Can We Really Tell Lossless From MP3? [Archive]

infuzion

11-18-2009, 08:14 AM

http://news.slashdot.org/story/09/11/18/0123217/Can-We-Really-Tell-Lossless-From-MP3

Unlike most slashdot posts, many of the comments are actually educational.

I still prefer WAV for extreme pitch-shifting as a DJ, but I did use MP3s for much of my recent mix, & some were lower quality from eMusic. I doubt most here can tell the difference, since I did some minor post-production.
http://www.sidewaysfive.net/forums/showthread.php?t=948

Brandon

11-18-2009, 08:22 AM

I've had this argument with people before, but most people can't start to hear audio degredation until 128k. I can start to hear the hiss/tin of files (even more so when they're VBR) at 180K, but people that claim to be able to tell the difference between a 320k and a 192k file are full of shit. From what I can tell, you can only hear that difference on a loud system and the difference between 320k and lossless can only be distringuished on a high-end audio system.

dmtree

11-18-2009, 08:27 AM

i think there is a slight difference between lossless and even 320mp3 - you can sometimes tell on a high-end system, but generally the differences in production quality between tracks are much more pronounced.

Brandon

11-18-2009, 08:33 AM

the differences really come out when someone records a vinyl record to MP3... but a pure digital recording is a lot harder to tell a lossy from lossless format.

Erik Mitchell

11-18-2009, 09:34 AM

people that claim to be able to tell the difference between a 320k and a 192k file are full of shit.

Disagree. I can def hear a difference especially with the hi hats. They sound compressed and not as bright. I can even tell in the headphones. :nerd:

And am I the only one having a problem reading the actual article?

Disko

11-18-2009, 09:56 AM

i agree. at 192k in most songs you can hear some loss in the higher frequencies. especially on good speakers. not always, and its definitely not obvious. but like mitch said, it aint as bright.

dave

11-18-2009, 09:59 AM

i believe a particular playback system's ability to resolve minute details is the key.

Brandon

11-18-2009, 11:08 AM

Disagree. I can def hear a difference especially with the hi hats. They sound compressed and not as bright. I can even tell in the headphones. :nerd:

I think that has to do with the production quality of the tracks. Most hats don't start getting tinn-y until the BR goes below 180. From what I've been able to tell, the only diff between 192 and 320 is purely in spectral size/loss and the sound differences are hardly noticable unless you're on the audiophile gear. Or maybe I just have shitty Sony cans and need some new ones. Or maybe I'm just getting old and deef. *shrug*

dmtree

11-18-2009, 01:02 PM

i can usually hear it in how full the bass sounds and maybe a bit of the high end too. i'm used to jagged/basscouch setups that have a very clean (Danley/Yorkville) sound and full spectrum of bass, so i can usually tell minute differences in how the system is sounding. but again, the difference is usually from one DJ to the next - if one was using 192 or maybe even 320 mp3s and the other is playing lossless, there's some improvement. but not really enough to justify the jack-up in price for wav quality tracks, imo.

Hazefire

11-18-2009, 02:20 PM

Idunno, I can tell the difference sometimes, but like Brandon said, it depends on the production of the tracks. I definitely notice the difference on stuff with really high production values that has been mastered well, but less so otherwise. It just sounds less... compressed, lol. That being said, I still buy my shit in 320 especially on beatport. $1 more? fuck dat.

And while we're on the subject, why don't more people use flac? Shit is awesome, would save on bandwidth, etc...

Erik Mitchell

11-18-2009, 02:44 PM

Now I kinda wanna test this by converting a few wav files to mp3 and comparing them... Hmm...

dmtree

11-18-2009, 06:49 PM

Now I kinda wanna test this by converting a few wav files to mp3 and comparing them... Hmm...

I did that once with Beyer 770 headphones and could tell the drop-off using high-bitrate VBR (levels were noticeably lower), but not really 320.

That was with my shitty laptop soundcard though. Haven't tried it since I got the Firebox.

infuzion

11-19-2009, 07:07 AM

Now I kinda wanna test this by converting a few wav files to mp3 and comparing them... Hmm...Just for you, I posted my DJ mix as a 24bit 44.1KHz source file that I compressed into MP3 (VBR next to highest quality) & OGG (8.5 quality). I maded the mix in Live7, saved as 16bit, then mastered in WaveLab4 with 40Hz or so rolloff, small high boost, light punch, & a limiter.

I'd be interested both if anyone can hear the differences in the files, and also if anyone can tell what of the source songs I used were WAV or MP3.

http://depositfiles.com/files/rw9rcodac : reEightATgmail_SquareButtonGray_TechOutMyHouse2009 .ogg : quality 8.5 85.3 Mb (http://depositfiles.com/files/rw9rcodac)

http://depositfiles.com/files/frkkve4ko : reEightATgmail_SquareButtonGray_TechOutMyHouse2009 .mp3 : VBR 2nd to highest quality 82.52 Mb (http://depositfiles.com/files/frkkve4ko)

http://depositfiles.com/files/1pmcfe8kc : reEightATgmail_SquareButtonGray_TechOutMyHouse2009 .WAV : 24bit 44.1KHz 722.81 Mb (http://depositfiles.com/files/1pmcfe8kc)

Brandon

11-19-2009, 07:31 AM

A guy I know in Portland did a test on NWTekno with people who claimed to tell the difference... he pulled a fast one and made ALL the test samples 320k and said they were different things. Tricky, sure, but he proved a point: People here and believe what they want to and our keen abilities to tell slight differences in digital encodings aren't as accute as we'd like to believe.

Erik Mitchell

11-19-2009, 07:49 AM

A guy I know in Portland did a test on NWTekno with people who claimed to tell the difference... he pulled a fast one and made ALL the test samples 320k and said they were different things.

Haha sneaky sneaky! I'd still be interested to do an actual test though... I also agree that a lot of times it's shoddy mastering or mixdown that's the culprit. Those will sound shitty whether in 192 or wav format hehe

I'd be interested both if anyone can hear the differences in the files, and also if anyone can tell what of the source songs I used were WAV or MP3.

Nice man, I'll try to check these out when I get home from work, can't do depostifiles here...

Brandon

11-19-2009, 08:23 AM

Haha sneaky sneaky! I'd still be interested to do an actual test though... I also agree that a lot of times it's shoddy mastering or mixdown that's the culprit. Those will sound shitty whether in 192 or wav format hehe
And that's the point he was trying to make. I emailed him before he announced the results and asked 'Are these all the same file? The sound identical', but he denied to keep the results more impactful. We then did an actual test and people failed that one even worse. So, I have a hard time believing people that say they can tell the difference on most music.

The ONLY music I think it's rather easy to hear the nuances are with alternative/rock because the trap cymbal gains a bad tin/hiss as the encoding rate comes down. Even at 320, I can hear some slight variance on those cymbals... but, again, this might be due to the mastering and original recording used to make the files.

If you want to catch file degradation, listen to the hats/cymbals.. they seem to project changes best to my ears.

infuzion

11-19-2009, 04:31 PM

The ONLY music I think it's rather easy to hear the nuances are with alternative/rock because the trap cymbal gains a bad tin/hiss as the encoding rate comes down. Even at 320, I can hear some slight variance on those cymbals... but, again, this might be due to the mastering and original recording used to make the files.A remarker in that slashdot link said he can tell with classical & jazz, perhaps likely since they tend to have more dynamic range & often recorded with high-end audio in mind.

dmtree

11-19-2009, 06:39 PM

makes a difference which mp3 algorithm you're using too. not sure if there's anything better out now, but I've always used LAME, since it was the best one.

Daver

11-20-2009, 08:57 AM

Yeah from what i understand you don't hear much loss on a 320 mp3 unless its on a high-end system. That being said someone once gave me numbers. .wav has 1/3 the dynamic range of analog, and 320 mp3 has 1/10. Some of these tests are bogus too, unless you have sat down and TRAINED your ears to look for the sometimes subtle differences between compression rates.....you are not going to know the difference. You have to know what to look for.

brandon bass

11-20-2009, 09:10 AM

if the track is mastered well the answer is no, if it is mastered like shit the answer is yes. whew, glad we solved that one. carry on with what you were doing now and hopefully nobody ever posts this again AGAIN

dave

11-20-2009, 09:13 AM

.wav has 1/3 the dynamic range of analog, and 320 mp3 has 1/10.....

i doubt this, digital has the potential for much higher actual dynamic range than most analog mediums...

the RIAA curve came about as a workaround due to the physical limitations of the medium (vinyl) with respect to dynamic range...

or am i in left field here?

don't get me wrong, i'm all about analog... but fact is, digital has a higher potential in this arena....
question is, will it sound as good or better?

Daver

11-20-2009, 01:40 PM

i doubt this, digital has the potential for much higher actual dynamic range than most analog mediums...

the RIAA curve came about as a workaround due to the physical limitations of the medium (vinyl) with respect to dynamic range...

or am i in left field here?

don't get me wrong, i'm all about analog... but fact is, digital has a higher potential in this arena....
question is, will it sound as good or better?

Correct me if i'm wrong but analog is a "true" medium and what you are losing, or potentially losing with digital is dynamic range and character...........using an analogy, i thought analog was like having a perfect circle in front of you, and digital the impossible attempt at recreating that circle, however close you might get.

infuzion

11-20-2009, 01:52 PM

Correct me if i'm wrong but analog is a "true" medium and what you are losing, or potentially losing with digital is dynamic range and character...........using an analogy, i thought analog was like having a perfect circle in front of you, and digital the impossible attempt at recreating that circle, however close you might get.Both analog & digital in general are NOT "true"; both are approximations of the live real deal. However, "better" sounding might not be true; that's why people shell out big money for tubes, exciters, mastering, etc.

It could be that a decent rate of MP3 might sound "better" on a weak (cheap earbuds, laptop speakers, etc) system since unplayable frequencies are filtered out.

But we as DJs must also think about is the extra cost of WAV vs MP3 worth it, due to pitch-shifting, additional effects....

budros

11-20-2009, 02:14 PM

But we as DJs must also think about is the extra cost of WAV vs MP3 worth it

YES, it is.

this is a subjective debate that will never have a clear answer, no matter how many blind taste tests you throw at it.

that being said

i can tell the difference between 320s encoded by different retailers

and most of you have damaged hearing ;)

i'll take the blind taste test on lossless any time! FLAC4LIFE!!

dave

11-20-2009, 02:14 PM

you do not loose potential dynamic range going from analog to digital,
in fact, i believe it is the other way around...

Daver

11-20-2009, 02:17 PM

Both analog & digital in general are NOT "true"; both are approximations of the live real deal. However, "better" sounding might not be true; that's why people shell out big money for tubes, exciters, mastering, etc.

It could be that a decent rate of MP3 might sound "better" on a weak (cheap earbuds, laptop speakers, etc) system since unplayable frequencies are filtered out.

But we as DJs must also think about is the extra cost of WAV vs MP3 worth it, due to pitch-shifting, additional effects....

well here is my thought, take the attack you hear on a violin while someone gradually applies more pressure. You simply can't recreate that with a computer, the "resolution" of an attack like that is just way beyond what a computer can recreate. I thought this was the same way with analog, the dynamic character that is captured is for lack of a better word, the "true" character of the instrument. Now we can get higher and higher resolutions with 1's and 0's but wouldn't it technically never reach the quality of analog?

To us it might sound the same but i view it as a case of diminishing returns; and that once we go beyond a certain bitrate you stop hearing a difference but technically you are still not reproducing the EXACT quality of analog. MAybe I'm totally wrong on this but I thought this is what the whole debate and movement for saving vinyl was about?

BeatFreq

11-20-2009, 02:39 PM

Well sure some information always gets lost in the conversion from analog to digital, but frankly, not much music gets produced in a strictly analog environment any more, so even with a vinyl record, you're often getting an analog representation of a digital signal (either produced in a DAW or recorded from DAT), so I'm not sure how much of a benefit there is to gain there. It seems kinda like taking a film photo of a digital print because film offers higher resolution - well, sure, but your resolution is only likely to be as good as the lowest resolution in the chain.

And yeah, I generally find 320 mp3 to be fine. Sure, I'd love to have WAV quality everything, but it's impractical in terms of storage. With most music on most systems, I can't tell much of a difference below 192 anyway.

Brandon

11-20-2009, 03:50 PM

i can tell the difference between 320s encoded by different retailers
Now THERE is some truth. Some stores have clearly different encodings than others (usually a quieter file, muffled almost?).

But practicality is something to consider too. Given that a WAV file is 4-5x as large as a 320k, it's just not feasible if you use your lappy for more than just DJing, IMO. That HD would fill up quiiiiiiick.

infuzion

11-20-2009, 09:05 PM

Originally Posted by infuzion
But we as DJs must also think about is the extra cost of WAV vs MP3 worth it

YES, it is....
i can tell the difference between 320s encoded by different retailers
i'll take the blind taste test on lossless any time! FLAC4LIFE!!You asked for it; here is my DJ mix with very light mastering. Half in .WAV, other half in MP3, a few MP3s are medium-quality VBRs... I dare you to guess which is which! I'll even give you the 24bit master just to give you a head start! ;)
(BTW DownloadThemAll! works great with depositfiles)

http://depositfiles.com/files/1pmcfe8kc : reEightATgmail_SquareButtonGray_TechOutMyHouse2009 .WAV : 24bit 44.1KHz 722.81 Mb (http://depositfiles.com/files/1pmcfe8kc)

...take the attack you hear on a violin while someone gradually applies more pressure. You simply can't recreate that with a computer, the "resolution" of an attack like that is just way beyond what a computer can recreate....Now we can get higher and higher resolutions with 1's and 0's but wouldn't it technically never reach the quality of analog?Well, with analog recording, you also lose resolution; the mic diaphragm can't perfectly pick up the sound hitting it, you lose a very slight signal information in the wires due to restance/RFI/EMI, & amps & mixer add their own coloring, as do the tape heads & the tape itself.

Digtial does not read every sub-sub-harmonic due to it being 1 & 0s like you said, but I think "someday soon" the rate & bit depth will be so great you won't be able to tell the differnce, & the digital will actually be more accurate than analoge. Like how computer graphics are now; high-end digital still cameras can pick up the fine hairs on someone's arm when doing a full-body shot, & the red video camera (http://www.red.com/cameras/tech_specs/) is better looking & more realistic than cameras 4 times its price.

cardinal

11-22-2009, 08:15 PM

cardinal

11-22-2009, 08:16 PM

you do not loose potential dynamic range going from analog to digital,
in fact, i believe it is the other way around...

I don;'t think the argument here is as much dynamic range more than it is flaws in A/D conversion

could be wrong.

Hazefire

11-22-2009, 09:08 PM

I don;'t think the argument here is as much dynamic range more than it is flaws in A/D conversion

could be wrong.

yup. another thing I've heard brought up in this whole debate is sampling rate, and it not being high enough on mp3's or somesuch. The wikipedia page on sampling (http://en.wikipedia.org/wiki/Sampling_rate) states the following:
The Nyquist–Shannon sampling theorem states that perfect reconstruction of a signal is possible when the sampling frequency is greater than twice the maximum frequency of the signal being sampled,[4] or equivalently, when the Nyquist frequency (half the sample rate) exceeds the highest frequency of the signal being sampled. If lower sampling rates are used, the original signal's information may not be completely recoverable from the sampled signal.
For example, if a signal has an upper band limit of 100 Hz, a sampling frequency greater than 200 Hz will avoid aliasing and allow theoretically perfect reconstruction.

This explains why mp3's, with a 44.1 khz sampling rate, tend to get muddled high hats and treble.

s.p.e.c.t.r.e.

11-22-2009, 09:56 PM

yup. another thing I've heard brought up in this whole debate is sampling rate, and it not being high enough on mp3's or somesuch. The wikipedia page on sampling (http://en.wikipedia.org/wiki/Sampling_rate) states the following:

This explains why mp3's, with a 44.1 khz sampling rate, tend to get muddled high hats and treble.

the standard wav format is 16bit 44.1 khz. sampling rate isnt the issue w mp3's its the compression, and usually what starts degrading first is the high frequencies. at 320k i doubt 1 in 100 people can tell the difference between an mp3 and a wav