The Light of Day

The fallacy of “odds of success”

From a dead-on post by Steve Pavlina:

Often such seekers [of success in any arena] will look for a certain statistic to help them assess the risk: What percentage of people who attempted a similar venture actually succeeded to the degree I’d like to experience? For example, if you want to earn $5000/month as a blogger, your question would be, “What percentage of bloggers who try to generate full-time income actually earn $5000/month or more?” Suppose it’s on the order of 1%. You then interpret your odds of success as the same figure.

What does such a statistic have to do with your personal chance of success? Nothing at all.

To me this is like asking, “What are my odds of success in kung fu?” If you’re committed to becoming a black belt in kung fu and are willing to put in the time and training, you’ll probably do just fine. But if you’ve never studied martial arts and are looking for a fast and easy road to success, you’ll be sorely disappointed.

In many fields you only see a 1% success ratio because the other 99% are merely taking up space. They’re just dabblers, not serious contenders. You’ll often see this 1% figure in fields with a low barrier to entry such as blogging, acting, or music. You’ll find a small percentage of people who are really committed to mastery, but the rest have virtually no hope of notable success.

(Emphasis mine.) This is of what I was trying to convince my friends and family 20 years ago. But that’s neither here nor there, since what good would it do me? It’s myself that I need to re-convince of this. I was right about something — yes, teenagers can be — and in a way I’ve actually been too open to disempowering points of view, at least as far as my outward presentation of myself is concerned. Fortunately I haven’t stopped the actual groundwork at any point in my life, though I may keep a lot of it either hidden from view, or just plain too tangled/incomplete/obscure to understand without any background knowledge. What I need to resuscitate are my above-ground aspirations. Come out of the closet, if you will.

Bottom line, it’s not all an odds game. Not by a long shot (er, no pun intended).

The unknowable black art of mastering part 2

Blah blah blah linear frequency blah blah logarithmic amplitude blah blah bright pretty colors.

There’s no particular reason why they have to be in the colors of a rainbow, mind you, that’s just my own personal eye-candy; in fact it’s probably detrimental, because I wind up sitting here going “ooh pretty” instead of moving along to the next step. (No, everyone, I really am working! It’s mental work!)

Anyway, the kindergarten version of this … if I can translate it into kindergartenese … “mastering”, as we all understand (and as none of us are the least bit prone to confuse with “mixing”) is the process of taking an otherwise complete, already mixed-down recording, and running it through some final nit-picky EQ and other processing to make it “play well with others” … particularly YOUR EARS.

The top half is an analysis of an old, crappy-sounding song, and the bottom half is a clean, full, big-sounding modern song. Left to right is low pitch to high pitch, but since the pitch scale here is linear, it means it’s not spread out the way we hear it. Most of the important stuff to our ears is smooshed into the left part of the picture.

The stuff that drops way off at the far right is in a range that only small children can hear. It’s interesting to consider that since music fans tend to be far younger than mastering engineers, they can hear stuff that the mastering engineers can’t.

The useful thing about all this is that any recording can be matched to any other recording’s profile. So what the top song lacks in highs (notice how much faster it slopes downward), it can re-gain by modeling it after the bottom song’s profile. And it can do this without creating a shrill, ear-hurting “spike” because it’s aware of the peak levels across the spectrum for both songs. It’s not just a dumb treble boost — it’s a precise auto-tailoring to a model.

The intriguing thing about these graphs, for me, is how straight of a line they both seem to form. I’m not sure what to make of that. This is the first time I’ve ever converted the values to decibels before graphing them.

The boxless box set starts to take shape (possibly in the shape of a box!)

Assuming I can continue to stave off demons, dragons, and other disasters with the same level of tenacity that I have thus far, I foresee the assembly of a “boxless box set”, possibly titled “Everything I’ve Ever Done” or something equally cryptic.

It has to do with organization. In a way, I’m not sure why. On the one hand, if you collect 100 songs into one place you can guarantee them a “life raft” for the day when the body finally surrenders to entropy. On the other hand, depending on how things are organized, certain things will still manage to be drowned out by the rest of the collection. If they were “lost” and then “found” they might attract more interest than if they were sandwiched between a few unpopular obscurities. But then again, isn’t “sandwiched between a few unpopular obscurities” just another form of “lost”? Bottom line, I can’t guarantee them all a bright future with a two-car garage, but I can at least make sure they all make it onto the school bus.

They will be sorted vaguely by what phase of my life they were initially written in, regardless of when they were recorded or remixed. There are just too many songs that I’ve put away for a few years, done revisions and touch-ups on, put away for a few more years, done more revisions and touch-ups on, and so on. Bottom line is, whenever they first came out, no matter how crude they were, that’s when they’re from. And that’s a loose rule, because if something just “goes with” something else, it stays with it.

The general categories (what could be “discs” if the crazed fans insist on having a physical object in their hand) could be:

• Pre-historic (anything written before Insomnic Hallucinations)
• Unfinished Business (mostly songs written in high school)
• Through Forbidden Black Doors (started just after graduating high school)
• Open the Window/Leave of Absence
• Broken Wheel to Rival Big Bang song cycle (title?)
• Orphan songs (never been included on an “album”)
• Stuff with/by other people that I made significant contributions to

At least two of the items on that list would take up 2 discs each, which would be a total of nine discs. The Open the Window/Leave of Absence thing I’m not sure how to sort out, because they overlapped. I dumped Window on principle (because I didn’t like the “why” behind it), but then most everything on it was actually pretty good, so I remixed most of it for Leave of Absence and felt perfectly fine about it.

“Pre-historic” might not be enough material to fill a disc. Since I have working versions of several songs from a childhood album idea, I figured I’d assemble them into a skeletal mock-up of that, to give my inner thirteen-year old a “thank you” gift, as well as setting some kind of foundation for everything else. The “stuff by/with other people” might be thin too, so maybe they can share a disc. I guess I’d really only feature things I played lead guitar on.

Maybe I could have a disc of nothing but sketches, things that never got turned into full-blown songs but ought to be. I could call it You Finish The Damned Thing or something like that.

The unknowable black art of mastering

I don’t think mastering is really as subjective as people say it is.

I think it’s about providing a balanced spectrum to the ear, so that you can take a lot in without being fatigued. Having a poorly-mastered recording is akin to having a lot of potholes in the parking lot at an art gallery; the paintings are still intact, but you will enjoy them less because you’re worried about your car.

The ear adjusts to imbalances, or at least the brain does, so if you have a recording that’s bottom heavy, it can figure out that it needs to shift its attention to a weaker part of the spectrum. But this is work for the brain to do, and you’re putting more strain on yourself for no aesthetic advantage. If a listener is going to make any adjustments at all, consciously or otherwise, wouldn’t you rather they go from more strained to less strained as their mp3 player’s random shuffle leads them from someone else’s song into yours?

The visual equivalent of mastering is adjusting the brightness and contrast of a picture. I think there is rarely any advantage to not using the full black-to-white range — if you want it to be a “dark” picture, then yes, have larger dark areas overall, but at least one tiny area should hit full-on white (and vice versa for bright pictures). Gamma and similar adjustments are great for all the stuff in the middle, but there need to be points of reference at both of the extremes.

This brings my analogy back to mastering, sort of. What is the equivalent of “gamma” in mastering? Well, where in pictures we have colors (red, green, and blue being the primaries), in sound we have frequencies from low to high. (Think of frequency not as a simple tone, but as a thin slice of music that “rings” at that pitch.) For any given frequency, we really want it at some point to be loud and at some point to be quiet. To have a continuous full-volume tone at some pitch for the entire duration of a song would be torture. But to never have much energy in that part of the spectrum would leave the sound “hollowed out”, like there was a hole in it. So at any frequency you need to have things happening; you need to have a range of louds and quiets. “Audio gamma” would be how much or how little of the medium-volume sound is closer to full volume and how much is closer to silence. The maximum volume (for that frequency) isn’t really an aesthetic choice, it can be “loud enough to get your attention without hurting you”. The aesthetic choice is how much of the time to spend closer to that maximum.

As I go along, I’ll be using existing music as a model, to automatically adjust the overall sound of my own. Stay tuned!

Pathetic, that’s what it is

Well, I’m back from the hospital (pictured above is an abstract representation of my mental anguish over the last week). Sorry I didn’t write a post to tell you I was going to the emergency room! I don’t know where my priorities were.

My right foot — yes, the same one that was broken a few months back — was pwned by a severe staph infection, probably as a joint result of scratching maniacally and soaking in stale/dirty water. My goal now is to not contract anything new for the rest of December, until my health insurance kicks in. (That’s right, boys and girls, I’m still vulnerable!)

Okay, I did lie about the picture. That’s a still from a follow-up experiement with the auto-morphing code. Although you may get a neat looking frame here and there, by and large cross-breeding between two simultaneous video sources is a failure. For one thing, it takes absolutely forever to run for even a short amount of footage, because it has to analyze two images every time it generates one frame (I left it running for several weeks; it was still going the entire time I was hospitalized!). For another thing, because of the way individual frames are interpreted, very small differences in the input can yield very large differences in the output, so the end result is extremely jumpy.

So the final verdict is that it’s really only ideal for filling in missing frames in a single video sequence. Analyze few, generate many.