My blog has moved!

You should be automatically redirected in 5 seconds. If not, visit
http://www.mattmarksmusic.com.
Please update your bookmarks and RSS reader.

Friday, April 25, 2008

Phase Artifacts

So, I'm currently mastering a track which, if you have any experience, you know can be an insane adventure. I've been mastering my own tracks for a couple years, but I'm only now attempting to learn how it actually works (got me a book an' everythin'!). I recently downloaded a demo of Izotope Ozone which after 30 minutes already has my track louder and sounding better than after like a day of Mastering in my DAW, Ableton (yeah, I know... why would you ever master in Live...).

The thing is, mastering is way more technical than composing, arranging and general mixing. I'm scrubbing the rust off of the part of my brain that was at one time engaged for learning science (not that I was ever very good at it, mind you). But man, some of this shit flies higher over my head than Lindsay Lohan in a space shuttle.

While reading about the analog-modeling equalizer in Ozone, I came across the term phase artifacts. Guh? Ok, so the term artifacts in audio tends to refer to unwanted distortions or clippings that result from applying effects or maxing your signal. But phase? Phase in music and audio can refer to about 10 different things.

So I do a google search on phase artifacts. For fun, here are the top five reults:

Correction for interferometric synthetic aperture radar atmospheric phase artifacts using time series of zenith wet delay observations from a GPS network

Mitigation of tropospheric InSAR Phase artifacts through differential multisquint processing

Missing first points and phase artifact mutually entangled in FT NMR data—noniterative solution

Removal of phase artifacts from fMRI data using a Stockwell transform filter improves brain activity detection.

Although shade-off and halo patterns occur as a natural result of the phase contrast optical system, they are often referred to as phase artifacts or image distortions. In all forms of positive phase contrast, bright phase halos usually surround the boundaries between large specimen features and the medium.

I'm out of my league here people.

Seriously though, apparently none of these have to do with audio. I guess I'll tool around and eventually realize that it's some basic sonic concept given a fancy name. Music tech geeks are all about giving fancy names to things to make themselves feel more important.

Here a couple definitions while I'm at it:

DAW - Digital Audio Workstation = program for audio. That's it. Calling Garageband or Logic or Protools a DAW is like calling Microsoft Word a Digital Word-Processing Workstation. Why not call iTunes a Digital Audio Listening-Station.

Processing = Doin' stuff. Ya know, like digging lint out of your belly button is technically a process. I love when laptop musicians talk about "processing" audio:
"I'm applying some realtime processing to an incoming audio signal routed live into my DAW"
Translation: I just dropped reverb on my vocals.

Ok, well back to mastering. Oops, I mean um, applying a sonic revolution to my digital audio artifice.

3 comments:

Anonymous said...

yo, i work for izotope and saw your post come up through google alerts. hopefully i can clear up the phase artifacts thing simply.

when you see the word phase in regards to audio, think time. when comparing the phase relationship between two channels, what you are comparing is their relationship in time.

ozone's stereo separation module takes advantage of time delays to spread stuff out: because your brain uses time relationships between what's coming in your left and right ears to parse where that sound is located, if you use small delays between left and right you can make something that used to be straight down the center sound "wider." this is kind of wound into the Haas effect (which is worth looking up).

so where do the artifacts from phase relationships come from? mostly it is a problem if you ever have to convert your stereo tracks to mono, where phase cancellation will manifest as comb filtering (essentially a series of notch filters spreading across your mix...what you typically think of as the sound of 'phasing'). however, there can be some disorienting, undesirable effects in stereo mixes as well.

ozone's metering can tell you at a glance if you can expect problems due to phase. on the reverb and stereo imaging modules (the two places you're most likely to run into this sort of problem), there are phase and correlation meters. if the top of your phase image is leaning to the left, or if the correlation appears to be closer to -1 (again, the left side of the meter), you've got problems!

if you wanna hear how bad it is, bonk on the 'show channel ops' box under the phase meter and check 'mono'. this will sum your mix to mono within ozone.

the reason you read about phase artifacts while you were boning up on the equalizer is because of pre-echo, a general problem that can happen with digital (linear phase) filters or phase distortion, a general problem that can happen with analog-style (minimum phase) filters. there's a lot of technical reasons why phasing happens with filters, but mostly they are evident with big EQ changes... which, in the mastering stage of the game you really shouldn't be making (if your EQ picture is really that bad, you need to go back to the mix stage).

i hope this addressed your issue and brought some light to the often intimidating world of mastering. mastering is just as artistic as the rest of the music making process, but does require a different set of ears... a set that listens for technical problems in a mix as well as aesthetic problems. if you have any other questions, feel free to contact me at the name above plus @izotope.com



also, your music is rad. you take good advantage of space & timing and it's very nice.

Babby Daddy said...
This comment has been removed by the author.
Anonymous said...

oops, made a mistake. i forgot that ozone's phase meter is different from some other phase meters. where i said the top part of the phase meter will lean to the left, i actually should've said the picture on the meter will start to look like a squashed tomato, or at least closer to a horizontal line than a nice, full round (mostly vertical), properly phase aligned mix would.