The “flat” sound that some people describe mp3s as having has nothing to do with them being mp3s. This is a misconception that I’ve seen repeated about as many times as people claiming that vinyl sounds better.

The “flat” mp3 sound (often attributed to ALL digital formats) is typically is due to the fact that when mp3s were first adapted, harddrive space cost a lot more, and they encoded them at low bitrates such as 128kbps, with most encoders cutting out 16kHz-20kHz range altogether. So yeah. That’s pretty bad. Unlike vinyl or cassette or CD, mp3s can be encoded in more than one way. And most of those ways were shitty when the format first came out. Unfortunatley, that’s when most people ripped their music, and music piracy helps keep such shitty encodes circling the globe.

This is why I’ve found myself buying a CD even though I’d already downloaded the mp3s of it. Because I wanted to make PROPER mp3s that actually SOUND GOOD.

The encoder itself matters, too. The same bitrate with different encoders that use different acoustic mathematical models will result in different sound quality. Such models have improved greatly over the years. I recommend LAME.exe. Many power users consider it the best encoder out there.

Let me play you a lossless recording and a LAME.exe, VBR-encoded mp3 over my 5.1 system compared to a WAV or FLAC of the same song. (But only after making sure they both come out at the same apparent volume; people often pick whichever one is the loudest one otherwise.) When you can correctly choose which is which 75% of the time, I might consider you an exception to the normal abilities of the human ear.

I hear a lot of people try to make up a silly claim that mp3s sound flat because they are in fact only discrete points of the audio, with math filling in the rest. And our “brain can tell the difference”.

Sorry, bub. That’s just not true. Your brain may be able to tell the difference between some files on your harddrive, but that is because they are bad files, not because they are digital. Reality is simply a matter of our brain being fed certain signals. There are most certainly upper limits in the resolution and processing power of what we can perceive — even if science may have the wrong limit stated (or not know them at all), and even if human beings exhibit a wide variance in perceptual talent. (There may be 1% that need double the bitrate to be fooled!) Yes, some people definitely hear at a higher resolution than others, just as some people have a greater tonal range. Eventually, however, technology will be advanced enough (and storage space ridiculously cheap enough) that it will be quite possible for everybody to carry a recording of something in such great resolution that no human being on the planet could ever distinguish it from the original analog production of those same sounds.

We’re not there yet, but how many of you have done a controlled study? Did you have someone play you back the 2 files? Did you make sure they both had the same volume? {mp3 encoding often mucks with the volume levels slightly}. Did they record the results? Were they randomized? What percentage of the time were you able to tell? So far, I know few people who have done a test like this. I did in 2000 — and years later decided that the results I had recorded for myself were wrong. I thought 160kbps was good enough. Then I thought 192kbps was good enough. I was wrong. I now encode at the highest bitrate with the highest quality. The point being — I’ve been willing to correct my own personal assumptions several times over the years.

Further complicating things is this:
http://news.slashdot.org/story/09/03/11/153205/young-people-prefer-sizzle-sounds-of-mp3-format

They did surveys and found that people who grew up on inaccurately-encoded mp3s preferred the sound of that to lossless recordings!
In other words, they preferred something worse!

This explains all those vinyl people right off the bat. They like the “warmer” sound because it is what they are used to. And it’s hard to perform a double-blind test when comparying vinyl to other formats, because vinyl is such a broken format that it’s impossible to NOT tell it’s vinyl. I’ve never not been able to hear the needle, to hear pieces of dust. I’ll carve an exception for people who use laser-based vinyl record players (which do not use needles), but they are still subject to the RIAA equaliation curve which causes vinyl to actually represent the full tonality of sound IN AN INFERIOR WAY. But because they can always hear the hiss and crack, in a survey, vinyl will win. It doesn’t mean shit. It means as much as the college kids in the study above preferring shitty 128kbps mp3s over lossless FLAC. People are dumb and don’t even know what they want. I didn’t know what I want and changed my own mind a few times.

The conclusion of the study for me personally is that when a single human being claims to like one format better than another — it really doesn’t mean much in terms of the format being better. People like worse things if that’s what they are used to. Aggregate controlled surveys are what truly dictates which format is better.

I grew up on vinyl, but I know CD is an improvement on it, because I personally think the people who say they prefer the “warmer” sound of vinyl are as full of shit as the students in the slashdot article linked above. The same thing applies to those who think that digital music is automatically “flat”. No. It depends on far more than whether it is digital/mp3 or not.

[P.S. If you haven’t done a blind, controlled study with volume-normalized encodes at the highest possible bitrate — please do not leave a comment about your opinion. I’m only interested in the opinion of people who understand proper testing methodology, and have gone through a proper test.]Mood: hungry
Music: Voivod – Nuclear War (live) (demo) (198412)