of the topmost quality. My portables will always have aac/mp3 though.

[@hendok828]

i'm convinced lol thanks! I though I needed to spend a bit more money to get and audible upgrade in music quality...I'll stick with my current setup!

[@MarshMellowChronic]

I'm still surprised at how many people trying to justify keeping their music in flac and stating that your gear was not good enough. Sometimes it just seems like they want to pay more than necessary, they should take this as a breath of fresh air because the space it frees up with no loss in quality.

[@Selrisitai]

What?

[@user-vi3tb3bw5t]

Just because you can't hear a difference doesn't mean there isn't one. It's very subtle, but it's there. There is a small but noticeable difference in overall "oomph" in vocals and overall resolution when I compare flac files to aac. I think if someone loves listening to music with high end heaphones then why not get the best quality version of the music you want to hear...And not ALL songs I am able to differentiate between the two..But certain albums like Love Is the Thing By Nat King Cole or Kind of Blue by Miles Davis...Absolutely noticeable difference between the lossless and lossy counterparts.

[@miffachan123]

J audiofooled

[@user-vi3tb3bw5t]

CloudSound go ahead and believe what you want, probably using earbuds and expect a difference.

[@Cobalt985]

Audio quality is not the benefit of lossless, *re-encoding is*. I keep my music in FLAC so that I can encode to different qualities and formats for every device. This came in handy literally today because I uploaded a lot of music to YT music and because it was in FLAC instead of MP3 it didn’t have to get re-encoded to AAC.

[@Ordo.Corinthivm]

Bruh, let be honest here you can't tell the difference between lossless and lossy audio. People who said they can is either got some superhuman hearing or just too embarrassed and unable to admit the truth that they cannot hear the difference. The lossy audio was mathematically engineered to sound the same so that a human ear unable to distinguish.

[@joeandjoot]

I love the reactions of your friends here

[@shimtest]

this has me interested in AAC since I did a similar test years ago with mp3 and it was very easy to tell the difference

[@treetoon_]

yeah mp3's in 128kb/s are far worse in quality than aac

[@HiFiInsider]

Great video, but I think the chain is not transparent enough to hear the difference.

[@lachlanlikesathing]

Thanks Mike! Honestly while the environment wasn't ideal I have to say that I thought the WASAPI output, the O2/ODAC coupled with the Fitears was some pretty darn serious equipment, both in terms of cost and the respect this gear gets on Head Fi. Tom definitely seemed to perform well with this chain, though not enough to get a statistically significant result, so more tests definitely seem in order. There are definitely some tracks that will trip up 128k AAC (there are a few instances of people reporting good ABX scores on 128k AAC) but I have to say, if you asked an average audiophile before they did this test they would probably tell you that any half decent equipment (let alone $1000+ IEMS) would easily show you the difference between 128k AAC and lossless. What this test demonstrates is that this simply isn't true - it is not easy, and it took Tom 25 minutes and a great deal of careful listening to achieve a non-significant result. You can imagine how poor these results would be if we were using 160k AAC or 256k AAC! And yet you have companies pushing 96khz / 24bit audio etc etc. You've definitely got a lot of shiny equipment though Mike so I'd be super keen to give the ABX testing a go yourself! :D I'm trying to find tracks that will trip up the 128k AAC algorithm myself.

[@robertsteinberger]

***** why so much? you just need a pono, or a good interface, and certainly not itunes, which already does stuff to your when it plays it.

[@Katzelle3]

Robert Steinberger *Fiio X5 if anything Forget that Pono bullshit.

[@RoyalHam]

Pulls out a pair of HD 800s like no problem. This dude has some legit equipment and he has some serious audiophile tendencies that I can appreciate :')

[@simplygame5530]

I don't blame the audiophile even with this bcs the race is actually FLAC againt MP3.mp3s are terrible in comparison especially at 128kbps they sound good at 320 but still there is much difference when compared to aac.aac is really excellent codec

[@peskypesky]

I've done ABX testing many times with 320kb mp3 vs FLAC and could not reliably tell the difference. Most people can't.

[@tdotstudentbudget4728]

haha this was pretty entertaining to watch while still proving your point. very good work you did. You went to pretty great lengths to put this myth to rest

[@svenc4803]

Lachlan, why did you perform these ABX tests in such a noisy environment? I thought there were way too many distractions for anyone in that room to truly concentrate on the sound.

[@rohanes]

I wish I could to listen to Platinum Disco on $1000 IEMs/headphones someday...

[@zuyanhe]

I just done the AB test for the "Point Zero" using my HD700, there is a way IMO that makes the difference where 320kbps file has a less range of the sound highness than the 1411k wav format. There is no major difference in the mid range, and all instruments, just the 1411k wav file sounds more open and the image is larger than 320kbps file. you could try to see if you can tell the difference btw 128k and lossless using the track "Point Zero"

[@seltzertronics2538]

Love your channel, Lachlan. Great video. Was fun.

[@erikhaw7313]

Love the intensity of the testers, lol

[@199122vincent]

I also want to add one thing. It's actually more difficult to do this kind of testing with music tracks that you are very familiar with. When listening to the lossy file, your brain will "fill in the gap" because you have heard the track so many times in lossless.

[@jasonlisonbee]

That's still a plus for lossy. It's like saying real world benchmarks are useless for testing RAM speed, because RAM speed makes little difference for most real world usage.

[@imbralp9841]

So interesting! I'm actually writing about a part of this as my dissertation right now!

[@GR-ir2bu]

What about?

[@terryRBNF]

Just the fact that it takes them several minutes to even discern a tiny difference...

[@Akira1G]

8:24 Good taste in music

[@lrmcatspaw1]

Personally I did a ABX test from head-fi with a Fiio E17 and a sennheiser HD 380 Pro and I got it 100% right (with a trick one in there). That being said it was 3 songs only (so ill try a higher sample today), and I now have a HE-400 and an asgard for an amp. Also, in regards to the gear used, the ODAC sounds like a smart move since it colors almost nothing.

[@neon_one]

I had started to convert things to AAC and figured I'd just convert it to mp3 when I needed too, but anything AAC or M4a FDK says it's corrupted when I try to reconvert it again. A lot of other people had this same problem, and the fix is to use different software. Wouldn't be shocked if apple has something to do with this

[@Drive-n-Vibe]

to be totally honest, 128kbps MP3 files dont sound very good, and generally sound flat and lifeless but beyond that i don't see a difference

[@LSDionysus]

"sound flat and lifeless *BUT* beyond that i don't see a difference" = thanks for the meaningless statement.

[@marc_overmars]

@@LSDionysus I think he means that he doesn't hear a difference when you go above the 128kbps barrier.

[@Ciprian-Amarandei]

Can you repeat this test in a busy train station. I bet there is no difference in results

[@TissueFc]

This video is so interesting.Now I'm no doubt why I can't tell any different between MP3 320k and AppleLossless

[@ArielLorusso]

Why you do it in a loud place??

[@Seramics]

These guys has serious gears.... yet struggle to differentiate 128k AAC (which admittedly, is far superior to 128k MP3) and FLAC.

[@svtcontour]

Seramics I think it has to do with the huge amount of noise in the room...

[@Seramics]

svtcontour possible... but some of them use closed headphones and others, use in ear monitors (earphones) which can isolate ambient noise.... given their price tags, i would expect it to be quite good

[@svtcontour]

Seramics That's true. I've only used open ones before so I don't know how well the closed ones would work.

[@ryantoomey611]

+Seramics I agree I have some 128k mp3's and they sound noticeably horrible compared to the same lossless file.

[@six1996]

+Ryan Toomey hmmm are you using the same files? or you download it from 2 different source? u have to use 1 loseless file and than convert it to lossy. hear the different. because, i cannot tell the difference. u will start noticing a bad files when u reach 64k sample rate.

[@Malangsufi]

I always said in terms of lossy format ATRAC3plus and AAC have came close to perfection. They are only obvious on large setup. Iam superised no one in your group had ER4S. ER4S can really help in ABXing.

[@lachlanlikesathing]

Tom actually has an ER-4S, but I don't think he had it on hand that day. For sure I would use an ER-4S if I was doing this!

[@lassejensen4525]

I love the casual style of the video (:

[@duantorruellas716]

good job , very interesting , I learned those AKG headphones must be pretty good , well thought out trails , I would like to see that system test other audio equipment. good luck with future endeavors.

[@zenfreestyler]

128Kbps should be obvious enough of a difference from a FLAC if you ask me, I use to download torrents of music for free (I stopped years ago and deleted all the audio files), and a lot of them would've been 128Kbps, I discovered that there was such a thing as better quality and that 128 was quite terrible compared to even 160 at the absolute lowest, but especially 320Kbps, and actually all my CD's were ripped at 128, and of course had to re-upload them onto the hard drive once I discovered 320 was really as good as I need in terms of sound quality + file size. So once I started hearing the same audio files from 128 to 320 I noticed, clearer imaging and separation of instruments, brighter sounding, especially in the cymbals of drums, overall less boomy longer sustain/dynamic range in notes, especially noticeable with reverbs/delays/echos And of course less distortion, some of the 128 files would distort with or without EQing.

[@lachlanlikesathing]

I did this test precisely because I thought 128k AAC would be easy to differentiate. Why not try the ABX test yourself? You might be very surprised.

[@kirolluse4278]

I did the same and thought that the higher bit rate was much better quality. This was untill I was forced to lower all of my music to 128kbs in order to fit it all on my mobile device. I think what you were hearing was badly encoded music. The encoding software can REALLY make a difference.

[@LusoryPrime]

Kyrillos Elias Not only that, but the encoding libraries themselves become more accurate, or rather, more psychoacoustically tuned for human hearing over time. Dig out an encoding library from the early 2000's, encode a file, get a contemporary one from the same library and encode the same source and you'll probably notice a difference immediately.

[@zenfreestyler]

***** Yeah but they never sounded better to me, AAC is also named M4A right? Those files sound horrible at low bit rates, worse than MP3 I reckon.

[@thk4321]

Great video, would love to see another ab test with balance and single ended

[@nicksmith6130]

Interesting tests.

[@arthurwatts1680]

As others have pointed out, why the hell are the other guys in the room continuing to chat amongst themselves while one of their number is doing his level best to concentrate on the tracks he's listening to ?? That lecture theater (my assumption) is hardly an ideal acoustic environment for any testing, esp with open headphones, but even with decent isolation I can still hear loud conversations going on around me.

[@TriMuXx]

You can convert a 128k compressed file to lossless and not gain any quality, but you can convert a lossless original source down to 128k and lose quality making the lossless version better. So I would do my test that way instead even though it would take a long time to track down the song in an uncompressed master tape format and convert it to both lossless and lossy. Also, some songs aren't going to benefit because they lack dynamics, but that is the choice in style by the artist which is up to the artist anyway.

[@Apollon0077]

Brilliant experiment gentlemen, actual practice instead of theory can make things so simple! I actually got myself some B&W P7's and an amp, with some FLAC files, I used to listen to regular MP3's all the time. I wondered how much of this was snobby crap, or actual refinement. I'm glad to see that I'M not crazy and its all the damn same :-) Someone else did this with wine too, and the "experts" could absolutely tell between a really good wine and a cheap wine.

[@philips9345]

Considering the significant noise in the room, I don't see how this test could be considered valid. Kudos to Thomas for managing 8/10 through IEMs in that noise though.

[@daanw6270]

There is a difference. I can sometimes hear a very slight difference but I still often pick the wrong one because it's difficult to tell which of the 2 is supposed to sound better. Most of the time I can't tell and so I will just stick to 320 mp3 or 256 aac. I do have some flacs but I have enough space so I dont care really. I am just never really going to upgrade if the bitrate is good enough. Even if it's lossy.

[@Pinkskull69420]

>Platinum Disco Based, my friend.

[@Schmidteren]

I just tested myself by random playing similar songs one of each not knowing which is which. Got right the first time. :p But hard to tell. But when alot of stuff was going on, the voice was louder and better on one. And that was the FLAC. Also some small details can also be missed if you really listen for it. But it can be hard. Doing test like this is bad compared to at home, relaxed with your own setup. Also probably gonna ned some really good gear to be able to tell the difference. But I used 320kb/s mp3. 128 is easy to tell from 320.

[@ceremony777]

I think, you should try "aac 128 vbr" - ITU standart for hi-fi transparency for ideal result. Good luck.

[@Stegmutt]

So 'statistically significant' now means 'a percentage I arbitrarily chose?'

[@lachlanlikesathing]

Stegmutt A 95% confidence interval is a typical for these kinds of trials.

[@Stegmutt]

You were correct and I was wrong- statistical significance is traditionally measured at 1%, 5% and 10%. On an unrelated niggle, I think you conflated 'confidence interval' with 'significance level.' It sounds like you've chosen a 5% significance level for your experiment. In contrast, a confidence interval would give a range of values that is some percent likely to contain the 'actual' population value.

[@lachlanlikesathing]

Stegmutt You are probably correct! I'm just an amateur at these things so I'm not quite so confident (haha) with the terminology. The basic understanding of the 95% figure is to be able to say, "I am 95% confident that the subject can differentiate between A and B, and is not randomly guessing". Thank you for replying civilly, and I hope the video was entertaining / helpful :)

[@Vageta1999]

Stegmutt it’s confidence intervals. Even high school statistics tell this

[@Crog2]

"Test track: 'Stroll Through The Sky' Joe Hisaishi 15:01:31 : Test started. 15:07:05 : 01/01 50.0% 15:10:37 : 02/02 25.0% 15:15:21 : 03/03 12.5% 15:17:13 : 04/04 6.3% ... " He has started pretty well but then has gotten confused. He would be able to get a valid ABX test If he would repeat the same test in comfort place like his home without rush and nobody around him. For a beginner 5 correct trials ( 5/5, p = 3.2%, hence less than 5%) is acceptable. A success of blind tests depends more on "how a listener controls him/herself and his/her placebo" rather than "hearing capabilities". Oh, and AAC 128 kbps VBR is considerably better than CBR.

[@lachlanlikesathing]

Thomas is actually thinking of repeating the tests at home actually, but hasn't had the time. I selected 128k CBR simply because it should have been very easy to tell the difference if you listened to what so many people claim. Turning on VBR in iTunes is slightly different because it uses 128k as minimum bitrate and then increases the bitrate dynamically, which for my purposes didn't exactly seem fair,

[@Crog2]

***** This is an example of hard to encode material www.hydrogenaudio.org/forums/index.php?showtopic=19682 www.rarewares.org/test_samples/

[@pheezus]

I think you introduced too many variables by letting them choose to use their own headphones

[@MaxLohMusic]

lmao get real, logically that's not a confounding variable, unless they changed headphones in between tracks.

[@bleedy13]

Dissapointed in you Lachlan. You need 30 Woo Audio Firefly amps connected to Deep Blue that are powered by the chi of Shang Tsung, the true ruler of Outworld. Ever self proclaimed audiophile is doing it!

[@larrymary8866]

Love your hair, dude.

[@mg9261]

BROOO

[@richarddickson9010]

I couldn't tell the difference between 1000+ bitrate ALAC and 320 bitrate MP3 running from an iPod touch 5G using RE-400's.

[@aspenmallery]

It's a practicality thing. The codec is close but not perfect. However the file size difference is substantial, especially if your talking about thousands of songs. I wanted to point out you would hear the difference a lot more if you didn't have chatter in the room with distractions around you. Those headphones are nice but are open ear. You can bet for sure there was ambient noise. I don't know the particular song but if it doesn't have enough musical elements in the mix then it's going to be harder to differentiate. A more realistic test would be using a speaker system with large woofers to reveal dynamic range and stereo imaging in a controlled environment. But in the short- the fidelity impact is negligible if storage space is a issue.

[@50LightSabersInAPack]

This might sound wierd, but for me one of the reasons I like to listen to lossless audio is because of OCD, and the fact that I'm somewhat of a purist. Although it's difficult to perceive a difference (I reckon there is something just perceivable just at the back of my head, it's difficult to describe though but I don't know whether that's confirmation bias), the fact that I'm spending time listening to the tracks whilst not receiving to my ears the best audio my auditory systems are capable of processing, makes me think my time isn't being invested to the best possible extent. The thing is aswell, I don't think ABX tests are entirely accurate - I don't think a comparison between the same songs with different compression methods is an accurate way to determine whether or not you can hear a difference. In a real world scenerio, you're not going to play the same song twice with different formats, but rather you're going to listen to a song once and move on.. So how do you know whether what you're hearing is a part of the song, or a result of compression?

[@BadGuyGoodAudioReviews]

Great video!

[@TheDesertRoseRanch]

With storage so cheap, why would you NOT go lossless? You need to see the "null value" videos that show (and let you hear) the trash left behind when the original lossless file is compared to the mp3.

[@IAmNeomic]

+TheDesertRoseRanch Even as someone who works with audio on a daily basis on Hi-Fi equipment, that's all stuff that is buried in a mix. The only time you DO hear it is when it's isolated like that, and even then, you usually have to normalize the volume of it to really hear everything. Most people don't have analytical, hi-fi setups with sound proofed rooms and DACs isolated from ground wiring, so that stuff would be impossible to hear within the song itself. Literally the only reason I keep lossless files is because I work with audio.

[@HalilDemirkaya]

if we can't detect the difference in these tests. For example, there is practically no difference between using sony wh1000xm3 on iPhone with bluetooth and listening to wired lossless format music on ATH M50X, right?

[@CYNICdiogenes]

The room is noisy, and they are not listening to highly detailed and nuanced music. I'd like to see them in a totally silent environment with a DDD recording of a big symphony that includes a full chorus and soloists.

[@thatunnamedredshirt]

If you have to go to that extent to hear the advantage of a WAY more expensive setup, you've already failed.

[@deykno87]

Great test no more fiio x3 for me,I'll stick with the sansa clip zip

[@LimitIsIllusion]

I doubt anyone could tell which one was which, but I can imagine someone telling them apart

[@noelchristianventura2426]

I hate anything below CD quality. Try listening to MP3s and WAVs of Mili's songs (Chocological, Nine Point Eight, and Yubikiri-Genman) and you'll hear a big difference between the two file formats. (That's how it worked out for me, at least.)

[@TheOverhaul]

What uni is this Lachlan? Do you have a headphone/audio society there or something?

[@lachlanlikesathing]

Nah we just have a group of friends through head-fi / uni that are all audio enthusiasts!

[@MusicFillsMyLife]

***** Which university Lachlan!

[@llamafrhd]

I got 100% when I did mine. It said that there was a 0% chance I guessed 😎

[@BastianUllr]

In my experience, my headphones are mostly useless detecting these differences. Thought, they are pretty noticeable when playing on the stereo. All comes down there to how dynamic becomes, and the improved imaging. I have little flac as production matters the most. Lots of music won't get any advantage from loseless files... like... MOST of the music, won't. Oh, and also, the version of the record. Some remasters are terrible... so for some records you need to hunt the right release. It won't be a difference from heaven to earth, but when it's noticeable, it's mostly clear.

[@GiannisSimeonidis]

One question! These guys knows when they start test which is A and which is B (such as A is 128AAC and B is Lossless)?

[@Rex-wm2dq]

hey lachlanlikesathing who are you anymore?

[@metalchocobo410]

LMFAO Platinum disco!!! i did NOT see that coming

[@Schmidteren]

Need music with alot going on to best tell it. Then listen for particulare things. Then listen to other particular things and go through one and the other. How I did and could tell the difference.

[@feres5020]

Josh is a cool guy!

[@starmaster48]

great. is there differences mp3 players and audiphil mp3 players ?

[@Katzelle3]

Yes, the price.

[@sleepbarrel]

ali sezer One is for people who like to enjoy music. The other is for people who need others to tell them how to enjoy. Im sure you know which is which ;)

[@4punkdude]

These guys seem like a joke but seriously, they're pretty serious! Liked.

[@jeffgendron1959]

Not sure about those headphones, DAC or amplifier in this test. I can say I've heard many amp/DAC combinations and some can mask the audio pretty badly. Certain headphones can be awful in some cases and sometimes the headphone amp can be mismatched to the headphones. All of this works against accurate testing results. I say you find someone with a great hi-fi system. Someone who knows their system clearly shows the differences in the files types and perform the ABX test on that set up. Mid-fi or just "strung together" audio components wont cut it. But never the less this is an interesting video series. Also great audio recordings that have low noise floor can show a lot of quality differences. Some music shows less. So there are many variables to try and prove anything definitively in this type of test. Also too much noise in the background. It should be completely silent for audio and critical listening.

[@RubenStuveling]

Cool test guys!

[@jetlagfrias]

Thank you guys for being honest. Unlike the audiophiles wanna be

[@TheBurleybrawler]

I think it's very hard to tell the difference, but I think you can be trained to tell alot easier, have you tried Philips Golden Ears Challenge?, after finishing it I became a bit more intune with what to listen for.

[@GTX790SLI]

what university? Univeristy of Melbourne? ANU? University of Sidney?

[@donfritz8966]

Wow, diverse group you chose for your scientific study.

[@GroovisticoTRD2]

I think, you should start with testing the test itself for reliability by ABXing something which have audible difference a priori. I've tried to ABX a flat sample and the same sample with EQ bump of 1.5db at 1kHz and I made only 7 of 10. So, maybe this test is not as reliable as many people tend to think?

[@lachlanlikesathing]

Groovistico Alex Level differences can actually be tricky to pick up. You can isolate that here: www.audiocheck.net/blindtests_level.php?lvl=3 vs say www.audiocheck.net/blindtests_level.php?lvl=1

[@GroovisticoTRD2]

***** I've scored 10/10 in +/-1db test. It sounds to me, that all 3 samples start with the same loudness and up and down samples only change loudness somewhere in the middle of playing. Thats kinda strange. But it is even easier that way, because you don't even need to go back to the reference sample.

[@PresentKA]

Why aren't test subjects chosen as people who have more sensitive hearing?

[@lachlanlikesathing]

PresentKA How do you know they don't have sensitive hearing?

[@PresentKA]

There is no reason to make me think that they have sensitive hearing.

[@sleepbarrel]

PresentKA Why do they need sensitive hearing? As long as their ears work fine i'm pretty sure they can enjoy their music and tell appreciable differences. Its not about whether there is a difference between the files, but whether this difference matters. Im pretty sure even the artists wouldn't care about the minute differences in sound quality between different file formats of their music as long as the main appreciable details are there.

[@aqwandrew6330]

@@PresentKA golden ear fckin syndrom

[@NiohNiohYT]

8:55 What is he talking about? How come Miracle is "fairly forgiving on the poor mastered material"? How did he come to that conclusion and what determine if some particular iem/headphone is "forgiving on poorly mastered material"... Nonsense. Nice video though, hope we get more of these videos that bust nonsense myths! Thanks for uploading this!

[@six1996]

+Petar Stankovic have u tried it? i have some experience with those earphoe, mh335dw, and parterre. unlike sony mdr ex1000 or shure 535, they are forgiving on poorly mastered materials. if u haven't heard those thing, u can start try it.

[@Jetpistolhaki]

i want to see Lachlan do the test :)

[@Eliafant]

Hi, nice effort, but I don't think you should do a test like this while people are talking in the background. Or were you trying to mimic a situation on the go for headphones for mobile use?

[@lachlanlikesathing]

No it was more that I didn't want to be a killjoy and tell everyone to shut up. Of course, a custom IEM like the ones Tom and Josh were using should back background conversation negligible, but it definitely is closer to being a real life listening scenario rather than perfect testing conditions. I think what's obvious from the test though is that 128k AAC is clearly not garbage.

[@Eliafant]

***** I guess it depends on what you are trying to show. If you wanted to demonstrate that aac 128k is not garbage I think you succeded. But who really believed it was? I am convinced however that there are differences especially in treble resolution between lossless and lossy. Whether this plays are role depends heavily on the situation (home listening, on the go, critical listening, background music etc.) and equipment used. Also AAC is said to be better than Mp3 at the same bitrate, probably with a 128k MP3 the results would have been at least a little different. In the future I think these discussions will get less an less important as storage capacities are rising and lossless streaming services emerge (see Wimp for example).

[@lachlanlikesathing]

Actually I think the current explosion in the market of high end portable media players that play things like 192khz 24bit lossless files points towards people NOT actually realising that you can throw away 90%+ of the data in a lossy track and you still won't be able to hear the difference in a typical listening condition. I am sure the 128k results would be different because I've passed ABX with 128k MP3 in my earlier videos. But I hate the idea that people are essentially being sold snake oil because the high end audio industry wants to encourage the perception that lossy files are not good enough for portable listening. Considering that even big companies like Sony are now moving towards this kind of marketing, I do not like people being encouraged to believe in these things without putting them to basic testing.

[@Eliafant]

***** Yes you are absolutely right. Often times I wonder if anything above 16bit 44.1khz (CD Quality) really makes sense on a portable setup. Still I always remind myself that some music reappears REMASTERED as a high res download, so it is always good to have the compatibility with your devices, even portable ones. These tracks may not sound better because they are higher res but because of the remastering process.

[@MrRunescape1]

All that gear man, are all those yours?

[@lachlanlikesathing]

Nope it was stuff that different people have :)

[@biboKralle]

Haha. It's like "We're doing a listening test, so off-camera-bro's... have at it! Whatever! It's fine! MAKE SOME NOOOOISE!" :D :D

[@jpahlstrom]

Hey, Trying to find info about apt-x if it can reproduce 16 bit or 24 bit music files like flac; or if it only is just as good as a 3,5mm wiredcable. I guess and hope it's better. And if I have a phone and headphone with apt-x there is no way of actually knowing if the apt-x is activated, or it is just normal less hifi bluetooth. Do you know about a 3rd party company that offers a phone app for showing this? I have a sony z2 and a unopened philips fidelio m2bt. Thanks!

[@backtrace8668]

> platinum disco incredibly based

[@brunobassi2440]

Vorbis 500kbs like lossless

[@PranavKadamares]

so on a 100 dollar iem without trying too hard to hear the difference AAC == FLAC

[@kynancecove]

How about getting the guys at the back to be quiet for the Trials as I remember doing the same with an old friend of mine when I showed him the superiority of the SONY Codec ATRAC3 Vs MP3 it took many plays for him to hear the better definition in mid to treble especially at the end of the day unless it is obviously better than FLAC then one could not tell people to go spend ones£'s or$'s.Anyway I personally am dubious of Apples claims stay with cd of FLAC i'd say or better Vinyl My Dad is definitely an AUDIOFILE who says yes on can hear above the ceiling of the limit21hz so who knows flac is enough spend cash on other things do all aing vs bing in a quiet room!

[@kakashihatake-pw1ti]

Good work gentlmen!

[@antonwibowo8452]

Md5 test flac 😊

[@nextlifeonearth]

If I did the test it would be in a silent environment with open back headphones like the Hifiman HE-560. I would bet I can hear the difference. I did a small blind test and I was 9/12 with K240s so with planar magnetic cans I bet I'm going to nail it.

[@lachlanlikesathing]

CapnTates Why not try it? ABX tests are pretty easy to set up and fun (kinda)! www.foobar2000.org/components/view/foo_abx

[@nextlifeonearth]

***** My proper dac is elsewhere and I don't have planar magnetic cans XD those are bloody expensive.

[@NLS87]

CapnTates yeah, we all really believe that...

[@xemy1010]

I cannot tell any difference between 192kbps AAC (iTunes encoder) and the original lossless file in every case I've tried. The highest frequency I can hear (without cranking up the volume beyond safe levels) is around 19,300Hz, and 192 AAC regularly preserves audio data up to 19,600Hz so I know I'll always be presented with the full audible frequency range. And unlike MP3s, AAC from 128kbps and upwards has zero audible artifacting (MP3 can sound "watery" at lower bitrates), and is technically superior in both efficiency and audio quality. AAC is the way to go! I really feel like it makes FLAC, ALAC and WAV obsolete (except if you want to keep those around for more lossy conversions in the future). TL;DR - 192kbps AAC reigns supreme :D

[@JerryLikeTheMouse]

I agree, except I think AAC is safe and transparent at a lot lower bit rates, but with variable bitrate. There is no reason to use constant bitrate, if you meant 192kbps CBR. I converted much of my music library into ~160kbps VBR AAC. I did ABX testing and AAC is almost transparent for most material even at 110kbps. It's pretty amazing, and I think 160kbps VBR is safe.

[@xemy1010]

I'll give 160kbps VBR AAC a shot. If you haven't tried Opus, that's what I've switched to since writing this post - transparent for most music at ~112 kbps VBR, and transparent for everything all the time at 128 kbps VBR. Almost hard to believe initially but it looks like Opus is the state of the art in lossy compression, especially at lower bitrates. Full CD quality at 1MB / minute or less is quite satisfying. If you use Windows, download LameXP, try the 96 kbps VBR setting for Opus - depending on how complex the upper frequency range of the music is, half the time I have a tough time ABXing Opus and lossless even at that setting.

[@JerryLikeTheMouse]

I gave Opus a shot, and based on what I've read and heard, I agree it is state of the art. My problem with "free" and "open" formats is that "free" and "open" in practice do not translate to real freedom, because of the lack of support. For me, it's just too inconvenient. My current phone and my next phone are unlikely to offer native support for Opus. The reason I switched to AAC is because I realized that in this day and age, it is finally about as widely supported as MP3, and is thus its true successor.

[@UKgamer87]

Im pretty sure i could tell the difference with my sony xba1

[@lachlanlikesathing]

Try it then! :) It's quite easy to do at home.

[@PeterMacPL]

What preset was used in AAC 128kbps ? CBR or VBR ?

[@HeavyProfessor]

I don't listen to anything but 32-bit FLAC 400+ kbps

[@Koobello589]

> 32-bit > FLAC > 400+ kbps ?!

[@noelchristianventura2426]

HeavyProfessor That must sound so pristine and lively! :)

[@Iamakoreancanadian]

FLAC does not support 32-bit bit depth...

[@BadButNotSad]

A whole new dimension is opened up xD

[@tiberiu_nicolae]

I only listen to live bands that I pay to follow me around