VVC 2013: audio editing panel
I thought I had made notes for this panel, but apparently not, which makes sense since 
nestra and I didn't do much planning; she can talk about music theory in her sleep, and all I did was show up with two .mp3s and a handout with lyrics and numbers on it.
As it turns out, I think it's good that we didn't over-prepare, because the audience we got was frankly not the audience I was expecting, and since the plan was basically "so, yeah, we'll talk about stuff" it was pretty easy to switch direction on the fly. I did have a moment of "...but what are you all doing here?" when it became clear that at least half the audience had already done quite a bit of audio editing; I had proposed the panel and roped
nestra into participating with the intended audience of people who wanted to edit audio but didn't know where to begin! But it was great to be able to crowdsource software recommendations and general advice.
nestra went over some basics -- time signature, beat, measure, downbeat, upbeat, musical phrase, the fact that math is at the heart of music -- and there was enough nodding that I should have guessed right then that we weren't dealing entirely with novices, but sometimes I am slow (and also I am used to my students, who nod supportively even when they have no clue what's going on, bless them).
But I did finally catch on, so we paused and regrouped and I tried to get a sense of what people actually wanted/needed to know, which... I'm not sure how well I did, but the group had some good back-and-forth. And then we moved on to the .mp3s: the original version of The Dandy Warhols' "We Used to be Friends" (3:20) followed by the edited-for-the-Veronica-Mars-credits version (0:30). The goal was to not get bogged down in where one could or should cut and instead look at how it was cut and what principles we could pull out from those decisions in terms of how to make edits appear seamless (sound) and, less obvious but just as important, make the trimmed version sound balanced (structure).
We tackled structure first. Obviously the easiest thing is to cut an entire self-contained chunk of the song: a repeated chorus, a verse. To trim more or differently, the basic principle is to cut things in half -- which is what the VM credits editor did: the credits version uses half of a verse and half of a chorus. Figuring out the musical patterns that characterize a song is the first and most important step toward editing the song, because the editing choices need to preserve rather than disrupt those patterns. In this song, the lyrical lines and musical phrases map onto each other, for the most part (one of the panel attendees cleverly spotted an exception right in the first two lines); each line is eight beats, each verse and chorus has four lines. This pattern is not universal, but it is typical -- like, the single most common pattern in pop music. (And not just pop music either; a similar pattern is found in Beethoven's violin concerto and in fact vast swaths of Western music generally.) These verse and chorus units can usually be split in a couple of different ways; the VM edit uses the first two lines of the first verse and the first and fourth lines of the second chorus, which, again, isn't going to work for every song but is a pretty reasonable starting point: first two, last two, or first and last are likely to work; using the middle two will almost never work because doing so inverts the underlying musical pattern. The bigger picture is important too: using half a verse but the whole chorus (or vice versa) is likely to sound unbalanced -- but there are ways around this: consistently using half a verse and a whole chorus creates a new pattern, based on but not identical to the original, that sounds intentional and therefore okay. But using half a verse, then a full chorus, then a full verse, then half a chorus... in most if not all cases, that is just going to sound weird, even if the individual cuts are made with laserlike precision.
Then we talked a little bit about sound. Techno and other electronic or synth-heavy genres are super-forgiving, for the most part, because most of the sounds aren't organic and the beats tend to be quite sharply delineated; I've sometimes done this kind of editing right in Premiere because all you need is a couple of hard cuts in the right places and you're done. Piano- and guitar-based songs are trickier, and guitar solos are deeply annoying but still do-able, as I can attest. The key is to match downbeats -- not necessarily to cut on the downbeat, but line up the downbeats -- put the two clips on different tracks and then literally line up the spikes in the waveforms -- and then try cutting at the low points of the waveform right before the spikes. If that doesn't work (or would work but there are vocals in the way on the upbeat or whatever), nudge the cut points around -- keeping the downbeats lined up! -- until something works. (At this point we started drawing waveform mockups on the pad, which... let's just say I'm no artist.) And if the transition point between clips needs some smoothing, as often happens, there's always the possibility of a crossfade -- either a linear crossfade or, if the software allows it, logarithmic or sinusoidal cross-fades, both of which I use all the time when audio editing; they are two of my favorite things about Adobe Audition.
And I think that's it!
Originally posted at Dreamwidth || Read
comments on Dreamwidth
nestra and I didn't do much planning; she can talk about music theory in her sleep, and all I did was show up with two .mp3s and a handout with lyrics and numbers on it.
As it turns out, I think it's good that we didn't over-prepare, because the audience we got was frankly not the audience I was expecting, and since the plan was basically "so, yeah, we'll talk about stuff" it was pretty easy to switch direction on the fly. I did have a moment of "...but what are you all doing here?" when it became clear that at least half the audience had already done quite a bit of audio editing; I had proposed the panel and roped
nestra into participating with the intended audience of people who wanted to edit audio but didn't know where to begin! But it was great to be able to crowdsource software recommendations and general advice.
nestra went over some basics -- time signature, beat, measure, downbeat, upbeat, musical phrase, the fact that math is at the heart of music -- and there was enough nodding that I should have guessed right then that we weren't dealing entirely with novices, but sometimes I am slow (and also I am used to my students, who nod supportively even when they have no clue what's going on, bless them).
But I did finally catch on, so we paused and regrouped and I tried to get a sense of what people actually wanted/needed to know, which... I'm not sure how well I did, but the group had some good back-and-forth. And then we moved on to the .mp3s: the original version of The Dandy Warhols' "We Used to be Friends" (3:20) followed by the edited-for-the-Veronica-Mars-credits version (0:30). The goal was to not get bogged down in where one could or should cut and instead look at how it was cut and what principles we could pull out from those decisions in terms of how to make edits appear seamless (sound) and, less obvious but just as important, make the trimmed version sound balanced (structure).
We tackled structure first. Obviously the easiest thing is to cut an entire self-contained chunk of the song: a repeated chorus, a verse. To trim more or differently, the basic principle is to cut things in half -- which is what the VM credits editor did: the credits version uses half of a verse and half of a chorus. Figuring out the musical patterns that characterize a song is the first and most important step toward editing the song, because the editing choices need to preserve rather than disrupt those patterns. In this song, the lyrical lines and musical phrases map onto each other, for the most part (one of the panel attendees cleverly spotted an exception right in the first two lines); each line is eight beats, each verse and chorus has four lines. This pattern is not universal, but it is typical -- like, the single most common pattern in pop music. (And not just pop music either; a similar pattern is found in Beethoven's violin concerto and in fact vast swaths of Western music generally.) These verse and chorus units can usually be split in a couple of different ways; the VM edit uses the first two lines of the first verse and the first and fourth lines of the second chorus, which, again, isn't going to work for every song but is a pretty reasonable starting point: first two, last two, or first and last are likely to work; using the middle two will almost never work because doing so inverts the underlying musical pattern. The bigger picture is important too: using half a verse but the whole chorus (or vice versa) is likely to sound unbalanced -- but there are ways around this: consistently using half a verse and a whole chorus creates a new pattern, based on but not identical to the original, that sounds intentional and therefore okay. But using half a verse, then a full chorus, then a full verse, then half a chorus... in most if not all cases, that is just going to sound weird, even if the individual cuts are made with laserlike precision.
Then we talked a little bit about sound. Techno and other electronic or synth-heavy genres are super-forgiving, for the most part, because most of the sounds aren't organic and the beats tend to be quite sharply delineated; I've sometimes done this kind of editing right in Premiere because all you need is a couple of hard cuts in the right places and you're done. Piano- and guitar-based songs are trickier, and guitar solos are deeply annoying but still do-able, as I can attest. The key is to match downbeats -- not necessarily to cut on the downbeat, but line up the downbeats -- put the two clips on different tracks and then literally line up the spikes in the waveforms -- and then try cutting at the low points of the waveform right before the spikes. If that doesn't work (or would work but there are vocals in the way on the upbeat or whatever), nudge the cut points around -- keeping the downbeats lined up! -- until something works. (At this point we started drawing waveform mockups on the pad, which... let's just say I'm no artist.) And if the transition point between clips needs some smoothing, as often happens, there's always the possibility of a crossfade -- either a linear crossfade or, if the software allows it, logarithmic or sinusoidal cross-fades, both of which I use all the time when audio editing; they are two of my favorite things about Adobe Audition.
And I think that's it!
Originally posted at Dreamwidth || Read
comments on Dreamwidth