mp3 sucks

[Tricky-Loops]

My own songs I publish as 192 kb mp3, I cannot upload huge WAV files, anyway!

Assuming you mean 192 CBR (constant bit rate), consider getting LAMEdrop or similar and using one of the 'V' presets. The V presets create VBR (variable bitrate) mp3s. Aiming to create VBR mp3s with an average bitrate of 192 kbps will yield much better results. More modern codecs such as AAC and Ogg don't even have CBR modes.

Both, I use CBR as well as VBR. I'm not so sure whether VBR is really better as it might introduce artifacts when the bit rate change is too abrupt

[cron]

My own songs I publish as 192 kb mp3, I cannot upload huge WAV files, anyway!

Assuming you mean 192 CBR (constant bit rate), consider getting LAMEdrop or similar and using one of the 'V' presets. The V presets create VBR (variable bitrate) mp3s. Aiming to create VBR mp3s with an average bitrate of 192 kbps will yield much better results. More modern codecs such as AAC and Ogg don't even have CBR modes.

Both, I use CBR as well as VBR. I'm not so sure whether VBR is really better as it might introduce artifacts when the bit rate change is too abrupt

It absolutely won't. The frame size of mp3 is really too short to talk about random jumps, but even if it weren't you'd more likely to notice the opposite if anything. Artifacting happens when an individual frame can't allocate enough bits to correctly encode the source. In CBR, the number of bits per frame is constant (discounting the bit resevoir), meaning that each frame receives the same number of bits regardless of how many it needs. The result is that, while some frames have more bits than they need, the frames that don't get enough are the ones that will lead to those 'jumps'. The V presets on the other hand are tuned in such a way that they target constant perceptual quality from frame-to-frame by allocating the correct number of bits that each frame needs (or least the same number above or below). As such, each frame should be of broadly the same perceived quality.

[metalifuxx]

The big difference is how it is mastered and how squashed each track is etc. Try listening to an original master of music from before the loudness wars started (classical, jazz/blues/classic rock, music that has more accoustic non electrified dynamic instruments/parts) in FLAC HD, 24bit 96Hz or higher, and crank it up on a good set of speakers or headphones. You'll hear the difference between that and mp3. Whether the quality difference is enough for you to care is debatable. You'll hear a lot more depth and spatial perception in the spaces recorded in, reverbs, delays. Dynamics are preserved so much better, and you really hear that smooth high end sheen and solid non muddied low end. But if your listening to modern music from the 90's and later that was limited and sqaushed to hell giving you a big distorted square wave form, or some older music that was remastered for "loudness", you will hardly hear a difference at all between mp3/lossy and lossless high quality/sample rate FLAC/WAV.

[Compyfox]

I think I'm slowly logging out of this discussion, as there is a lot going on that heavily reminds me of "Hydrogen Audio" - especially after comments like "I can hear the difference of 320kbit MP3s compared to WAV files". Or the other comment "it sounds great on headphones, but crap on a HIFI"

What is next, people believe that MP3 is the main reason for the "Distortion of Sound", as depicted in the documentary of the same name, while totally disregarding the "Loudness War" and the existence of "louder releases on purpose - specificially for MP3"?

Sorry - this is just not my type of discussion then.

Also who downloads to listen to a track? >.>

I do - especially for analysis purposes.

If I care, I would download the .wav because 128 is in some cases so degraded. I used to not mind using SC to share music but I quit. I made something a bit more bottom-heavy one day and I couldn't stand it. and I wasn't about to remaster for Soundcloud, I'm not a fan really.

There are several factors involved I think. The major factor is still "can the CODEC take it?", meaning "is the signal strength too much?" (meaning both digital peak maximum and average signal strength).

If that is not the case, then the next factors play huge roles: what bitrate was used for encoding, were there HP and LP filters engaged, is it Stereo or Joint Stereo (read: Mid/Side), Constant Bitrate or Variable Bitrate (which also has influence of a better handling of the frequency masking)?

The most default settings these days are 192kbit VBR joint stereo - and that is a really good format to boot for a lot of purposes (if the content is not too loud). A VBR 320kbit file, if done correctly (meaning: not overcompressed, suitable digital peak headroom) is in A/B tests nearly indistinguishable.

I mean... look at DAB (digital audio broadcasts) - the stream is in MP2 (constant bitrate, roughly 160kbit), and it's accepted up and down. And radio stations (in the digital era) were using DAT, then MP2 (loudness reduced) and now even MP3/AAC as playback to boot. And people enjoy that - it's suitable enough for them.

Over 80% of the time, music is consumed either with headphones or a kitchen radio. The remaining 20% are then the "audiophile" guru's?



FLAC is often named as "alternative", but FLAC is basically just a "compression algo" like 7zip, and players are not as widespread. Also, this format was (to my knowledge) always an open source project that never really took off in terms of support. I also realized through the years, that certain settings of FLACenc can create more harm than "good" (too strong compression for example).

IMO, I still see a future in Fraunhofer's (second) attempt to create lossless MPEG audio streams: HD-AAC. But... there need to be hardware/software players for that (according to white papers, it uses a "fallback" system and plays regular AAC-HE if there is no suitable player at present), and the CODEC also needs to be more widespread. Else it's just another dead on arrival like AAC-LC was for a long time (until the first hacks and reverse engineered tools came out).



But the main question is: will that please the audiophiles?
My answer would be plain and simple: no, they won't be pleased.

Sometimes I wonder why "phonograph records" or "analog tape" is still the most preferred and enjoyable medium, if they have a ton of flaws. I guess people just miss the "noise" and the "imperfections" of these media and therefore demonize digital technology which can show every flaw - every time on playback.



And this is as far as I'll go.

[jancivil]

Yes, there is a clusterfuck of factors and a lot of subjective language leading to this facile 'mp3 bad' assessment.
The thing that made me just finally move away from SC was definitely about <too much> bass. I don't know what they do, but if I were to create a 128kilobitspersecond file, it should be a mite better than what that was. If I really had to rely on SC I would have made a couple of renders targeting it but it isn't anything to me. Always I was noticing the way cymbals were squashed, but this dull, yoogly bass vs what the thing actually is, I wasn't going to share it.

The one thing cron brought in about pre-echo makes sense to me technically but I have never noticed it. Except for my stuff, the majority of what I have in iTunes is 'classical' records and most of this I converted to 192kbps before iTunes had any better. {IIRC, that is.} I'm not going through these files really critically like I do with what I make, but it sounds like the records to me, and a lot of this is me buying CDs from records I spent a lot of time with.

Now with my own music, I have LONG tails which I care deeply about often enough. There is no problem with my 320VBR 'best' copies. The state of the art appears to be very refined and I don't think this loss is very meaningful for the listener.
OTOH I have a bit where in the 'mastering' I had really added gain to a tail for something almost imperceivable I wanted. And it brought up the noise floor. Which wasn't bad, and I would rather have it than not because I like the flavor. I just went to dither it and I believe I like it better. Other than that, I have zero concern with my delicate, gorgeous reverb tails. If it was going to be amplified for a big theater, yes. Once someone went for bringing up the fader playing something of mine on (actual, analog) radio and I learned something here, but that much noise is a wash.

Audiophiles tend to live on magical puffed air and a placebo effect. I have clearly tricked myself through belief, so I wouldn't make very bold assertions before blind tests in the first place. I definitely don't believe in FLAC over mp3, or ALAC over mp3 for listening to music. I have one big fan that believes in CD quality audio so I cater to it, but I wouldn't use the space myself, in any case.

[2ZrgE]

People can claim as often as they want that they hear a (quality) difference between uncompressed and Mp3@320 - without proper A/B-testing that's just opinions (and of course often snobbery) without further relevance. We all like our golden ears, right?

But that topic always leeds to funny threads.

[Deep Purple]

I used to not mind using SC to share music but I quit. I made something a bit more bottom-heavy one day and I couldn't stand it. and I wasn't about to remaster for Soundcloud, I'm not a fan really.

I'm kind of resigned to the fact that those places do that to music - I think they're bowing down to pressure to make music sound 'better' because it's played on poor quality mp3 players, through phone speakers, laptop speakers, etc. It's an improvement I can do without.

The setup in my car is just as bad - more modern radio-friendly recordings (and the radio itself) play back fine, but it screws up the sound of many older recordings, especially in the way that it boosts the low end and seemingly scoops some of the mid. It also has some kind of noise reduction system with a slow attack and VERY fast release. Not sure what they were aiming for, but it means that tracks with a loud start never have the impact I know they should, and it often cuts off notes at the end of tracks before they've properly faded.

What is 'quality' to one person is nothing of the kind to others.

[ntom]

Maybe I'm just a fuckin plebe who doesn't understand shit about codec compression, but I have never experienced any issues in my mixing or mastering related to artifacting from an MP3.
That's not to say that I am not aware of said artifacts. They are easily identifiable on a low bit rate MP3. But have I had to get special drivers or codec compressors that change can use VBR or CBR? No. I've used the built-in render Codec on my DAW (which I believe uses LAME) and it's simple to adjust: You choose whether mono or stereo and at what CBR it renders at. End of story.

Never have I had issues with tails or ceilings due to the limitation of the format, and I have done classical songs to noisey, overly compressed songs.


But like I said, I'm just a plebe who doesn't understand anything. So take it with a spoonful of salt.

[Katelyn]

You know, I've heard of DJ's at clubs using MP3's on their personal laptops. I can't find it but somewhere someone said to me that mp3 is just intended for personal listening, but when I said that I had this fact in mind.

From what I can tell mp3 is fine for most genres with no noticeable difference... perhaps it is just that I had heard that synth so much when working on the track that my brain over exaggerated the effects it had on the end of the reverb/delay of the notes.

[cron]

I often played mp3 at clubs. Never came across a sound system that didn't sound like shit no matter what you put through it. If I was back to back with a vinyl DJ I'd sometimes end up soft-limiting and/or deliberately clipping the output of my laptop to match the kind of volumes expected, mixer and sound system pegged to the red all the while. Couldn't tell you what playing mp3 in a club would sound like on a high quality system that can handle the levels involved (i.e. a club that's a bit slicker than the dives I used to play at), but I expect it'd be fine. Exposure to very high volume doesn't exactly freshen the ear.

[manytone]

Well I don't know what age you are but some of us were around long before the Internet was even invented. MP3 was and is, small and portable. It's become the accepted standard. Flac would be better now but can you imagine trying to get the entire world to change, just so they could hear your music in a better quality format? Kids these days even believe that MP3 sounds better than ''wav'' anyway. Etc...

I remember lugging around a CD player with a single CD in it as a young teen a decade ago when I wanted to listen to something if that gives you an idea. I can see the convenience but at a certain point for some music....

well there's part of your problem no offense, it's all relative. Many of us remember lugging around portable tape players, mine was 8-track and the blue one of these until I switched to cassette then I had a bunch of boomboxes.



but I did also have what was then pretty much state a of the art 8-track recorder (cartridge not multi-track) and later decent cassette decks. The world changed when CDs came out because there was no loss of the quality of the sound from just playing the music. Tape, both cassetes and 8-tracks, had a limit on how long they would last and the truth is so did LPs. Sure the first few times you listened to a tape it was good, but if you didn't keep your heads clean it could get the tape dirty, forget about it you got a tape close a magnet.

If you were smart you recorded your own tapes from an LP so you could preserve the LP and not have to keep buying the album. When a tape wore out you dug out the LP and re-recorded it. Again if you were recording it yourself you could go with a decent tape (for cassette my choice was a 90 minute maxell udlxII) but if you bought prerecorded tapes instead of buying the lp and recording the tape yourself you were getting the lowest quality tape there was which might give you 20-25 listens before you started hearing issues.

Capstans would pinch the tape too hard and it was very common for the tape deck to eat the tape. If you left a tape out that was in the middle of the tape it would pick up dirt and other stuffs and that spot would soon have noticeable audio issues so you needed to make sure that you always rewound (or fast forwarded) to the leader to protect the recorded material. But even doing that only meant you were delaying the inevitable because every time you played a tape some audio quality was lost. Then there was the hiss, even with noise reduction it was still a problem and you had to mark your tapes with what noise reduction you used so you could play it back with the same or there would be no NR.

Now when it came to recording your own music the same principles held true, if you had a 4-track you had to make sure you didn't over work your song because every time you played it it lost quality. If you were like me doing many takes, using a second tape deck to mix down to as opposed to bouncing tracks you saved that sound some but still the rerecording process lost some audio quality. You also had to make sure you used a 60 minute tape, 90 minutes was risky and 120 was suicide as the tape was thinner and often broke under the constant rewinding you did when recording. But rarely would anyone come close to the quality of a bad MP3.

Like I said it's all relative, you had your first taste as it were with a format that had little to no loss from playing and very little care is needed with a CD so I can see how MP3 seems like a step backward for you. However for me MP3 is a far cry better than listening to pops, squirggles, dead spots on tapes, lots of hiss (and I mean lots of hiss) so for me MP3's are a huge step forward, especially as I was very hard on LPs and tapes. There are many albums I have owned on multiple LPs, multiple 8-track tapes, multiple cassette tapes but only one cd. Naturally those were ripped years ago into my collection so I have the MP3 as well (ripping is a lot easier and less time consuming than recording a tape).

It all depends on how you look at it I guess, but before MP3 never could I listen to thousands of songs without ever having an audio issue and do so on something as small as a modern cell phone.

Great post Hink. Brings back a lot of the same memories as I went through the same things. Perhaps we're just spoiled nowadays.
A Good song is a Good song. It can surely sound different on the different mediums and formats available. But.... A Good song is a Good song.

MP3 definitely had it's place back in the day and I think it still does today. I'm enjoying the technology ride I've experienced over my music making days and appreciate being alive during this era. I can surely see you are too Hink!
Paul

[jalcide]

I have become totally skeptical that anyone hears the difference.

Even if you can't hear it, or pass a blind listening test, you can feel it in the way of listening fatigue after a couple hours.

I spend a lot of time listening to electronic music, techno, trance, etc. while I code. My own library is all lossless, but I also listen to the same and similar songs on Plug.dj, letting others play music, all of which is lossy.

It becomes noticeable over time.

Admittedly, most of the songs are sourced from YouTube and ripped from MP3s. So it's not a fair test, but I've also noticed it with my older MP3 library that was carefully converted from CD, before I made the decision to only buy lossless.

The transients are more delicate and crisp and the stereo image is more three dimensional.

Some of the stuff MP3 removes from the signal is really noticeable.

There is no way everyone's ears and brain match the averaged, psycho-acoustic model MP3 relies on. Even if there were a custom model for each person, it would still be a game of tradeoffs.

[mutantdog]

One of the many reasons I favour Spotify now is due to their use of OggVorbis which to my ears sounds a lot better than mp3. In either case I tend to notice the lossy nature and artifacts more in non-electronic music, particularly live percussion elements. I think a lot of mastering engineers tend to roll off the high frequencies a lot more these days to minimize issues in the conversion process which kinda helps i guess.

[nasenmann]

curiously, the one thing where i ever noticed a huge difference between mp3 and wav is chiptunes. despite being lofi, the unfiltered waveforms sounded very different. at 128kbps, it was a world of difference.

i'm really not picky with compression. even some 128kb encodings sound pretty fine to my ears, though there may be noticable differences.

i once listened to this 96kbps mp3 of an instrumental with mellow bell sounds, and it sounded perfect. it really depends on the material used. the intricacies of encoding i know nothing about though.

[jancivil]

I have become totally skeptical that anyone hears the difference.

Even if you can't hear it, or pass a blind listening test, you can feel it
The transients are more delicate and crisp and the stereo image is more three dimensional.

You act like you're stating facts, but there are no facts demonstrated. The last statement here I think is a placebo effect and you're primed yourself to "feel" it, it's psychology. I'm betting your "older mp3 library" isn't 320 variable/best. I would get physically sick and tired of real lossy files in no time at all. I test renders over and over, I feel quite sure I am not suffering any such thing.

Have we not established that there are different quality mp3s? I'm still seeing talk about internet streaming quality which is usually 128.