|
synthetik posted:This is great: And it's interesting that "Made with high quality parts" is the only sentence that makes any sense. The rest is 
|
#
?
Jan 11, 2013 23:55
|
|
|
|
| # ? Apr 18, 2024 12:47 |
|
|
Where do we stand on people saying they can tell the difference between 320mp3s and flac/wav/aiffs? I personally can't tell. At lower bitrates yeah but not 320s.
|
|
#
?
Jan 14, 2013 13:38
|
|
|
88h88 posted:Where do we stand on people saying they can tell the difference between 320mp3s and flac/wav/aiffs? I personally can't tell. At lower bitrates yeah but not 320s. Generally speaking, it's bullshit. Most people can't find differences at 196.
|
|
#
?
Jan 14, 2013 16:27
|
|
|
Ask them to do... wait for it... a double blind test.
|
|
#
?
Jan 14, 2013 18:59
|
|
|
Combat Pretzel posted:Ask them to do... wait for it... a double blind test. "Oh, er, I'm er, busy and stuff yeah..." *incoherent mumbling* That's usually the response.
|
|
#
?
Jan 15, 2013 13:20
|
|
|
I can definitely hear the difference in sound between a CD and an LP though. CDs make a kind of "PLIK" sound whereas LPs are more "KLOP".
|
|
#
?
Jan 15, 2013 13:26
|
|
|
88h88 posted:"Oh, er, I'm er, busy and stuff yeah..." *incoherent mumbling* Either that or "the cold clinical setting of a double blind test masks the difference in emotion that I can detect with my magic elf ears".
|
|
#
?
Jan 15, 2013 13:56
|
|
|
I only really notice differences between high bitrate, really high bitrate, and lossless when I'm listening to classical music, or something else with a fairly wide dynamic range. On modern music where everything is one volume all of the time and the instruments are just sonic mud, 192 and FLAC are impossible to tell apart. If you're listening to Vivaldi or Schubert you can definitely pick up a distinct improvement in the timbre of the instruments at a higher bitrate, because that stuff gets recorded as it's actually played (e.g. there's an actual wide variation in the volume and pitch of the instruments and cutting off the high and low ends of the spectrum actually dulls the sound).
|
|
#
?
Jan 15, 2013 15:56
|
|
|
128 kbit MP3s sound like poo poo in my car though, high quality Spotify streaming is completely fine and I don't notice a difference compared to FLAC except when they're mastered differently. Almost all the Genesis albums on Spotify have severe DRC issues so obviously my versions are better.
|
|
#
?
Jan 15, 2013 16:12
|
|
|
Even if you can't hear the difference, lossless codecs have three things going for them that none of the "audiophile" poo poo does: 1. The files are quantifiably different from 320kb MP3 in a way that might matter someday. Hypothetically, you can imagine some awesome future encoding algorithm that works noticeably better from lossless sources than lossy sources. Though that's a longshot; AFAIK you can almost always transcode a high-bitrate MP3 into anything else and without noticeable degradation over, say, FLAC. 2. Generally lossless files take fewer cycles to decode, for what that's worth. Usually not much, just like #1. However,... 3. Storage is cheap. It still makes a difference in mobile devices, but at home, the tradeoff in using 7GB to store your music versus 70GB is splitting hairs when storage is $50-75/TB. I keep my music at home in FLAC. It's probably overkill, but it's cheap overkill, especially compared to someone selling shiny rocks to tape to your speaker cables for $2000 to "sweeten your mids" or something.
|
|
#
?
Jan 15, 2013 18:49
|
|
|
Saving in a lossless format also allows you to transcode your music into as many different formats as you want, with no loss of quality. MP3 is probably here to stay, but imagine that Ogg Vorbis or AAC or Musepack suddenly gained massive amounts of traction and displaced MP3 as the lossy format of choice for everyone. You'd have to transcode your files from the lossy MP3s to another lossy format. And if a format change happens again, you'll have to do another lossy transcode. It's either that or rip every single CD you own all over again, provided they are still readable. I know the situation I described is quite unlikely, but by ripping to FLAC, I have ripped my CD collection for the last time. And I plan on keeping those same FLAC files for the rest of my life, so who knows what'll happen? E: vvvv off-site backups, bro 
KozmoNaut fucked around with this message at 21:13 on Jan 15, 2013 |
|
#
?
Jan 15, 2013 18:59
|
|
|
KozmoNaut posted:I know the situation I described is quite unlikely, but by ripping to FLAC, I have ripped my CD collection for the last time. And I plan on keeping those same FLAC files for the rest of my life, so who knows what'll happen? I do! They'll eventually corrupt due to faulty storage mediums.
|
|
#
?
Jan 15, 2013 21:04
|
|
|
Yeah, I rip everything to ALAC. If I'm syncing an iPod then iTunes will transcode it to 256kbps AAC on the fly. If I'm listening on my iPhone it's all coming from iTunes Match, so I'm getting Apple's 256kbps AAC masters streamed directly to my phone. That means I use even less data and I can still have my lossless files at home.e
|
|
#
?
Jan 16, 2013 01:30
|
|
|
pgroce posted:Even if you can't hear the difference, lossless codecs have three things going for them that none of the "audiophile" poo poo does: I've hosted at least a dozen can-you-tell-the-difference-between-mp3-and-wav contests that literally zero people have ever passed, but if I didn't have so much music or if storage was cheaper I'd be keeping it in FLAC just for archival purposes. Maybe one day I'll want to make art out of the waveforms or something and the MP3s will look like poo poo because they're missing everything above 20k.
|
|
#
?
Jan 16, 2013 01:42
|
|
|
longview posted:128 kbit MP3s sound like poo poo in my car though, high quality Spotify streaming is completely fine and I don't notice a difference compared to FLAC except when they're mastered differently. Almost all the Genesis albums on Spotify have severe DRC issues so obviously my versions are better. This is unlikely to be the case, or you're using a bad encoder. Most samples of most types of music are transparent to the majority of people at 128abr, with 160/192 being the end of it. This isn't like "oh, so this is 128 I can hear that right away" -- it's hard to tell the difference in optimum listening conditions, flipping back and forth between both sources. Try it, there's plenty of free ABX (double-blind) listening tests you can do on your computer. The chances of you doing this on one sample, with road noise, whilst focusing on driving ... well it's small. You've placebo'd yourself :P A few types of music have problem samples, with nothing more troublesome than the harpsichord, but outside of those it is very very very hard to be given one sample, and be able to accurately determine whether this is an encode or the CD, for bit-rates over ~140abr. That said, if you can rip to lossless then you may as well, as the reasons MP3 ever existed (storage space being expensive) have long since passed. So what of the audiophile, who can't hear the difference in an ABX? Well, as mentioned, it's "too clinical" so they can't be expected to do it. But, better than that, many will do a spectral analysis on the computer of source vs original, and then go "I heard weaker high-notes, and this is shown in this graph here" Yes, they really do this.
|
|
#
?
Jan 16, 2013 03:52
|
|
|
Khablam posted:Well, as mentioned, it's "too clinical" so they can't be expected to do it. But, better than that, many will do a spectral analysis on the computer of source vs original, and then go "I heard weaker high-notes, and this is shown in this graph here" A spectral analysis of the output of a speaker in test conditions would be somewhat acceptable, surely?
|
|
#
?
Jan 16, 2013 13:04
|
|
|
Neurophonic posted:A spectral analysis of the output of a speaker in test conditions would be somewhat acceptable, surely? Not really. While there will be detectable difference in the spectral analysis, that is not the point. Actually I'm not even sure if there would be a difference once you get to the actual speaker output. The THD on even the best speakers is miles and miles higher than even the most cheap-rear end amplifiers. The point is that MP3 at a decent bitrate (128 ABR for most music, 192 ABR for virtually anything else) is completely indistinguishable from the original to even the most sensitive human hearing. This has been proven again and again with double-blind tests and yet audiophiles insist they can hear (or "feel") a difference. None of them have been able to back this up with a double-blind test.
|
|
#
?
Jan 16, 2013 13:33
|
|
|
Spectral analysis of MP3 encoding would probably show a lot of "error" but that doesn't take into account the psycho-acoustic models lossy compression uses.
|
|
#
?
Jan 16, 2013 16:10
|
|
|
longview posted:Spectral analysis of MP3 encoding would probably show a lot of "error" but that doesn't take into account the psycho-acoustic models lossy compression uses. You can very easily tell an MP3 from uncompressed audio by looking at spectral analysis.
|
|
#
?
Jan 16, 2013 16:13
|
|
|
Twiin posted:You can very easily tell an MP3 from uncompressed audio by looking at spectral analysis. Does anyone know if the psychoacoustic model encodings are robust enough for DSP filters? I just know from image editing that starting with a decent looking HQ jpg something as simple as a shift in HSV can make the gigantic compression errors glaringly obvious, so i imagine it's similar for audio.
|
|
#
?
Jan 16, 2013 17:26
|
|
|
There are also almost no 128 kbps MP3s out there that weren't also created with a lovely encoder years ago that guarantees they'll sound terrible.
|
|
#
?
Jan 16, 2013 17:46
|
|
|
RoadCrewWorker posted:Yes, that's exactly what he said. They said 'probably'. I'm confirming, and saying absolutely. It's trivial.
|
|
#
?
Jan 16, 2013 17:54
|
|
|
RoadCrewWorker posted:Yes, that's exactly what he said. Generally speaking, we are much better able to tell lossily-compressed images from lossless. JPEGs don't really apply an equivalent of a psychoacoustic model, they just work on the principle that areas of low contrast can be 'blurred' more than areas where we expect high contrast, such as an edge. The psychoacoustic model is so effective because our perception of music / sound is in the time domain - whereas viewing image data is an analysis of the frequency domain. A spectral analysis of sound/music effectively reverses this, making our ability to perceive (see) the differences apparent. It's slightly more complicated than that in application, but the gist is there. Consider that when you are listening to digital audio your brain is listening to data which was stored at 44,100 samples per second, each with any one of 256 levels, per channel, and you hear that in real-time. With a random JPEG you might be looking at, say, someone's eye which is made up of a few dozen samples, and you have as long as you want to perceive it. It's not hard to see why it'd be easier to notice compression in that scenario. Most psychoacoustic models work in a similar way, which is to 'blur' areas of the sound where it's confident you cannot perceive the lack of signal resolution - a good chunk of which is music / noise that is -30db (or so) below the main sound that you'd barely hear even without the rest of the music distracting you. The reason audiophiles hate compression so much, is they convince themselves they can hear the difference in the oxygen content of the power cables (0.0000000001% kinda stuff), yet when a codec removes 80-90% of the sound signal they're unable to accurately tell you which is which. It's pretty funny, and pinning them to the wall on this is endlessly amusing (until they ban you for discussing ABX tests).
|
|
#
?
Jan 16, 2013 18:13
|
|
|
So I just read an album review and the reviewer used a bunch of audiophile-like terms to describe the music, such as "lush/lushness", "detailed", "soundscape", "palette", "scenery", "bright, "piercing midrange" and "brittleness". This got me to wondering, is it valid to use such terms to describe actual music as opposed to describing the noise produced by a bazillion dollar "audiophile" sound system? Also, here's a still from the classic 90s TV show Rocko's Modern Life where I'm pretty sure they're taking the piss out of audiophiles: 
|
|
#
?
Jan 16, 2013 19:36
|
|
|
This is also why the old trick of subtracting the MP3 from the uncompressed wave and playing the error signal to show how awful MP3 is is such a terrible "test", it ignores everything about how lossy audio encoding works.
|
|
#
?
Jan 16, 2013 20:25
|
|
|
cheese-cube posted:So I just read an album review and the reviewer used a bunch of audiophile-like terms to describe the music, such as "lush/lushness", "detailed", "soundscape", "palette", "scenery", "bright, "piercing midrange" and "brittleness". This got me to wondering, is it valid to use such terms to describe actual music as opposed to describing the noise produced by a bazillion dollar "audiophile" sound system? I'd say that yes, they are a whole lot more valid when talking about music than when talking about equipment. All of the adjectives you list are basically results of the mastering process. It's about the interplay between instruments, you can have a "rich" sound without going being overwhelming, just as you can have a "sparse" or "delicate" sound. They're completely useless when talking about CD players, amplifiers or cables, though.
|
|
#
?
Jan 16, 2013 20:26
|
|
|
I think it's fair enough to a point when you're trying to describe the kind of audio quality you get from a system - a professionally mixed and mastered song will be carefully tweaked to sound decent on anything, from a crappy portable radio to a car stereo to a real home audio setup. But you're only going to get the real nuances with good equipment, and it can be those little details that really being the music to life. But you can only go so far in saying "poo poo sounds nice and clear without being harsh, and the midrange isn't muddy at all" before you get into wine-tasting descriptions. I think the problem is sometimes like giving things a score out of 10, if something gets a 10 and then something even better comes along, where do you go from there? For audiophiles it's a case of inventing new metrics to claim things are objectively better, so you can say I like the sound of this more because it's more resonant in the rear of the soundstage, like the wise audio guru
|
|
#
?
Jan 16, 2013 22:10
|
|
|
The only valid place for that kind of descriptions is with speakers or headphones, since they are by far the biggest variables in the quality of sound output you'll be getting. For amps, cables, CD players etc., those same descriptions are completely meaningless. You have your "flattering" speakers, the ones that generally make most music sound "nice". Most people want this kind of speaker. They don't want to analyze the timbre of every little hi-hat hit in detail, but they want their rock music to sound full of energy while still being able to bring out the acoustic guitar and vocals on softer tracks. They put a little more "oomph" in the bass and generally soften the treble a bit, for instance using silk dome tweeters rather than metal dome tweeters. I guess you could call these "fun" speakers or "party" speakers or "everyday" speakers. Then you have your "analytical" speakers, the ones that have a completely ruler-flat frequency response, studio monitors strive towards this ideal. You'll hear every little detail of a bad mix or bad compression, which is why they're no good for actually listening to your music for enjoyment's sake. Brilliant for studio work and mixing, but most studios run their mixes through a cheap set of Radio Shack speakers before sending out the master, though, to make sure it sounds good on radio. It's the target market after all.
|
|
#
?
Jan 17, 2013 08:26
|
|
|
KozmoNaut posted:....but most studios run their mixes through a cheap set of Radio Shack speakers before sending out the master, though, to make sure it sounds good on radio. It's the target market after all. And then we run it thru about $15k of processing gear ourselves to tailor the sound to how the Program Director or station engineer likes it. Not as much fun processing audio for radio lately with everything looking like a square wave - all the brickwall limiting used in pop/rock music makes for a mess to process, so much so that de-clippers are being built into some new processing gear to try and clean things up. I enjoy detail and nuances in audio - but I find that audio processing can make a bigger difference than the speakers used when it comes to bringing out certain details. Maybe if these audiophiles bought a AirAura or Optimod 8600 instead of magic wooden blocks and cables, they would be happier with their sound system.
|
|
#
?
Jan 17, 2013 08:51
|
|
|
bigtom posted:And then we run it thru about $15k of processing gear ourselves to tailor the sound to how the Program Director or station engineer likes it. Not as much fun processing audio for radio lately with everything looking like a square wave - all the brickwall limiting used in pop/rock music makes for a mess to process, so much so that de-clippers are being built into some new processing gear to try and clean things up. I know, and I hate that mastering techs are basically being forced into murdering the dynamics for the sake of radio, something you guys are more than capable of doing yourselves, probably with better results. I wish they would master the originals as good as possible, with rich dynamics and a full sound and let the end users (radio etc.) make the final adjustments to fit their usage profile. After all, you can always further compress a track, but it's impossible to "uncompress" it.
|
|
#
?
Jan 17, 2013 09:37
|
|
|
KozmoNaut posted:I know, and I hate that mastering techs are basically being forced into murdering the dynamics for the sake of radio, something you guys are more than capable of doing yourselves, probably with better results. It's not like they don't already make separate radio cuts of songs anyway. One really egregious one lately is Some Nights by Fun. There's a spot where he holds a note on the word "lie" for about 15 seconds on the album, and on the radio they just chop it out super awkwardly without even bothering to blend it, meaning you can actually hear the rest of the track skip cadence. If they're going to send that crap out to be played on the radio, I don't understand why they can't re-process that from a real master that goes onto the CD or iTunes.
|
|
#
?
Jan 17, 2013 15:39
|
|
|
LeftistMuslimObama posted:It's not like they don't already make separate radio cuts of songs anyway. One really egregious one lately is Some Nights by Fun. There's a spot where he holds a note on the word "lie" for about 15 seconds on the album, and on the radio they just chop it out super awkwardly without even bothering to blend it, meaning you can actually hear the rest of the track skip cadence. If they're going to send that crap out to be played on the radio, I don't understand why they can't re-process that from a real master that goes onto the CD or iTunes. Because the number of people who truly care about that isn't worth the money to do it correctly.
|
|
#
?
Jan 17, 2013 15:48
|
|
|
bigtom posted:And then we run it thru about $15k of processing gear ourselves to tailor the sound to how the Program Director or station engineer likes it. Not as much fun processing audio for radio lately with everything looking like a square wave - all the brickwall limiting used in pop/rock music makes for a mess to process, so much so that de-clippers are being built into some new processing gear to try and clean things up. Don't some do that intentionally though? Like how radio stations have a 'house sound' to make their broadcasts more recognisable and enjoyable (which I'm guessing involves a lot of multi-band compression), isn't there a bit of a tug-of-war with mastering engineers giving the stations less room to gently caress with the one true sound?
|
|
#
?
Jan 17, 2013 17:24
|
|
|
baka kaba posted:Don't some do that intentionally though? Like how radio stations have a 'house sound' to make their broadcasts more recognisable and enjoyable (which I'm guessing involves a lot of multi-band compression), isn't there a bit of a tug-of-war with mastering engineers giving the stations less room to gently caress with the one true sound? We do - I use a great amount of AGC (automatic gain control), multiband compression, parametric EQ, clipping, limiting, phase rotation, and reverb on my radio station to give it a distinctive sound from anything else out there. But since my station plays music from the 60's, 70's & 80's I can get away with it since the loudness wars in mastering didn't kick in till the 90's. The problem is that music producers have discovered these same tools, and have gone mad with them to the point where we in broadcasting have to try to undo it with DSP - else everything ends up a very crunchy mess coming out of the radio. If the labels would release an un processed version for radio airplay and another for iTunes/commercial release, it would make life much easier. But for some reason, they won't. Somewhat tangentially related - radio people are not immune to the audiophile syndrome - many a busybody PD or general manager were given dummy boxes to supposedly tweak the stations "sound." They would tell the engineer how much "brighter" and "alive" the audio was...when they were doing nothing but fiddle with a box hooked up to nothing.
|
|
#
?
Jan 17, 2013 20:18
|
|
|
bigtom posted:The problem is that music producers have discovered these same tools, and have gone mad with them to the point where we in broadcasting have to try to undo it with DSP - else everything ends up a very crunchy mess coming out of the radio. If the labels would release an un processed version for radio airplay and another for iTunes/commercial release, it would make life much easier. But for some reason, they won't. Well that's what I mean really - aren't some of them doing it on purpose, so it's harder for you to mess with their vision or whatever? If they can master it so you barely have any room to move, they can control how people hear the song and limit what you can do to change it, so their songs sound more consistent and 'pure' and you don't get to change them for your own benefit. Obviously you're not sitting back and letting that happen, but I thought I'd read something about labels taking that approach
|
|
#
?
Jan 17, 2013 20:36
|
|
|
Learning that radio stations gently caress with the audio before it goes on air explains why most radio stations drive me up the wall with their sound quality.
|
|
#
?
Jan 17, 2013 20:40
|
|
|
Now you can take it with you. http://catacombosoundsystem.com/
|
|
#
?
Jan 17, 2013 22:28
|
|
|
Combat Pretzel posted:Learning that radio stations gently caress with the audio before it goes on air explains why most radio stations drive me up the wall with their sound quality. Yeah, I just figured it was a limit of the stereo signal FM could reproduce .. but this is intentional??  Granted, some of it sounds just fine, but I hear terrible sound garbage all the time on radio (in the UK, for reference).
|
|
#
?
Jan 17, 2013 23:47
|
|
|
Yeah it's meant to be loud and attention-grabbing, so it's compressed to hell and given a 'cool sound' so you'll prefer listening to songs on that station, or that's the idea anyway. But dead air is anathema to commercial radio, so the opposite is... an endless torrent of sound that never stops, even for talking or letting a song fade out! It's always a weird experience to first hear a song on the radio, get familiar with it, then hear it elsewhere and discover how it actually sounds.
|
|
#
?
Jan 17, 2013 23:58
|
|
|
|
| # ? Apr 18, 2024 12:47 |
|
|
Combat Pretzel posted:Learning that radio stations gently caress with the audio before it goes on air explains why most radio stations drive me up the wall with their sound quality. Legally, we have to - the FCC has somewhat strict regulations on minimum and maximum modulation (aka "loudness") levels, with 75% being the minimum and 100% the max on FM. Also, being louder helps reception in fringe coverage areas as well as in the car with road noise. FM also has a nasty 75 millisecond pre-emphasis curve that makes highs very hard to process. Most people however, have no ears and should not be let near an audio processor - you can get in the weeds very easily and create a mess on your hands. A friend of mine flies around the country setting up audio processors for radio stations that buy from the company he works for - it's a skill that many don't have a touch for, and outside of the major markets most stations sound like crap. Music producers are control freaks - Phil Spector hated to do anything in stereo because he felt his songs wouldn't sound right if the listener had stereo speakers that were not placed correctly. I understand where they are coming from, but still - digital audio has a massive amount of headroom in it, and brickwalling everything just turns it into mush. Bob Orban & Frank Foit, two of the most knowledgeable people on the topic of audio processing, wrote about how to make a radio ready mix about a decade ago - interesting reading if you are into producing music: http://www.orban.com/support/orban/techtopics/Appdx_Radio_Ready_The_Truth_1.3.pdf. I personally like heavy processing for certain music - http://jammin.servemp3.com is my web station if anyone is curious. And yes...we HATE dead air. Running a tight board is a skill - fading songs out while firing jingles while keeping your finger on the button to fire the next song is a art form. Or at least it was in the 1980's... bigtom fucked around with this message at 00:40 on Jan 18, 2013 |
|
#
?
Jan 18, 2013 00:30
|
|