Sunday, August 31, 2014

deathtokoalas
right.

so, a compressed audio file (an mp3) reduces it's size by throwing away information on the high and low end. running that through an eq that exaggerates the low and high end can compensate for this mildly. it's not as good as an uncompressed source through a flat eq, but it's better than a compressed file through a flat eq. but, when you put an uncompressed file through an eq designed to compensate for compression loss, it's going to sound absurd.


there's actually been a change in mixing philosophy recently, where producers in certain genres are mixing their records to sound optimally through compression, throwing all kinds of wrenches in the whole thing...

at the end of the day, you're right: there's an eq in foobar and there's probably an eq on your mp3 player, but not all eqs are created equally (the one in my sansa is kind of weak and was not able to boost the lows as powerfully as my bass boost headphones were).

ExpensiveGarbage
Soooo..... you're just sitting here having a conversation with yourself?

deathtokoalas
the truth is we're all really conversing with ourselves.

ExpensiveGarbage
Yes, this is I can agree with. 

ImActuallyABanana
well what you have to understand is that when an mp3 "throws away" information in the highs and lows, that information is gone. Beats audio exaggerating the lows and highs isnt exactly a good things. Beats audio is boosting the severed highs and lows... its not adding, its boosting.. so basically, your boosting highs and lows that sounds like crap.. you get me ?

deathtokoalas
if you look at the response on the phones, it actually cuts the lowest registers and boosts around 100. this is consistent with an "mp3 bass boost". i have a pair of sennheisers that does the same thing, but it's marketed as "bass boost for mp3 response".

what an mp3 throws away is mostly outside the audible range, but it sort of bleeds over into the audible range in various ways. i don't have the numbers off hand, but it's not going to cut at 100. it'll highpass around 20 or something and then slowly reduce the information up to around 50 iirc. so you want that cut down bottom to cut out the noise, and then the boost a little higher to exaggerate the lost lows.

so, no. you can't boost information that is gone. it's just gone. it's not boosting a "severed" high or low. what's lost is lost. what's is doing is exaggerating what still exists in order to compensate.

it's not some kind of magical de-mp3 uncompression algorithm or something, it's really essentially a trick. but if they're anything like my sennheisers, it's a trick that works.

stated another way, you can mangle the signal under 50 and boost at 100 and get the effect of reconstructed bass, even though it's not what's actually happening.

if you're listening to really deep bass, you'll possibly be able to tell. but the average bass range is simply not that low. you're getting overtones and murkiness on the bottom. the overtones are gone, they can't be reclaimed, but you can fudge it by boosting a little higher.

i doubt the phones have them, but a little reverb box might be able to reconstruct a bit of those overtones. you'd be fighting with purists, though. fuck, i'd fight with you over that...

ImActuallyABanana
thats mostly what i said.. if you ever listen to an mp3, you can clearly hear that isnt not just the unheard frequencies being gone.. if that was the case, then mp3s would sounds great.. but its not. mp3s sound like shit in general no matter what bitrate they are. they lose alot of the frequencies that matter as well and with those frequencies completley gone, boosting the highs and lows will do absolutley nothing for sound quality

deathtokoalas
obviously, you can't make an mp3 sound like a flac by using an eq. that's not what i was saying. but, you can make an mp3 sound "punchier" by boosting the bass, and it will sound better to most people.

you can try it yourself. find a wave editor. i still use a copy of cool edit i "found" somewhere back in about 1998, but i think the standard free tool to do this nowadays is audacity. rip a cd to wav. pick something that's not deep bass music. then, run it through a high pass filter around 20 hz. you'll immediately notice that the rumble is gone, but that the track doesn't sound particularly different. then put it through a parametric eq that slopes from -50% at 20 to +50% at 100 and then back down again to 0 around 150. you'll immediately notice that it's punchier than the file that's just cut at 20, even though it's missing the subbass.

it's not a perfect comparison because the mp3 is losing some information up to about 50 (iirc). but it gets the point across.

ImActuallyABanana
ive done this alot and i can honestly say that to me, being in audio for 6+ years, i have never heard an mp3 sound even slightly better with any type of eq... its like compressing distortion, no difference. in reality your just turning up Crap. Ive had to take mp3 samples of things and put them in a new song (24bit/44.1k) and when you blend in that mp3 well in there, thats when an mp3 is tolerable. but yes your right, to some people(meaning regular consumers), boosting those frequencies on an mp3 would sound better. to them. not to us of course. i just hate mp3s in reality.

Randy Parsons
The compression rate has a lot to do with it.  Unless you use a really bad compression  people simply can't tell the difference between compressed and uncompressed.  Even so called audiophiles.  They have done studies, although they are kinda hard to find with Google.  As for EQ.  EQ to the room or to your 10$ earbuds or whatever makes it sound good to you.  Beats certainly didn't invent the EQ.   Many cheap earbuds have pretty horrible bass.  I would always prefer to EQ myself instead of some 'preset' that Apple and Beats does.  Not sure if Apple still does this on their ipods.

deathtokoalas
for those that are curious, the highpass (sort of.) does occur after the fourier transform (for the lowest frequencies), but this isn't the right place to get into the mathematics of the issue and such a pedantic discussion does nothing to address my argument. if you want to look into this, though, understanding the way the waveform is split in encoding will help you understand why more information is lost at 40 hz than at 100 hz and why there's still enough of the signal around 100 hz to boost for the signal loss in the subbass. i don't think i implied that mp3s literally highpass, and i apologize if you've misread me in that way.

fwiw, i wrote the following piece of music using wavelet and fourier transforms in partial requirement of a graduate level course in mathematics on data compression algorithms:

i just want to ask if you did an a/b with compressed v uncompressed sources regarding the bass?

the reason is that mp3 compression kills the bass. some time in the middle of the last decade, i actually bought a pair of sennheisers with a bass boost (and they were advertised and marketed this way) that was purposefully designed to counteract the mp3 compression. if you tried to use them in the studio, the bass was totally muddy and washed out. but, if you plugged some studio phones into the mp3 player, the bass sounded thin - reflecting the source.

http://www.youtube.com/watch?v=XkVZwj4pZ7A

a discussion of “beats” headphones

deathtokoalas
but, just how bad are they? i've never heard them.

i'm not quite an audiophile. as a composer and heavy listener of complex music, i'm very concerned about quality, but i know most of the claims made by audiophiles are unable to pass double-blind tests and you usually want to take what they say carefully. i've been into high end sennheisers since i was in grade school, so i know what good and bad headphones sound like. i just can't think they're really that bad, not even with the understanding that they're mass marketed for profit rather than manufactured for audio purposes.

one of the things that's recently impressed me with my sennheisers is their ability to pick up the bow noise on a cello part. just excellent reproduction. but, is somebody listening to top 40 pop looking for that kind of response, or do they just want the bass to drown everything out? if your expectations are less (and dre's expectations may not have been high), you're going to test the phones in less stringent ways and ultimately not be able to tell.

there's going to be a cut-off point in terms of quality where virtually everybody says "this is shit", but that cut-off point is much lower for the consumer market than it is for the audiophile one.

so, how bad are they?


the other thing you have to keep in mind is that headphones nowadays are mostly bought for portable use, which means the source is probably compressed. are people really using flac? i doubt it. music that's....found...online is still generally compressed.

i was looking for a pair of sennheisers a few years ago for portable use and searched around a little and realized they're actually producing a completely different line of phones for it than their studio models, which are created specifically to compensate for the loss inherent in compression. i ended up going with a bass boost model. when i plugged my studio sennheisers into my mp3 player, i could hear the problems with the compression - on certain albums that i've heard hundreds of times and can play back in my head from start to finish. the bass boost alleviated that mildly (although it didn't help with the high end). conversely, i couldn't use the bass boost in the studio for obvious reasons, it just muddied everything.

i'm going to guess that the beats are probably a bit boomy, but it's a valid question as to whether that's a purposeful response to the compression and whether people are using wonky metrics.

but, i'm really just curious: how bad are they?

Stegmutt
Innerfidelity's review is a good place to start, although maybe a bit technical: http://www.innerfidelity.com/content/monster-beats-dr-dre-solo 

deathtokoalas
that wasn't very technical at all.

it's beginning to confirm my suspicion that these phones are made for compressed audio, though - exaggerated low end, and a rolled off high end, where it doesn't matter due to the compression killing it.

i again have to reiterate that high-end companies do this, but they market their products properly. the result sounds awful from a flat source, but really brings the compression back to life on the bottom.

i can't argue they're not overpriced (my bass boost sennheisers were $100 (canadian) not $350), but are these reviews using them to do what they were engineered to do....?

Berenizes Gutierrez
when it comes down to it beats are perfect for the average consumer that really doesn't care about being able to hear every aspect of a song

deathtokoalas
i'm really curious if somebody with a pair could test with compressed audio, specifically. that's not the way you're supposed to test audio equipment, but it's the consumer reality right now. it would maybe help if beats had specified, as sennheiser did. but, it's easy to draw the conclusion from the frequency response that that's what they're going for.

luwiigi427
Marques does explanation of what Beats is actually doing with the audio /watch?v=Cdbn_pmxFic

He also does a side-by-side comparison with the M50s /watch?v=et_PWifUd1w

The only thing he doesn't do (which is probably what you really want out of this) are legit audio tests/comparisons. To be fair, I'm pretty sure he doesn't have the equipment to actually do it.

deathtokoalas
right.

so, a compressed audio file (an mp3) reduces it's size by throwing away information on the high and low end. running that through an eq that exaggerates the low and high end can compensate for this mildly. it's not as good as an uncompressed source through a flat eq, but it's better than a compressed file through a flat eq. but, when you put an uncompressed file through an eq designed to compensate for compression loss, it's going to sound absurd.

there's actually been a change in mixing philosophy recently, where producers in certain genres are mixing their records to sound optimally through compression, throwing all kinds of wrenches in the whole thing...

at the end of the day, he's right: there's an eq in foobar and there's probably an eq on your mp3 player, but not all eqs are created equally (the one in my sansa is kind of weak and was not able to boost the lows as powerfully as my bass boost headphones were).

so, maybe they're overpriced.

but maybe it's a clever way to make up for sales lost to torrents, while making the pirated product sound better at the same time.

i mean, i'm not as big as dre. nobody's torrenting my work. right now, i'll argue that if you like it you should throw me some cash, 'cause i'm just getting by and it'll help me keep going.

but if i was dre, and i had some money in the bank, what would bother me most about torrenting would be that people aren't hearing the thing i spent hours and hours creating properly. i mean, i didn't spend all day tweaking the reverb so you could fucking compress the file afterwards and lose it. if i had to resign myself to compressed audio as the de facto standard, i'd be yelling at people to tweak their eqs properly...

it makes sense when you analyze it.

even if it's overpriced.

when you keep that in mind, the whole "as the artist meant it" thing does make sense - presuming you'd otherwise be listening to compressed audio on a flat eq. more accurate would be "closer to how the artist meant it to sound, considering you're BUTCHERING THE FUCKING SOURCE".

i think it was kef (a medium to high-end british speaker manufacturer) that initially came up with the "as it was meant to sound" line in relation to producing very flat speakers.

Stegmutt
I don't think the solo hd's were designed with compressed files in mind- they were just badly designed. From a manufacturing standpoint, it doesn't make sense to deliberately create an acute roll off spanning the lower two bass octaves and have a spike from 100 to 300hz. This probably has more to do with Monster's incompetence than any strategic design. But this is all in the past. By most accounts, the Solo 2's are excellent performers and a great value. Hopefully Beats bringing engineering in house has turned around their product line and they will offer good products at competitive prices.

deathtokoalas
but it does if you're trying to compensate heavy bass compression - which is just as applicable to xenakis as it is to dubstep.

somebody really ought to do the test.

Fabian
sincerly, they sound good :/

Stegmutt
No, from a manufacturing standpoint it doesn't make sense to have a roll off in the lower two octaves. It's not like they save money by having crappy response in the octaves most bassheads crave.

deathtokoalas
on a clean source, yes, but an mp3 is going to cut off so much from the lows that turning it up below 100 hz is just going to introduce noise. it's 100-300 where you can still reconstruct it. that's probably also why there's such a steep cut about 7500 - you're just going to get static on an mp3 if you boost it there.

fwiw, it's below 114 hz where mp3s start getting noisy.

Miguel
About 2 years ago. I had a friend who purchased Beats -- the ones that sit on the ear and not cover them. I tried them out and I was astonished how similar it sounded to my $40 sennheiser ear buds. Sure the bass was higher but it was missing a certain clarity to it. Sure my friend didn't have a dedicated amp or use flac or even a phone with a great sound chip. But this is the typical consumer. Buying "high end" headphones just for the brand and less for actual quality. But whatever. The sound was obviously not horrible but horrible due to the price. It'd be fine if they were $35.

deathtokoalas
and here's another question: was their ability to compensate for mp3s a part of the reason apple bought them and then changed the design?

one of the reasons i don't own an ipod is drm. now, i understand they've changed it a little. but i want a device i can put what i want on without proprietary software &etc.

if this pair of headphones shows up to make lossy audio sound better, it potentially hurts their product.

Stegmutt
I understand the concept of psycho acoustic compression, but once again, from a manufacturing standpoint, having a steep roll off spanning the lowest two octaves is bizarre. Sure, many consumers use lossy files, but it make no sense to deliberate manufacture poor response below 100hz. Or are you implying that beats assumes all their customers exclusively listen to 96kb mp3's that have no data below 100hz and then specifically designed their headphones to perform poorly in that range? That is a dubious assumption at best.

deathtokoalas
you don't have to go down to 96 to get bad low responses, you can hear it at 128-256, but the limitations of the technology are going to create weaker bass response even at 320.

i am explicitly stating that it seems like beats has assumed that their customers are listening to mp3s on their phone, and this is the reason they've created the frequency response that they have.

and they're mostly right.

and again: sennheiser did something similar with their "bass boost mp3 phones", which sound great through my sandisk that's sending out mp3s and horrific from my cd player.

MistahJuicyBoy
What bitrate are your MP3s? Lossless is a big difference, but it's not that drastic. Good headphones will always sound better, no matter what you're putting in

deathtokoalas
i rip to 320, but the stuff that i....find....is however i...found...it, which is often 256 or lower. 192 is probably most common.

in my view, you want your studio headphones flat. but flat phones will then reproduce the compression. just trust me when i say that that bass boost in the sennheisers (which i can confirm were engineered specifically for mp3 use) made a gigantic difference. but, as one would expect, it made higher quality sources sound muddy, as well.

Toxis
To be honest - they sound pretty nice, especially if you like deep bass, and as a person who listens to hardcore (gabber, not a punk one:) - they really really sound great - never owned them, too expensive, and I bet there are better headphones for the price (i.e. audio technica) but the beats do not sound crappy at all (again, I am not an audiophile, they sounded good for my taste) - the price is different beast - would I buy them - nope. would I use them if I had those - yes:)

heh, I do own scullcandy fix - and I never knew until this video they are same category as beats (the overpriced category:) - got these beacause my brainwavz died, like them a lot, especially because managed to get them for ~33$ - look awesome, sound awesome, fit awesome :)

energeez
hey smart lady , love your comments!  i own a pair of beats and use them for dubstep, club music, which i listen a lot to, but listen to everything, which i use other headphones for.  My theory for the freq.  is that they try to make them fun sounding (which i do find them fun), and how do you make something sound fun?  You cut off the sub bass, and increase the kick bass/ mid freq synth bass.  Thats my take.
disease research should be funded with progressive taxation. i mean, people are dying and you've gotta get people to do a stunt to raise money to fight it. it's barbaric.

\
they dance like elaine...