TAXI.com

I was just going to start a post about mp3s, and saw yours! Maybe we can merge these 2 threads!

I have some ideas:

First of all, I really dislike mp3s as compared to higher resolution recordings. But they are what they are.

My biggest complaint is most of them sound so COMPRESSED! And, lately, I figured out the problem. --- If you record your digital masters at too loud a level, especially tipping into the red, any computer using Vista etc. that makes .wma files, or even .wav files will slightly compress them to start with. THEN, any mp3 program that I've used further compresses the sound. I'm not talking about "lossy" compression, but audible compression.

So the trick is to make sure your digital masters peak at -2db or lower. I've really experimented with this over the last week or so, and am convinced of this. The hotter you make your mixes, the more audible "pumping" of compression you're gonna hear. As opposed to analog recordings, louder is not better when it comes to mp3s. Sure, try to get your levels as hot as possible to compete with everyone else, but when your mp3s sound like sh!t, what's the point?

My 2 cents,

Ern

-2 eh? I am gonna try that..

I am driving myself nuts trying to get a good level. No mastering compression is obviously bad because we need to compete, but too much is even worse because then things sound like crap. For me, it is easy to tell those extremes, my difficulty is, where to draw the line. I often feel flustered when I am at this point with a song, to the point where I throw my hands up in the air and ask for divine intervention. LOL

Yea, it gets that bad...

Rob

Consider these threads merged! Thanks for that idea about keeping the level kinda low for MP3's sake. I will try that and see. I, not surprisingly, usually do the final mix as hot as I can, so that makes some sense. A few db might make all the diff. What i hear more that compression though is a kind of bubbly distortion or something. Hard to describe audio with words, but it's almost like tape warble from a crappy analog machine. Or even, sometimes, it's "bubbly"!!! Weird...still, this is Friday, so a good day to experiment!

How are you recording your final stereo mix? I've found that mixing to a seperate stereo audio track within the session improves my final mix. That way your computer is not trying to mix and convert at the same time. I run all my tracks through a master fader with a mastering limiter on it and then to a stereo audio track that I label "mix". Then when I bounce to disk, whether it's as a 192kbps mp3 or as a wav file, it's just a single audio file that the computer is dealing with. I've found this method to yield better results than bouncing my whole session at once. eo.

A couple of other things to take into account:

1) The bit rate that you're converting at. For the TAXI hosting site the maximum is 192kbps. You could create two versions, one for TAXI and then another one at 320kbps to get the maximum quality.

2) Variable Bit Rate (VBR) vs Constant Bit Rate (CBR). Some software apps will give you the option to choose between those two settings. If that's the case, choose Constant.

3) The MP3 algorithm itself. There are at least a couple of different algorithms that are used to create the MP3 (maybe more). One of them is the LAME algorithm, which is free and is used in apps like Audacity and Reaper. The other is the Fraunhofer, which is the original algorithm and is not "free" is used in Pro Tools and other apps as well.

EDIT: My point in #3 is not meant to be an endorsement or indictment on any particular DAW.

All three of those factors will also play a part in how your mp3's sound as well.

hth,
- Big Blue

davem wrote:Consider these threads merged! Thanks for that idea about keeping the level kinda low for MP3's sake. I will try that and see. I, not surprisingly, usually do the final mix as hot as I can, so that makes some sense. A few db might make all the diff. What i hear more that compression though is a kind of bubbly distortion or something. Hard to describe audio with words, but it's almost like tape warble from a crappy analog machine. Or even, sometimes, it's "bubbly"!!! Weird...still, this is Friday, so a good day to experiment!

Yea, Dave, it does make it sound "bubbly" or "warbly," whatever you want to call it. But the levels go up and down unnaturaly, however slightly you overmaster your track levels. The higher the level on your masters, the more this phenomenon becomes!

I'm writing currently for a publisher, and he's getting masters WAY too hot. He's emailed all the writers, and wants the masters to come in at -6 db! --- I just made a CD from some of those masters, and that's way too low to my ears, but he's erring on the side of caution, I guess! I pumped the tracks up in my Masterlink to about -1 db, and the CD sounds great. But I guess for mp3s, a bit lower is better.

And so it goes,

Ern

Yeah, I've learned to leave 3dB headroom (-3dBFS) in my mixes when converting to MP3; I jes' dip the master fader after I'm done with everything else. I simply pop it back up to -.3dB for CDs. Using the LAME encoder, here. I think it's more an encoder thing than a Windows thing; the encoders were originally designed for film, where they have standards as to how hot a mix can be. Crest factor and average level are separate issues; you can still have your waveform look like a solid black line, if you want, heh.

I also find it helpful to gently roll off the highs above 10kHz, and not get too crazy with the low end; I was listening to some late eighties country CDs last night, (practicing me drums) and it's amazing how little bass energy they used compared to now. Frankly, I think this has to do with format changes, too, not just the volume wars-when those CDs were made, (Vern Gosdin, Desert Rose Band) they were also coming out on vinyl, where too much bass will pop the needle right out the groove.

mojobone wrote:I also find it helpful to gently roll off the highs above 10kHz, and not get too crazy with the low end; I was listening to some late eighties country CDs last night, (practicing me drums) and it's amazing how little bass energy they used compared to now. Frankly, I think this has to do with format changes, too, not just the volume wars-when those CDs were made, (Vern Gosdin, Desert Rose Band) they were also coming out on vinyl, where too much bass will pop the needle right out the groove.

Yes, and yes about a couple of things, Mojo: --- 1. People always complain about cymbals on mp3s. Too "swishy" sounding ("Not that there's anything wrong with that"--- Seinfeld ), which might be too much high end on a mix. Personally, I haven't found that on my mixes, but there ya go. --- 2. Bass used to be REALLY understated because of vinyl and the needle popping out of the groove. That's less of a concern with digital music, of course, but I always roll off the low lows anyway @ 40-50 hz.

It's funny that McCartney used to bitch at the EMI engineers that his bass parts had no low end. These guys WERE engineers, wearing white lab coats etc., and told the Beatles that they had to follow scientific standards. But Paul told them "Listen to Motown, for godssakes! James Jamerson has TONS of low end on his bass!" The lads got their wishes on boosting treble during the middle years, and FINALLY on bass (Abbey Road). I've got a vinyl copy from "The Original Master" of Abbey Road, and the bass is almost TOO much! Paul must've been pleased.

BTW, a bass player friend of mine has always griped about "There's no high-end!" on Fragile by YES. Maybe he's right, but it sure is a great sounding recording, even to this day!

Ern

Interesting thing about the bass on those Motown records, I suspect James Jamerson's dead strings (he used the same bass and the same set of strings on every record) had very little transient energy at the front end of the note; thusly avoiding the threshold of any limiters and compressors further down the chain....he could get more lows, cuz he had no highs, sorta thing. The kick drum, conversely, was kinda pointy, with more beater impact than note. For non-soundtrack stuff, I dump everything below 'bout 50Hz; nuthin' down there you wanna hear, jes' air-conditioning rumble and a semi changin' gears a block or two away, heh.

mojobone wrote:Interesting thing about the bass on those Motown records, I suspect James Jamerson's dead strings (he used the same bass and the same set of strings on every record) had very little transient energy at the front end of the note; thusly avoiding the threshold of any limiters and compressors further down the chain....he could get more lows, cuz he had no highs, sorta thing. The kick drum, conversely, was kinda pointy, with more beater impact than note. For non-soundtrack stuff, I dump everything below 'bout 50Hz; nuthin' down there you wanna hear, jes' air-conditioning rumble and a semi changin' gears a block or two away, heh.

For me, it's garbage trucks in the alley, and Lear jets going out of Van Nuys Airport heheh!

As far as Jamerson, I think you have it, Mojo. He maybe used a mute on the bridge to dampen the sustain, but I've heard more often that he opened up his Fender bass and put sound-deadening material in it to make his sustain really short. If you listen to some of his parts soloed on YouTube, they're really dirty and noisy, but NO sustain! I think he had a direct box that he could dial in as much "fur" as he wanted, but the lack of sustain was totally from his bass and how he played (I've heard with just his index finger --- how he could play so many notes with one finger is beyond me!). What a great musician ---

Ern