Slashdot Mirror
Mastering Engineer Explains Types of Compression, Effects On Today's Music
Stowie101 writes "Today is Dynamic Range Day, which is an event to educate the public about the 'Loudness Wars' that are compressing and harming the quality of today's music. Ian Shepherd, a mastering engineer and founder of Dynamic Range Day, explains why music lovers should avoid MP3 files. 'The one that springs to mind is to avoid MP3, especially if it's 128 kbps. Apple uses a more advanced technology called AAC, but if someone can get lossless files like FLAC that's a better place to start.' Shepherd says it's actually harder to make a good 'lossy' encode of something that has been heavily musically compressed. Very heavy dynamic compression and limiting makes MP3s sound worse, so the loudness wars indirectly make MP3s sound worse."
Hearing the difference now isn't the reason to encode to FLAC. FLAC uses lossless compression, while MP3 is 'lossy'. What this means is that for each year the MP3 sits on your hard drive, it will lose roughly 12kbps, assuming you have SATA - it's about 15kbps on IDE, but only 7kbps on SCSI, due to rotational velocidensity. You don't want to know how much worse it is on CD-ROM or other optical media.
I started collecting MP3s in about 2001, and if I try to play any of the tracks I downloaded back then, even the stuff I grabbed at 320kbps, they just sound like crap. The bass is terrible, the midrange...well don't get me started. Some of those albums have degraded down to 32 or even 16kbps. FLAC rips from the same period still sound great, even if they weren't stored correctly, in a cool, dry place. Seriously, stick to FLAC, you may not be able to hear the difference now, but in a year or two, you'll be glad you did.
I've heard one engineer complain that he mixes the music correctly, with loud and soft passages, but the musicians then demand he make it sound louder. They are not satisfied until the quiet passages are just as loud as the loud passages.
So basically a CD with 90 db range is compressed to about 10 db (plus clipping off the top of the max volume scale).
My AC stalker: " I personally agree with your posts most of the time, but that won't keep me from modding you troll"
Neil Young made the same argument last month in Wired. The interviewer was a douchbag, so I'm not going to link to it, but Neil was right, and first.
-- I care not for your foolish signatures.
All this switching back and forth between dynamic range compression and data compression makes my head hurt.
So to clear things up... dynamic range compression is a form of signal processing that is usually used to make the average level of a signal louder, hence the loudness wars.
Data compression probably doesn't need to be explained to this crowd. But you know... MP3s and stuff.
I am greatly enjoying 'Adopt an Audiophile... And Beat Some Sense Into Him' Month here on Slashdot.
I've been a musician for many years, and I have a nice studio set-up so that I can hear music as clear as possible. Yet I have amassed... umm, through various ways thousands of mp3s as well as flacs and oogs. Do I like the quality of lossless files better, yeah. But does that make me want to get on some flac-only crusade and not listen to mp3s? Not at all. Maybe it's because I'm of that age where I remember scratchy records, or pressing a transistor radio against my ear to hear the latest Jackson-5 or Stevie Wonder cut that was playing on the radio. For me it's the notes, melody, rhythm, lyrics that matter, that's the true musical information. From my music collection I have grown in my musical sensibilities immensely. I don't think it would be possible to have the library I have if everything was lossless just from the standpoint of space and perhaps download time.
So of course, lossless is better than lossy by definition, but mp3s still bring me to where I want to me in terms of getting the music the artist wanted to convey.
The problem is not that bad quality records exist, but that good quality records do not.
every other technology I use takes advantage of MP3. Asterisk can't use FLAC. Which would be hilarious if it did because the standard codecs are about the worst way to transmit music anyways. A phone call is terrible for quality.
Phone calls don't use MP3. The wired phones are uncompressed 7-bit PCM, while the cell phones use a codec designed specifically for speech and barely stream faster than 4-5 kbit/s. (Yes that's right... 1/10th the speed of a 56k dialup connection.)
My AC stalker: " I personally agree with your posts most of the time, but that won't keep me from modding you troll"
Personally I wouldn't blame the degrading quality of modern music on compression. Even with a high dynamic range, there's a higher ratio of crap out there than during the disco era.. Of course, you may be standing on my lawn.
MP3s are still a wonderful compression and it's quite amazing how it has withstood the test of time. Large scale ABX tests have shown people are unable to tell the difference between a 256kbps mp3 and the original lossless recording. Over the past several years I've also noticed a trend for MP3s no longer to be encoded at stupidly low bitrates.
No I won't be avoiding MP3s. I much prefer an MP3 (even at 128kbps) than one of those wonderful "remasters" of an old album. Quite frankly there's nothing masterful about how the loudness war has managed to destroy modern music. The real shame is it doesn't end with the CD master. SACD, DVD-A and I guess now we can include the new supposedly magical itunes format have all tried to tell us the wonders of 24bit music, and yet the dynamic range of music rarely drops below -7dB.
When people download some backyard mp3 digitisation of a Red Hot Chilli Pepper's vinyl release of an album to get better sound quality, or when they download rips of the GuitarHero versions of Metallica songs to get some form of dynamic range you really know the industry has gone to shit.
The loudness wars have been a complaint of mine for some time now. The example video at the end of the article gives an EXCELLENT explanation. I only wish that more people would complain about this so that the quality of recordings would get better. Unfortunately most of the music of today sounds more like a Stephen Hawking lecture with distorted beeping and buzzing in the background and no actual music. When I was in school - music was part of the curriculum. I don't know if it still is but the kids of today are completely CLUELESS when it comes to music. They only seem to like songs about 'guns, money, drugs, niggas and bitches' because they SEE not HEAR these videos on MTV. They see some gun toting loser driving a ferrari, throwing stacks of cash around, surrounded by half naked crack whores and think - "Man that is the life I want to lead!" Their music tastes follow accordingly. If they actually listened to the lyrics - they might actually be disgusted.
When audio recording was first invented, quality was awful, but people loved it, because it was new and exciting, and nothing like it had ever existed before..
Year after year, quality improved.
We expected that someday, recorded music would become indistinguishable from live performance.
Then everything changed.
Convenience became more important than quality.
Storing 5000 mediocre quality recordings on an ipod became the norm.
Combine that with the excessive compression used to fight the loudness war, and it really makes an old-school audiophile sad.
You should be so lucky. These days most mastering engineers shoot for 4 dB of dynamic range at most because, otherwise, the soft passages will be lost in the car noise.
That is all.
but when I do, I buy the CD and make my own flac set from it. Then I can re-encode that to mp3 for portability, etc.
Wikipedia's article on the "loudness war" does a good job of explaining the problem.
http://en.wikipedia.org/wiki/Loudness_war
I used to work for JJ Johnston. He took a popular music track (I won't say which one) and ripped a .wav file from the CD, and then ran a simple Matlab script that tallied how many samples there were of each value. CDs use 16-bit samples, so there were 64K bins in this histogram. You would expect a pretty much Bell-curve shape to the histogram. With this particular song, over half of all samples were either +1 or -1 (i.e., 16-bit sample values of either +32767 or -32768).
That music is so horribly overcompressed that most of the wave forms are sawed-off into square waves. Square waves, in turn, add unpleasant harmonics, which make the music harder to enjoy, and make it louder (in the psychoacoustic meaning of "louder").
I'm hoping that "audiophile" versions of songs become available, not because I think I need all my music in 24-bit 192KHz but because I'm hoping the mix engineers will be allowed to do the mix properly, instead of mixing it far too hot.
I'm sort of afraid to buy remastered versions of old classic rock albums, because I'm worried they will actually sound worse than the originals!
steveha
lf(1): it's like ls(1) but sorts filenames by extension, tersely
Yes, the connections to the PSTN don't use the MP3 codec, and it would be very strange to use MP3 between phones on a PBX system. However, PBX systems like Asterisk do transcode MP3 files to play as MOH or even system sounds on a channel.
If you build a jukebox system to provide MOH, typically the end user uses MP3's to load their music, not FLAC.
Also, if you are doing anything scripted on a Linux system for dynamic content generation Sox does not fully support FLAC. FFMPEG does have support for it, but I am not sure about any others.
So while it is possible to convert FLAC to MP3, so it can be converted to g729 or whatever codec you prefer it does not make a lot of sense when the codecs actually used for transport to most PSTNs are terrible for music and audio fidelity in general.
Which is kind of my point. Unless you are talking about some in-house conference systems, even MP3 is wasted.
>>>soundwaves always peak at the maximum floor and ceiling levels
Hardly. The C64 has a volume control. 0 to 255 if I recall correctly, so the music could range from soft to loud (not maxed-out like today's CDs). Ditto other "chip music" produced by Atari 800s or Commodore Amigas.
I've tried sharing 64, amiga, and Super Nintendo music on facebook but most people think it sounds like junk. They don't appreciate that electronic sound. (shrug). BTW http://www.lemon64.com/ let's you hear 64 music directly over the web.
My AC stalker: " I personally agree with your posts most of the time, but that won't keep me from modding you troll"
More research into your platforms? You really have to make it something you look for, rather than something you expect to just appear.
Personally, I use MediaMonkey as a library app and have a Cowon S9 for music/movies on the go (battery life is fantastic with the screen off). They both support FLAC, OGG, and a bunch of other filetypes. Heck even my phone (Galaxy Nexus) supports FLAC and OGG.
Y'know, there's always someone harping all day long about how MP3 takes a steaming liquid crap all over your sound, and I cannot agree with them. I have a mid-range yet respectable sound system, worth maybe $4000 new. I listen to a LOT of music with an unforgiving ear for detail, and what I often joke as "digital audio memory". Anytime I listen to something, I'm comparing it to a very precise memory in my head. If the pitch is off by a hundredth, there's subtle (dynamic) compression, or phasing issues, I know immediately.
Back when we were peddling 112 and 128kbps MP3s (y'know, 15 years ago), it was pretty obvious that our encoders sucked. You could hear the nasty phasing all over the high end. Today, with most dedicated rippers using "LAME -V0" or 256/320kbps CBR, I'll say that it is impossible to tell the difference on 99.9% of all music out there. Yes, you theoretically lose some high-frequency information above 19khz, but hardly any adults can hear those frequencies anyway, as our range of hearing degrades with age. At 32, I have supposedly great hearing, yet I can barely hear 18khz, and 19khz I can't really hear but just "feel" as pressure on my ears canal. The parts MP3 encoders discard, most people can't hear anyway, and even if we could, it's so high in the audio spectrum that it's just headache-inducing whine. In practice, many mastering engineers will filter that out anyway, because those frequencies are nothing but trouble, they can mess with playback on cheap (read: common) stereos, and are basically a waste of signal which could be better allocated to the mids.
The compression artifacts themselves, they are nothing like they were 15 years ago. If you really want to see how much sound is lost from compression, take an uncompressed WAV, convert it to MP3, then back to WAV. Pull a spectrogram for both the original and processed WAVs, and compare these in a graphics editor. If you're lazy, you can grab the screenshots from here instead. If you're using photoshop, change the blending mode to "Difference" on one of them. Any coloured pixels are the differences, while black means both images are identical.
So, that's digital compression. The other big thing audiophiles bitch about is dynamic compression, and that is an all-too real problem. This is the "brick wall" sound people often cite as the cancer that's killing music. It is the process by which quiet sounds are made disproportionately loud, resulting in the average signal level being louder across the entire album. Most common audio is stored as 16-bit data, this means there are 65536 different intensities available, from silence to maximum, across what is often quoted as 96dbfs of range. Most modern pop music crunches all the sound into the uppermost 6db, so you're kind-of getting 1/16th of the fidelity (yes my math is flawed). This makes crappy speakers and earbuds sound "better" (still shit), and good speakers sound equally shit. It's the sonic equivalent of turning the brightness and contrast on your TV all the way up, now everyone has bright red skin and look like cartoon characters. If you want a painful example of this distortion, cue up Metallica's Death Magnetic, the official CD or iTunes version. Then go find the Guitar Hero version of the same album on TPB and compare. The pressed version is brickwalled, the Guitar Hero version was mixed much more reasonably, in-line with past Metallica releases. Then if you want to hear the opposite, something with very wide dynamic range, try ZZ Top's Eliminator, or Van Halen's 1984. Björk's albums also tend to have good characteristics. You're looking for quiet sounds amid the louder ones - they might be the little squeaks of guitar strings or drum skins, or the long fade of a cymbal.
Back to our buddy boy Ian Shepherd... one of his recommendations for good dynamic range is Daft Punk's Tron Legacy soundtrack. This is pretty much an admission that the man is completely full of shit. Don't get
-Billco, Fnarg.com
Heh, ironically, just today, my coworker started bragging about how much better his FLAC sounds than MP3s, so we did some blind ABx tests. After one listen, the difference was so obvious that I could almost immediately tell which was which for subsequent tracks. He was right. So there is a difference between some of these things. Some things being snakeoil doesn't mean all things are snakeoil. Loudness is real (the article has a movie to teach you to hear the difference, if you care).
The problem is you're not listening to it, but that doesn't mean the difference isn't there, it means you've closed your eyes. Or ears.
"First they came for the slanderers and i said nothing."
Anon Cow has a point, if not blunt. Most people have non discriminating ears till you really get some square waving going, so mp3 is fine for cd. It will get blasted from average home stereo speakers, 6x9 car speakers and crammed into Wal-Mart earbuds after being ripped with the V.U. pegging the red zone anyway.
I've listened to scratchy 78s , old 45s, modern vinyl, reel to reel, 8-track,cassette,DAT and hard drive on everything from audiophile to crap to P.A. systems.
Audiophile music has survived popular culture thus far. Mp3 isn't going to harm anydamnthing. The sky isn't falling. Relax.
I do however expect flac to become standard for bands smart enough to give away their music to promote themselves. What balls to fly in the chin of the music industry who doesn't dare pull a stunt like that.
*Repent!Quit Your Job!Slack Off!The World Ends Tomorrow and You May Die!
I think what audiophiles like TFA just don't seem to get is exactly that, the crap speakers we've been using since the turn of the last century. VERY few people have systems that would even qualify as low end audiophile, most are frankly just barely above crap if not crap. My home system is a 1980s cheap Korean home stereo with a couple of 10 inchers, does anyone REALLY think I'm gonna miss something by not using FLAC? or on the shitty earbuds on my MP3 player or that $30 2.1 system I use at work?
Then there is the other elephant in the room, how so many of us have listened at ear bleeding levels and gone to rock concerts and frankly just don't have that great a range left anyway. I've been playing bass for 30+ years now, guess where they always stick the bass player? Right next to the crash. Between that and my love of powerful bass amps i doubt i could hear the difference between FLAC and MP3 if you put a gun to my head.
Oh but I think you're wrong about FLAC because for 99.99% of the users out there MP3 is "good enough" and the whole point when you give away songs is that you want them HEARD and most people don't know WTF FLAC is or what it will play in, but MP3 is universal, it'll play in anything from that $10 thumbstick MP3 player at the checkout of the Walgreen's to the most expensive audio decks. if folks want better they are welcome to buy my CD, at $10 for 14 songs its not like we gouge on the things.
ACs don't waste your time replying, your posts are never seen by me.
I can hear the difference listening to Dave Brubeck on original vinyl vs. modern CD.
Yes, but we're talking mp3 here, nobody's claiming a 256kbps mp3 is as lossy or distorted as vinyl.
When Argumentum ad Hominem falls short, try Argumentum ad Matrem
I've been playing bass for 30+ years now, guess where they always stick the bass player? Right next to the crash.
First, wear hearing protection. Too few artists actually do this and it hurts them in the long run.
Secondly, why are you standing near the crash? I mean, you're in the band. Can't you stand wherever you wa-
bass player
Oh... nevermind.
Random Thoughts From A Diseased Mind (Not For Dummies)
This isn't a hardware issue, this is the issue of music fans thinking a CD is low quality if the volume doesn't red line their music player.
As an artist, if you make a recording and your CD/master is lower in volume than the typical commercial tripe people will assume it is of low quality. So you end up taking your CD/master to places like Music Masters where they apply the $50-60,000 compression units, and noise reducers to enhance its "red-line" potential.
This is the way the industry has been since at least 1993, and the only thing that has changed has been the introduction of lossy codecs. Lossy codecs have some interesting effects on their own where they tend to compress a sound, and ruin dynamics. Stuff that might only be a hidden layer on a CD can be very much front and center on an MP3/OGG.
Freedom is merely privilege extended unless enjoyed by one and all.
so that article links to a list of 9 good albums, one of which is Nirvana's Nevermind. note: the 2011 remaster is smashed to the wall, and, iirc, you could even hear clipping: http://i.imgur.com/i0Vag.png
.MP3 is capable of far, far greater dynamic range than any vinyl. Lossy compression has its issues, but dynamic range isn't one of them.
Though it does exhibit certain artifacts when the dynamic range and the bit rate are both low. Of course, vinyl has its own issues with hot signals.
Neither the article you linked to, nor the suspect Bauman paper it references measured a real 112db dynamic range from an actual record. They measured the noise floor of a pickup, did some measurements of the pre-amp, did some hand waving about how a skilled ear can add a magic 30 dB below the noise floor and claimed that as the theoretical dynamic range of actually playing a record on high quality equipment. It totally ignores the physical limitations of the vinyl and pickup. It also ignores the other issues with records such as rumble, wow and flutter, poor stereo separation, non-linear response, and other distortions. Their own measurements show the pre-amp SNR in in the 70s, which shows the 112dB claim to be BS.
I label any review as suspect when it uses fluffy wording like this:
As you might expect, the resolution of low-level detail was outstanding. Against a dead-black background, finely layered images floated in three dimensions on what was a somewhat wider soundstage than I'd become accustomed to
From your quoted article
...are represented by stylus motions of less than an ultraviolet wavelength (1/100,000,000 of a meter)—a dimension approaching the size of a complex organic molecule.
Yep, just relying on a few assumptions:
Vinyl can be cut 3 times more precisely with record presses than Intel can than silicon (32nm, 3.2/100,000,000 of a meter) with multi billion dollar factories.
This movement, after being transferred through you multiple millimetre long stylus, then into the magnetic coil, then through a meter non-zero impedance analogue cable and into your pre-amp, this signal will be stronger than the radiation caused by a star in another galaxy.
There exists a cleaning method that can remove all dust, germs, water drops, micro-organisms from vinyl without putting a 10nm scratch into it.
When Argumentum ad Hominem falls short, try Argumentum ad Matrem
Ian Shepherd's mentioning that one should avoid 128kbit/s encoded MP3. This is leaving out a critical piece of information. Luckily he mentioned himself that heavily (audio) compressed music (data) compresses very badly. This will be especially evident if you force the encoder to only allocate a fixed number of bits to a section, called "Constant Bitrate" (CBR) in MP3 encoders. "Busy" sections will get the same data allotment as quiet sections. This problem can be diminished by using "Variable Bitrate" (VBR) mode when encoding, which encodes to a specific target quality rather than file size. With that, (LAME) MP3s can still sound good enough around 128kbit/s, since the encoder is free to allocate more bits to critical sections and less bits to non-critical section.
In short, there is no reason to use CBR encoding, unless your target device is unable to decode VBR encoded files, or you absolutely need to know the exact bandwidth requirement of a stream. It defeats the whole point of lossy encoding, which is to reproduce the original with highest possible fidelity, not reach a target file size.
Hey, I'll have you know girls LOVE the bass player buddy! Its not our fault them damned whiny guitarist and pretty boy singers keep getting in our way and hogging all the spotlight! Oh and ladies? just remember those whiny little guitarists are all rush rush rush and singers will fight you for the mirror, but we bass players are great with our fingers and now how to build a nice slow groove. Just sayin, something to think about. Just don't think about it around me as my Cherokee princess is liable to scalp my ass. Word of advice fellas, ain't nothing more scary than a NAP (Native American Princess) with a case of the green eyed monster, talk about feeling like Custer!
Oh and as for earplugs? can't wear 'em, they throw the tone off so damned bad it makes me literally trip over my own fingers. I really need the full range of the 5 string to be able to groove and with the earplugs you lose the low mids, it sounds like you are trying to play while you have water in your ears, so not fun. I've learned not to stand anywhere near the crash and to put my bass amp cocked off to the side instead of blasting at me and instead have a custom mix in a monitor so I can control the volume a little better.
ACs don't waste your time replying, your posts are never seen by me.
112dB? Ha! Hilarious.
You're lucky to get 70dB out of audiophile grade vinyl. See http://wiki.hydrogenaudio.org/index.php?title=Vinyl_Myths
There's also an interesting discussion of the dynamic range of both vinyl and 16-bit/44kHz digital audio here:
http://www.hydrogenaudio.org/forums/index.php?showtopic=47827&st=0&p=425794&#entry425794
The dynamic range of vinyl does vary by frequency. For example, in that thread a poster notes he measured 84dB at 300Hz for vinyl. A 300Hz tone recorded to a 16-bit wave file with noise shaped dither exhibited a dynamic rage of 151dB!
Vinyl has extremely limited dynamic range in the bass - something like 30dB at 20Hz. The needle would pop out of the groove if you tried to record more than that. Vinyl also suffers from constant negative signal to noise ratio incidents, when impulse noise (clicks and pops from scratches, dust and defects in the groove, static discharge) completely drowns out the signal. Unacceptable, in any format.
See also this recent article, which, while skewering the distribution of 24-bit/192kHz audio, notes that 16-bit digital audio has an overall dynamic range of 120dB with dither:
http://people.xiph.org/~xiphmont/demo/neil-young.html
Vinyl's a shitty format for reasons apart from its inferior dynamic range, but that's not terribly surprising since it's like 100 years old, mechanical, and prone to a plethora of issues - rumble, wow and flutter, phase issues caused by the RIAA equalization / de-equalization process, scads of unwanted harmonics and harmonic distortion, ultrasonic noise, preamp hum, static clicks, etc., etc., etc.
Probably should have been replaced by some other analog disc-based format by the early '70s - maybe something based on RCA's capacitance discs, which wound up being used for video, and had scads of bandwidth - more than enough for near-flawless reproduction of the original studio master tapes. But at the time most industry attention was focused on the emerging lo-fi but convenient tape formats, first 8-track then cassette, as well as the failed competing quad systems. And then by the middle of the decade everybody knew a digital format was coming, with Sony and Philips working first separately, and then by '79 or so together on what would become the Compact Disc.
The funny thing is, I'm sure there's all sorts of gunk in a vinyl groove resulting in tiny stylus motions that are theoretically audible...but which weren't put there intentionally.
Whoops.
I grew up with vinyl and always thought it was a ridiculous audio format. Fragile, noisy, screeching high frequency harmonics, static, had to be cleaned every play, degraded, spindle holes always off center, warped, scratched, hum, rumble, wow & flutter, clipped, pure unadulterated shit. To add insult to injury a decent turntable and cartridge cost at least $300 bucks even in the early '80s (probably $600 in 2012 dollars), just to try and make that crap format sound reasonably acceptable.
By the early '80s and the arrival of Dolby C the lowly cassette sounded better in a decent deck, and that's pathetic when you consider cassette was invented as a dictation medium. Unfortunately most pre-recorded tapes were poorly made, and the only thing you had to record from was either FM radio or (you guessed it), f'in vinyl.
it makes me literally trip over my own fingers.
Some hate it when people use "literally" when they mean "figuratively".
Not me. I picture you literally tripping over your fingers. Which body parts were you playing your bass with before that happened?
First, if people ever give you shit about being a bass player, show them Victor Wooten. Too many bass players just jam on one string, move down a note, and repeat. There's a lot of potential with the instrument.
Secondly, the idea isn't to block out the music you're playing - it's to block out the volume. The standard solution I've heard of is to have noise-cancelling headsets with everyone's instrument piped into them but at a much lower volume. The point is volume reduction.
I see very, very few musicians that actually give a shit about their hearing, sadly.
Random Thoughts From A Diseased Mind (Not For Dummies)
What makes you think that whatever you're hearing has anything to do with that one specific variable (the sample rate) and nothing to do with any of the millions of other other variables (microphone selection and placement, preamps, mixers, A/D converters, processing, your D/A converter, amplifier and speaker system) inherent to the process of recorded music?
Typically, when a single person experiences reality in a way that the majority of people don't, we consider them delusional, yes.
If you want to not be considered delusional, pass a double-blind ABX test where A is a direct feed from an analog signal source and B is the same source being run through an A/D and D/A conversion at 44.1/16bit.
Make me a friend and I'll mod you up