WEBVTT
1
00:00:00.230 --> 00:00:02.270
You know, usually when we talk about a film or
2
00:00:02.310 --> 00:00:06.580
TV scene just perfectly matching a piece of music, there's
3
00:00:07.219 --> 00:00:10.619
this expectation of pure, effortless magic. Oh, totally.
4
00:00:10.660 --> 00:00:12.380
Like it just fell out of the sky. Right.
5
00:00:12.759 --> 00:00:14.859
Like you just take the visual, you drop the song
6
00:00:14.919 --> 00:00:17.850
on top and boom, you know, cinematic gold. Yeah.
7
00:00:17.969 --> 00:00:21.410
I mean, it feels completely intuitive to you, the viewer,
8
00:00:21.469 --> 00:00:24.230
sitting on your couch. The assumption is simply that a director,
9
00:00:24.870 --> 00:00:27.410
I don't know, heard a cool track on Spotify, clicked
10
00:00:27.429 --> 00:00:29.410
a button and just dropped it into the timeline.
11
00:00:30.050 --> 00:00:32.859
But then you actually step into the reality of audio
12
00:00:32.880 --> 00:00:37.539
post-production and sync licensing. And suddenly that effortless magic looks
13
00:00:37.740 --> 00:00:39.859
a lot more like a really high-stakes puzzle.
14
00:00:40.259 --> 00:00:41.979
Oh, absolutely. It's a massive puzzle.
15
00:00:42.520 --> 00:00:47.130
You're suddenly staring at this landscape of clashing audio frequencies,
16
00:00:47.530 --> 00:00:52.570
missing digital files, and legal tripwires that is genuinely murky.
17
00:00:53.049 --> 00:00:57.460
Yeah. The modern sync placement is basically the absolute definition
18
00:00:57.539 --> 00:01:01.859
of a creative and administrative minefield. Winning requires navigating both
19
00:01:01.920 --> 00:01:03.340
sides flawlessly.
20
00:01:03.640 --> 00:01:06.060
Which is exactly why our mission for this deep dive
21
00:01:06.519 --> 00:01:08.620
is to guide you through that exact minefield.
22
00:01:08.640 --> 00:01:09.439
Bring your hard hats.
23
00:01:10.060 --> 00:01:14.859
Seriously, we are unpacking the wild, rapidly evolving ecosystem of
24
00:01:14.939 --> 00:01:19.010
TV and film music syncs. And we're tailoring this specifically
25
00:01:19.049 --> 00:01:22.489
for all you indie publishers, songwriters, and producers trying to
26
00:01:22.530 --> 00:01:24.269
keep up with the 2026 landscape.
27
00:01:24.590 --> 00:01:26.290
It is a lot to keep up with right now.
28
00:01:26.590 --> 00:01:30.260
It really is. So we're synthesizing insights from across the
29
00:01:30.299 --> 00:01:33.780
industry today. We've got industry guides from CD Baby and
30
00:01:33.859 --> 00:01:37.959
Clint Music, technical manuals on AI stem extraction from AI
31
00:01:38.019 --> 00:01:39.239
Magix and the Postflow.
32
00:01:39.400 --> 00:01:40.980
Some really fascinating data in there, yeah.
33
00:01:41.280 --> 00:01:44.079
And fresh data from sound charts and Billboard from the
34
00:01:44.120 --> 00:01:48.489
latest Prestige TV finales. Okay, let's unpack this. Because before
35
00:01:48.530 --> 00:01:52.349
we even worry about contracts and technical asset delivery, we
36
00:01:52.370 --> 00:01:54.849
have to start with the creative DNA of a sync.
37
00:01:55.430 --> 00:01:56.769
Right, the actual music.
38
00:01:57.150 --> 00:02:00.260
Exactly. What actually makes a piece of music attractive to
39
00:02:00.280 --> 00:02:01.739
a music supervisor today?
40
00:02:02.079 --> 00:02:06.480
Well, that underlying philosophy has completely inverted over the last decade. Historically,
41
00:02:06.540 --> 00:02:09.840
there was this prevailing myth in the industry that sync
42
00:02:09.909 --> 00:02:12.770
music needed to be vague or universal.
43
00:02:13.360 --> 00:02:17.800
Right, like that standard mid-2000s indie folk or commercial pop
44
00:02:17.840 --> 00:02:19.300
playing over network TV dramas.
45
00:02:19.680 --> 00:02:23.900
Exactly. Those songs were literally engineered to broadly apply to
46
00:02:23.960 --> 00:02:26.659
anyone feeling a little bit sad or a little bit happy,
47
00:02:26.780 --> 00:02:27.020
you know?
48
00:02:27.259 --> 00:02:28.780
Just keeping it right down the middle.
49
00:02:29.680 --> 00:02:33.539
The goal was low-friction background music. But today, the sheer
50
00:02:33.580 --> 00:02:37.030
volume of streaming media has exposed viewers to highly nuanced,
51
00:02:37.159 --> 00:02:40.699
diverse storytelling— Directors don't want background music anymore. They want
52
00:02:40.759 --> 00:02:44.960
a co-star in the scene. Music supervisors actively seek out quirky,
53
00:02:45.099 --> 00:02:46.219
highly specific songs.
54
00:02:46.539 --> 00:02:49.909
It's essentially like trying to bake a generic vanilla cake
55
00:02:50.409 --> 00:02:52.469
that you think everyone in the room is going to tolerate.
56
00:02:53.110 --> 00:02:56.530
But the music supervisor is actually frantically searching the city for, like,
57
00:02:57.050 --> 00:02:59.090
a habanero-infused chocolate tart.
58
00:02:59.689 --> 00:03:00.210
Exactly.
59
00:03:00.849 --> 00:03:01.729
You have to be the tart.
60
00:03:02.229 --> 00:03:02.810
Be the tart.
61
00:03:03.330 --> 00:03:03.750
I love that.
62
00:03:04.069 --> 00:03:06.210
And if you look at the current guidance being handed
63
00:03:06.250 --> 00:03:09.780
down to DIY musicians from places like CD Baby, the
64
00:03:09.819 --> 00:03:11.020
warning is explicit.
65
00:03:11.500 --> 00:03:12.199
What are they telling them?
66
00:03:12.520 --> 00:03:14.759
They're saying do not round off your creative edges to
67
00:03:14.780 --> 00:03:16.280
be more syncable. Really?
68
00:03:17.099 --> 00:03:18.020
So keep the weirdness?
69
00:03:18.500 --> 00:03:23.120
Yes. Removing a strange time signature, an unusual instrument, or
70
00:03:23.159 --> 00:03:25.580
a highly specific lyric just to play it safe will
71
00:03:25.620 --> 00:03:28.319
likely remove the very element a music supervisor would have
72
00:03:28.379 --> 00:03:29.319
used to anchor a scene.
73
00:03:29.340 --> 00:03:30.770
Oh, wow.
74
00:03:31.050 --> 00:03:31.250
Yeah.
75
00:03:31.900 --> 00:03:34.159
And a brilliant example of this just played out in
76
00:03:34.199 --> 00:03:36.699
the season three finale of Netflix's Heartstopper.
77
00:03:36.919 --> 00:03:37.659
Oh, that was huge.
78
00:03:37.979 --> 00:03:41.099
It was. Looking at the latest weekly TV Music Sync
79
00:03:41.159 --> 00:03:44.300
research report from October 3rd, they closed out the season
80
00:03:44.319 --> 00:03:47.419
with this incredibly emotional needle drop using I Know the
81
00:03:47.520 --> 00:03:48.680
End by Phoebe Bridgers.
82
00:03:49.000 --> 00:03:50.479
Such a great track. It is.
83
00:03:50.879 --> 00:03:54.370
And this track from her 2020 album Punisher is the
84
00:03:54.560 --> 00:03:57.469
exact antithesis of a standard vague pop song.
85
00:03:57.629 --> 00:03:58.090
Completely.
86
00:03:58.110 --> 00:04:00.889
I mean, it's a nearly six minute chamber pop suite.
87
00:04:01.469 --> 00:04:05.349
It starts as this hushed, intimate, acoustic ballad about depression
88
00:04:05.469 --> 00:04:07.050
and driving up the California coast.
89
00:04:07.150 --> 00:04:08.969
And then it just shifts gears entirely.
90
00:04:08.990 --> 00:04:14.500
It literally detonates into a massive, metal-inspired, apocalyptic screaming match.
91
00:04:15.159 --> 00:04:18.420
We are talking chaotic horns, a massive choir, and a
92
00:04:18.500 --> 00:04:20.879
blistering lead guitar part by Nick Zinner of the Yeah
93
00:04:20.899 --> 00:04:23.459
Yeah Yes. It is wonderfully weird.
94
00:04:23.600 --> 00:04:24.240
It really is.
95
00:04:24.259 --> 00:04:27.420
I mean, Bridgers used to perform this on NPR's tiny
96
00:04:27.459 --> 00:04:30.769
desk from inside a fake Oval Office. And she smashed
97
00:04:30.829 --> 00:04:32.269
a monitor playing it on SNL.
98
00:04:32.569 --> 00:04:35.899
See, it's an uncompromising creative vision, which is exactly why
99
00:04:36.019 --> 00:04:37.959
it works so brilliantly on screen. Yeah.
100
00:04:38.259 --> 00:04:39.680
A generic track wouldn't have done that.
101
00:04:39.860 --> 00:04:43.660
Not at all. That screaming brass and guitar finale layered
102
00:04:43.720 --> 00:04:47.250
over a pivotal crossroad moment for the main characters in Heartstopper.
103
00:04:47.329 --> 00:04:50.209
It just created an overwhelming cinematic payoff.
104
00:04:50.269 --> 00:04:50.449
Right.
105
00:04:50.730 --> 00:04:54.410
A generic sweeping orchestral cue would have felt entirely unearned there.
106
00:04:54.879 --> 00:04:57.620
The specificity of the chaos in the track mirrors the
107
00:04:57.800 --> 00:05:00.180
internal emotional chaos of the characters.
108
00:05:00.600 --> 00:05:04.350
And we are seeing this demand for hyper-specific friction everywhere
109
00:05:04.379 --> 00:05:07.170
right now. Like, as we move into the fall streaming
110
00:05:07.209 --> 00:05:11.490
season for horror and psychological thrillers across platforms like Shudder
111
00:05:11.529 --> 00:05:12.009
and Hulu.
112
00:05:12.149 --> 00:05:14.029
Oh, the briefs for those are fascinating right now.
113
00:05:14.290 --> 00:05:19.379
Right. Music supervisors are actively issuing briefs for Gothic Americana
114
00:05:19.819 --> 00:05:21.439
and Minimalist Dark Folk.
115
00:05:22.100 --> 00:05:25.680
Yeah, they are actively hunting for high friction acoustic arrangements.
116
00:05:26.399 --> 00:05:31.980
They want like solitary acoustic guitars, bowed bass, haunting low
117
00:05:32.019 --> 00:05:36.290
register vocals, and really dark, non-romantic lyrics.
118
00:05:36.769 --> 00:05:40.990
Exactly. They want all those weird atmospheric imperfections left in
119
00:05:41.009 --> 00:05:43.829
the track to give the sound design room to breathe. Yeah.
120
00:05:44.250 --> 00:05:48.420
And what's fascinating here is that this desire for uncompromising,
121
00:05:48.660 --> 00:05:52.139
specific artistic vision isn't actually a new phenomenon.
122
00:05:52.319 --> 00:05:52.540
No.
123
00:05:52.839 --> 00:05:56.060
No, it's just becoming the mandatory standard. If we connect
124
00:05:56.079 --> 00:06:00.339
this to historical sync perfection, Blondie's 1980 hit Call Me
125
00:06:00.620 --> 00:06:02.040
serves as the ultimate blueprint.
126
00:06:02.379 --> 00:06:03.939
Oh, from American Gigolo.
127
00:06:04.199 --> 00:06:08.990
Yes. The legendary Italian disco producer Giorgio Marotta was scoring
128
00:06:09.029 --> 00:06:12.120
that neo-noir film. And he possessed an instrumental track called
129
00:06:12.379 --> 00:06:14.850
Man Machine that he initially pitched to Stevie Nicks.
130
00:06:15.189 --> 00:06:16.209
Oh, I didn't know that. Yeah.
131
00:06:16.750 --> 00:06:19.709
But contract conflicts prevented her from taking it. So he
132
00:06:19.750 --> 00:06:22.350
brings the instrumental to Debbie Harry. He asks her to
133
00:06:22.389 --> 00:06:24.930
write the lyrics. And rather than penning a vague pop
134
00:06:25.009 --> 00:06:28.240
anthem about love, she writes lyrics specifically tailored to the
135
00:06:28.279 --> 00:06:31.839
psychology of the film's main character, Richard Gere's male prostitute,
136
00:06:31.879 --> 00:06:32.500
Julian Kay.
137
00:06:32.879 --> 00:06:35.439
So she put herself completely inside the head of the
138
00:06:35.500 --> 00:06:37.199
character on screen. Exactly. Exactly.
139
00:06:37.720 --> 00:06:40.560
She pictured the opening visual, which is Julian driving his
140
00:06:40.620 --> 00:06:44.279
Mercedes convertible along the coast of California, and matched the
141
00:06:44.339 --> 00:06:46.160
psychological energy flawlessly.
142
00:06:46.360 --> 00:06:47.699
That's brilliant.
143
00:06:48.019 --> 00:06:50.310
The track is rooted in D minor, driven by a
144
00:06:50.350 --> 00:06:54.629
blistering 143 beats per minute dance rock tempo. It physically
145
00:06:54.750 --> 00:06:56.870
pushes the visual momentum of the car forward.
146
00:06:57.269 --> 00:07:00.029
And it paid off. I mean, it spent six consecutive
147
00:07:00.069 --> 00:07:01.790
weeks at number one on the Billboard Hot 100 in 1980. Right.
148
00:07:03.430 --> 00:07:06.790
Both Bridgers and Blondie prove that tailoring music to specific
149
00:07:06.850 --> 00:07:10.420
psychological states yields the most legendary cinematic payoffs.
150
00:07:11.569 --> 00:07:15.000
So let's follow that thread forward. Say you have captured
151
00:07:15.019 --> 00:07:19.100
that magic. You've embraced the weirdness. You've written the habanero tart.
152
00:07:19.759 --> 00:07:22.339
And the supervisor places it in a massive streaming show.
153
00:07:22.879 --> 00:07:24.300
You get the yes. Right.
154
00:07:24.519 --> 00:07:28.189
You get the yes. What actually happens in the aftermath? Obviously,
155
00:07:28.209 --> 00:07:31.050
you collect the upfront licensing fee, but What is the
156
00:07:31.310 --> 00:07:35.569
ongoing real-world value to the artist in a fragmented digital landscape?
157
00:07:36.089 --> 00:07:38.149
Well, we really have to examine the Shazam effect.
158
00:07:38.810 --> 00:07:40.110
Okay, the Shazam effect.
159
00:07:40.569 --> 00:07:44.009
Yeah, market data from places like sound charts consistently shows
160
00:07:44.050 --> 00:07:47.470
that Shazam is the ultimate early indicator of a cultural hit.
161
00:07:48.069 --> 00:07:50.339
Why is that? Just because people are curious?
162
00:07:50.680 --> 00:07:53.319
Because it captures active, high-intent discovery.
163
00:07:53.740 --> 00:07:54.019
Okay.
164
00:07:54.439 --> 00:07:56.500
When a viewer is watching a show and pulls out
165
00:07:56.540 --> 00:07:59.139
their phone to identify a track, they are crossing a
166
00:07:59.399 --> 00:08:00.839
massive behavioral threshold.
167
00:08:01.019 --> 00:08:01.959
Yeah, they're taking action.
168
00:08:02.160 --> 00:08:05.220
Right. They are no longer passively consuming background noise. They
169
00:08:05.240 --> 00:08:08.199
are demanding to know the identity of the artist. Spikes
170
00:08:08.259 --> 00:08:12.069
and Shazam counts almost universally precede streaming and radio growth
171
00:08:12.110 --> 00:08:13.230
by days or weeks.
172
00:08:13.750 --> 00:08:17.089
It's the purest signal that a track is rapidly gaining traction.
173
00:08:17.189 --> 00:08:17.670
Exactly.
174
00:08:17.990 --> 00:08:19.970
But let me push back on the scope of that
175
00:08:20.009 --> 00:08:23.069
for a second. Sure. If a music supervisor uses a
176
00:08:23.129 --> 00:08:27.240
song by an independent, completely unknown artist— A Shazam spike
177
00:08:27.379 --> 00:08:31.980
obviously translates to massive discovery. But supervisors frequently use tracks
178
00:08:32.039 --> 00:08:35.980
by massive, instantly recognizable legacy artists. Yeah, they do. Like,
179
00:08:36.139 --> 00:08:38.440
does a Shazam spike actually move the needle for a
180
00:08:38.480 --> 00:08:43.639
song everyone ostensibly already knows? Aren't discovery tools inherently for
181
00:08:43.679 --> 00:08:44.379
the undiscovered?
182
00:08:44.730 --> 00:08:47.570
It is a logical assumption, but it overlooks the mechanics
183
00:08:47.610 --> 00:08:49.149
of catalog revitalization.
184
00:08:49.169 --> 00:08:50.169
Okay, explain that.
185
00:08:50.490 --> 00:08:53.870
Syncs are the engine of back-end performance royalties. Every time
186
00:08:53.889 --> 00:08:57.059
that show airs in international markets... moves to a secondary
187
00:08:57.100 --> 00:09:00.360
streaming service, or is syndicated, it generates publishing royalties.
188
00:09:00.539 --> 00:09:01.279
Right, that makes sense.
189
00:09:01.539 --> 00:09:05.090
But more importantly, Async introduces legacy tracks to a younger
190
00:09:05.129 --> 00:09:08.029
demographic who might not possess any cultural memory of a
191
00:09:08.549 --> 00:09:08.850
1980s hit.
192
00:09:09.330 --> 00:09:10.710
Oh, that's true. Right? They weren't there.
193
00:09:10.970 --> 00:09:13.929
Right. So when those younger viewers Shazam the track, it
194
00:09:14.009 --> 00:09:18.269
feeds a cascading algorithm. The Shazam tags trigger an influx
195
00:09:18.340 --> 00:09:21.919
of Spotify searches, which pushes the song into algorithmic playlists
196
00:09:21.980 --> 00:09:25.830
like Discover Weekly, triggering a massive compounding effect across the
197
00:09:25.929 --> 00:09:27.409
entire streaming ecosystem.
198
00:09:27.830 --> 00:09:31.090
We actually have a flawless real-world example of this algorithmic
199
00:09:31.110 --> 00:09:34.450
cascade happening right now with HBO's The Penguin.
200
00:09:34.889 --> 00:09:36.740
Oh yeah, the needle drops on that show are wild.
201
00:09:37.379 --> 00:09:40.440
They really are. The music supervisors on that project, Jen
202
00:09:40.460 --> 00:09:44.059
Malone and Whitney Piltzer, are conducting an absolute masterclass in
203
00:09:44.139 --> 00:09:46.120
ironic high-contrast needle drops.
204
00:09:46.519 --> 00:09:51.019
Yeah, they are placing highly recognizable tracks underneath ink incredibly tense,
205
00:09:51.019 --> 00:09:52.480
gritty crime scenes.
206
00:09:52.480 --> 00:09:55.500
Like in episode one, Oz and Victor get into a
207
00:09:55.500 --> 00:09:58.509
car and Tolly Parton's 9 to 5 plays on the radio.
208
00:09:58.820 --> 00:09:59.169
So good.
209
00:09:59.500 --> 00:10:02.779
And in episode two, Bobby Darin's Call Me Irresponsible scores
210
00:10:02.799 --> 00:10:06.139
a brutal shootout. They are pulling in Gil Scott Heron's
211
00:10:06.179 --> 00:10:08.500
Me and the Devil and The Cures of Forest.
212
00:10:09.200 --> 00:10:13.960
By juxtaposing those evocative, familiar tracks against high-stakes mob violence,
213
00:10:14.480 --> 00:10:17.220
they instantly drive viewers to their phones to re-engage with
214
00:10:17.240 --> 00:10:17.580
the music.
215
00:10:18.139 --> 00:10:20.950
And the weekly sync data really illustrates the sheer power
216
00:10:20.970 --> 00:10:24.070
of that juxtaposition. Like in episode three of The Penguin,
217
00:10:24.450 --> 00:10:27.769
during a fast-paced vehicle chase sequence through Gotham's industrial district,
218
00:10:28.210 --> 00:10:29.769
they featured Blondie's Call Me.
219
00:10:30.169 --> 00:10:31.629
A track from 1980.
220
00:10:31.629 --> 00:10:35.250
Right, from 1980. And over the following seven days, Shazam
221
00:10:35.269 --> 00:10:37.889
recorded a top 10 global spike for the song.
222
00:10:38.409 --> 00:10:38.909
It's amazing.
223
00:10:39.269 --> 00:10:42.570
That isolated moment drove a massive wave of fresh streaming
224
00:10:42.610 --> 00:10:45.470
discovery among a younger demographic watching the crime saga.
225
00:10:45.919 --> 00:10:49.240
And the Heartstopper finale yielded an even more dramatic metric.
226
00:10:49.259 --> 00:10:50.220
Oh, yeah. What happened there?
227
00:10:50.860 --> 00:10:54.120
Following that sync, Phoebe Bridger's 2020 master of I Know
228
00:10:54.159 --> 00:10:59.259
the End experienced an immediate 410% surge in official U.S.
229
00:10:59.330 --> 00:11:00.610
Audio streams.
230
00:11:00.610 --> 00:11:00.690
410%.
231
00:11:00.690 --> 00:11:03.509
That's insane. That's insane. Yeah.
232
00:11:04.029 --> 00:11:06.529
Shooting straight to number one on the Billboard Top TV
233
00:11:06.570 --> 00:11:08.169
Songs chart for the week of October 3rd.
234
00:11:08.649 --> 00:11:12.289
Okay. Here's where it gets really interesting, though. Okay. You
235
00:11:12.309 --> 00:11:15.669
can write the brilliant, specific song. You can know exactly
236
00:11:15.730 --> 00:11:19.269
how massive that Shazam-driven streaming boost is going to be
237
00:11:19.309 --> 00:11:22.429
on the other side. But there is a heartbreaking reality
238
00:11:22.769 --> 00:11:24.480
hiding in the shadows of this business.
239
00:11:24.779 --> 00:11:27.100
Oh, boy. The administrative side.
240
00:11:27.299 --> 00:11:31.179
Exactly. You can win the creative lottery, get chosen by
241
00:11:31.220 --> 00:11:34.580
the music supervisor, and completely lose the deal on a
242
00:11:34.639 --> 00:11:39.649
purely administrative technicality. It's brutal. The biggest hurdle to actually
243
00:11:39.730 --> 00:11:42.929
getting your music on television isn't always writing a great hook.
244
00:11:43.750 --> 00:11:44.669
It's file management.
245
00:11:45.029 --> 00:11:48.350
It serves as the great unforgiving filter of the sync
246
00:11:48.389 --> 00:11:51.539
licensing world. Industry producers, like the folks at Clint Music,
247
00:11:51.919 --> 00:11:53.399
constantly warn about this bottleneck.
248
00:11:54.139 --> 00:11:56.860
Because supervisors are just too busy. Yes.
249
00:11:57.299 --> 00:12:01.710
Music supervisors are functioning under immense pressure. They are simultaneously
250
00:12:01.730 --> 00:12:06.129
juggling hundreds of audio cues, managing tight post-production deadlines, and
251
00:12:06.169 --> 00:12:09.509
appeasing demanding directors. So when a producer finally gets a
252
00:12:09.570 --> 00:12:13.389
yes on a creative pitch, the supervisor immediately requests the stems.
253
00:12:13.929 --> 00:12:16.350
And for those who haven't had to deliver a commercial project,
254
00:12:16.710 --> 00:12:19.929
stems are the individual isolated audio tracks that make up
255
00:12:19.960 --> 00:12:23.559
the final song. Right. So the isolated lead vocal, the
256
00:12:23.720 --> 00:12:27.899
isolated drum kit, the isolated bass guitar. The video editor
257
00:12:27.960 --> 00:12:31.320
needs those individual pieces so they can dip the volume
258
00:12:31.360 --> 00:12:33.929
of the vocals when the actors are speaking or perhaps
259
00:12:34.070 --> 00:12:37.049
use just the drum beat to build tension before cutting
260
00:12:37.070 --> 00:12:38.230
to commercial. Exactly.
261
00:12:38.570 --> 00:12:41.409
And the breakdown happens when a producer sends over a
262
00:12:41.549 --> 00:12:42.870
chaotic digital folder.
263
00:12:43.370 --> 00:12:45.220
Like what? What are they doing wrong?
264
00:12:45.460 --> 00:12:48.299
Well, the files might possess the wrong sample rates or
265
00:12:48.360 --> 00:12:51.639
bit depths, which is basically the audio equivalent of submitting
266
00:12:52.080 --> 00:12:56.080
a low-resolution, heavily pixelated photograph to a high-end magazine.
267
00:12:56.100 --> 00:12:56.799
Ooh, yikes.
268
00:12:57.139 --> 00:13:00.450
Or the files are lazily named things like, Final Drums V3,
269
00:13:00.870 --> 00:13:01.450
use this one.
270
00:13:01.889 --> 00:13:03.429
We've all named a file that way. We have.
271
00:13:04.120 --> 00:13:07.220
But if a music supervisor opens a folder and encounters
272
00:13:07.539 --> 00:13:11.500
an unorganized disaster, they do not have the time to
273
00:13:11.600 --> 00:13:12.500
act as your tech support.
274
00:13:12.990 --> 00:13:14.879
They won't just email you back and ask for the
275
00:13:14.919 --> 00:13:16.019
right ones? Nope.
276
00:13:16.970 --> 00:13:19.549
They will simply close the folder, delete the email, and
277
00:13:19.590 --> 00:13:21.529
move to the next producer on their list who has
278
00:13:21.570 --> 00:13:22.269
their act together.
279
00:13:22.570 --> 00:13:24.370
That is a brutal reality check.
280
00:13:24.850 --> 00:13:28.090
You pour your soul into the creative process, you secure
281
00:13:28.129 --> 00:13:31.090
the placement of a lifetime, and you lose the payout
282
00:13:31.149 --> 00:13:35.029
because your folder hierarchy was a mess. Professionalism in asset
283
00:13:35.070 --> 00:13:38.049
delivery carries the exact same weight as the songwriting itself.
284
00:13:38.509 --> 00:13:41.730
The music supervisors who consistently return to an independent publisher
285
00:13:41.870 --> 00:13:44.850
are doing so because that publisher is utterly frictionless to
286
00:13:44.889 --> 00:13:45.230
work with.
287
00:13:45.690 --> 00:13:46.570
That makes a lot of sense.
288
00:13:47.149 --> 00:13:51.830
If your stems are pristine, intuitively organized, and perfectly formatted,
289
00:13:52.350 --> 00:13:54.950
you transition from being just another artist to being a
290
00:13:55.190 --> 00:13:57.250
reliable problem solver for their production.
291
00:13:57.269 --> 00:14:01.990
Okay, I understand the demand for perfect organization. But I
292
00:14:02.029 --> 00:14:04.440
want to challenge this workflow for a second because we
293
00:14:04.539 --> 00:14:08.279
are living in 2026. Okay, hit me. Let's say I'm
294
00:14:08.320 --> 00:14:14.139
an indie publisher representing a brilliant, authentic 1960s analog soul recording. Sure.
295
00:14:14.559 --> 00:14:17.750
Or maybe I'm a modern bedroom producer who simply lost
296
00:14:17.759 --> 00:14:19.970
the original hard drive for a track I produced five
297
00:14:20.009 --> 00:14:23.389
years ago. I physically do not have the multi-track stems.
298
00:14:23.529 --> 00:14:24.009
It happens.
299
00:14:25.090 --> 00:14:28.250
Can't the film's video editor just run the stereo file
300
00:14:28.269 --> 00:14:31.019
through an AI tool and rip the vocals out? Like,
301
00:14:31.210 --> 00:14:32.870
separating a mix song used to be like trying to
302
00:14:33.009 --> 00:14:35.409
unstir a cup of coffee or unbake a cake, right?
303
00:14:35.669 --> 00:14:37.110
Yeah, that was the old analogy.
304
00:14:37.470 --> 00:14:41.690
But hasn't artificial intelligence completely solved the unbaking problem by now?
305
00:14:42.389 --> 00:14:44.490
Why am I losing the deal if the editor can
306
00:14:44.509 --> 00:14:47.129
just click a button and generate the stems themselves?
307
00:14:47.429 --> 00:14:50.090
To understand why relying on the editor to do the
308
00:14:50.149 --> 00:14:53.110
extraction is a trap, we have to unpack both the
309
00:14:53.230 --> 00:14:57.210
staggering technical capabilities of modern AI and the incredibly strict
310
00:14:57.309 --> 00:14:59.090
legal framework of audio post-production.
311
00:14:59.740 --> 00:15:00.960
Okay, so let's start with the tech.
312
00:15:01.419 --> 00:15:06.149
Technically, your assumption is valid. Artificial intelligence has made quantum
313
00:15:06.190 --> 00:15:10.769
leaps in stem separation. The data from AI magics is wild.
314
00:15:11.029 --> 00:15:14.909
We have moved entirely beyond the messy, artifact-heavy vocal extractors
315
00:15:14.929 --> 00:15:15.460
of the early 2020s.
316
00:15:15.450 --> 00:15:19.070
The ones that sounded like they were recorded underwater? Exactly.
317
00:15:19.490 --> 00:15:23.669
Today, neural networks are trained on unimaginably massive data sets
318
00:15:23.850 --> 00:15:27.750
consisting of fully mixed songs paired with their original isolated
319
00:15:27.870 --> 00:15:28.830
studio multitracks.
320
00:15:29.309 --> 00:15:32.549
So the software isn't just applying basic equalization to cut
321
00:15:32.590 --> 00:15:36.789
out the middle frequencies. It actually understands what an instrument
322
00:15:36.870 --> 00:15:37.409
sounds like.
323
00:15:37.769 --> 00:15:41.690
Yes. It learns the distinct spectral and temporal patterns of audio.
324
00:15:42.230 --> 00:15:45.269
When the AI analyzes a complex mix, it recognizes the
325
00:15:45.370 --> 00:15:49.370
unique harmonic structures of a human voice vibrating or the
326
00:15:49.429 --> 00:15:51.950
sharp transient attack of a drumstick hitting a snare.
327
00:15:52.370 --> 00:15:58.129
That's incredible. Cloud-based neural networks like lolol.ai... can now cleanly
328
00:15:58.210 --> 00:16:01.710
separate a single stereo track into up to 10 pristine
329
00:16:01.750 --> 00:16:05.980
individual stems. And tools like iZotope RX12 can do it
330
00:16:06.039 --> 00:16:06.899
locally on a machine.
331
00:16:07.240 --> 00:16:10.059
And the Sync Research Report actually highlights a massive trend
332
00:16:10.100 --> 00:16:11.039
here with that tech, right?
333
00:16:11.059 --> 00:16:11.559
It does.
334
00:16:11.580 --> 00:16:16.159
Supervisors are actively using this high-fidelity AI audio unbaking on
335
00:16:16.320 --> 00:16:17.460
original Vintage Masters.
336
00:16:17.960 --> 00:16:21.659
Right, because they take authentic 1950s and 60s jazz and
337
00:16:21.700 --> 00:16:26.009
blues recordings that literally never possessed multi-track stems. They were
338
00:16:26.149 --> 00:16:27.129
recorded straight to tape.
339
00:16:27.460 --> 00:16:27.620
Yeah.
340
00:16:28.039 --> 00:16:30.940
And they use AI to isolate the upright bass or
341
00:16:30.980 --> 00:16:35.289
the horn section. This allows editors to place authentic historical
342
00:16:35.409 --> 00:16:39.610
instrumental beds underneath dense dialogue scenes without the original singer's
343
00:16:39.669 --> 00:16:41.549
vocals colliding with the actor's voices.
344
00:16:42.029 --> 00:16:43.750
So the technology definitely works.
345
00:16:44.389 --> 00:16:48.019
The supervisors are already doing it with vintage tracks. But
346
00:16:48.299 --> 00:16:52.080
audio post-production editors working on major films require a very
347
00:16:52.120 --> 00:16:55.279
specific layout, don't they? They don't just want a folder
348
00:16:55.320 --> 00:16:56.460
of 10 different instruments.
349
00:16:56.919 --> 00:17:00.480
No, they don't. They require a highly specific configuration known
350
00:17:00.519 --> 00:17:02.820
in the industry as an M & E split.
351
00:17:03.539 --> 00:17:04.400
Which stands for?
352
00:17:04.900 --> 00:17:07.660
Dialogue, music, and effects. When a film is prepared for
353
00:17:07.880 --> 00:17:10.559
foreign distribution or sent to a trailer house to cut
354
00:17:10.579 --> 00:17:13.599
a promo, the dialogue, the musical score, and the sound
355
00:17:13.619 --> 00:17:17.160
effects must be isolated into three completely independent lanes.
356
00:17:17.880 --> 00:17:18.640
Ah, okay.
357
00:17:18.880 --> 00:17:21.500
This allows a studio in Germany, for example, to seamlessly
358
00:17:21.539 --> 00:17:24.740
drop in a German voice actor without erasing the explosion
359
00:17:24.779 --> 00:17:28.450
sound effect or the underlying music. Specialized audio software like
360
00:17:28.549 --> 00:17:32.990
iZotope RX12 Advanced or Moises now offers tools specifically designed
361
00:17:33.009 --> 00:17:35.950
to force a stereo track into this mandatory three-way split.
362
00:17:36.569 --> 00:17:40.470
Okay. So the technology is miraculous, and the editors possess
363
00:17:40.490 --> 00:17:43.789
the exact software needed to generate these M & E splits. Yep.
364
00:17:44.349 --> 00:17:47.150
So I return to my challenge. Why can't they just
365
00:17:47.210 --> 00:17:49.750
run my indie track through the AI if I forget
366
00:17:49.789 --> 00:17:50.470
to send the stems?
367
00:17:51.410 --> 00:17:55.430
Ah, this raises an important question. The brutal reality of
368
00:17:55.690 --> 00:17:56.940
copyright law. Ah.
369
00:17:57.289 --> 00:17:57.690
The law.
370
00:17:58.190 --> 00:18:02.089
Industry resources for video editors like the Postflow are incredibly
371
00:18:02.130 --> 00:18:05.549
clear on this boundary. Creating an isolated vocal, generating a
372
00:18:05.730 --> 00:18:10.049
drumless version, or taking an AI-extracted stem and relaying it
373
00:18:10.150 --> 00:18:15.089
under a different visual sequence without explicit contractual permission creates
374
00:18:15.130 --> 00:18:17.089
an unauthorized derivative work.
375
00:18:17.579 --> 00:18:17.880
Oh, wow.
376
00:18:18.039 --> 00:18:23.000
Yeah. Under U.S. copyright law, specifically 17 U.S.C. Section 101,
377
00:18:23.380 --> 00:18:26.859
The right to adapt, transform, or recast a copyrighted work
378
00:18:27.019 --> 00:18:29.640
belongs exclusively to the copyright holder.
379
00:18:30.059 --> 00:18:33.240
That is a massive distinction. So even if the production
380
00:18:33.259 --> 00:18:35.480
company is paid for the sync license to use the
381
00:18:35.519 --> 00:18:38.680
song in the scene, that license doesn't grant them the
382
00:18:38.720 --> 00:18:41.960
legal right to fundamentally alter the structural DNA of the
383
00:18:42.000 --> 00:18:43.809
master recording by extracting the vocals.
384
00:18:44.029 --> 00:18:47.390
Nope. The mechanism of extraction is irrelevant to the law.
385
00:18:47.910 --> 00:18:50.880
Whether you built that instrumental version with a razor blade
386
00:18:51.000 --> 00:18:55.460
splicing magnetic tape in 1975 or a neural network generated
387
00:18:55.500 --> 00:18:58.759
it this morning, you are violating the rights holder's exclusive
388
00:18:58.799 --> 00:19:00.240
control over adaptations.
389
00:19:00.670 --> 00:19:04.569
So it's legally radioactive for the editor to do it themselves.
390
00:19:04.910 --> 00:19:08.789
Completely. If an editor tries to slip an unauthorized AI
391
00:19:08.849 --> 00:19:12.859
instrumental into a broadcast, the platform's content ID systems which
392
00:19:12.940 --> 00:19:16.599
constantly scan audio against registered databases, will flag the copyright
393
00:19:16.640 --> 00:19:17.900
violation immediately.
394
00:19:17.920 --> 00:19:18.779
Oh, I bet.
395
00:19:19.119 --> 00:19:22.539
Major stock music libraries explicitly ban the registration of edited
396
00:19:22.579 --> 00:19:25.279
or derivative versions of their tracks for this exact reason.
397
00:19:25.920 --> 00:19:29.740
Film editors simply cannot legally unbake a licensed indie track
398
00:19:29.799 --> 00:19:30.740
merely out of convenience.
399
00:19:31.200 --> 00:19:34.299
Because a separated stem of a licensed cue can never
400
00:19:34.380 --> 00:19:38.529
exist as a standalone digital file without violating the underlying license.
401
00:19:38.970 --> 00:19:39.410
Exactly.
402
00:19:39.869 --> 00:19:42.549
Which brings us perfectly back to the warning about messy folders.
403
00:19:43.210 --> 00:19:46.519
The film editor is legally prohibited from unbaking the track
404
00:19:46.599 --> 00:19:49.970
for you. They must have the official authorized stems provided.
405
00:19:50.109 --> 00:19:50.910
directly by you.
406
00:19:51.130 --> 00:19:51.349
Right.
407
00:19:51.809 --> 00:19:54.369
And if you don't have them cleanly organized and ready
408
00:19:54.410 --> 00:19:58.039
to deliver the moment the supervisor asks, the production moves
409
00:19:58.119 --> 00:19:58.970
on to someone who does.
410
00:19:59.630 --> 00:20:03.099
The Sync Report explicitly notes that supervisors will actually choose
411
00:20:03.119 --> 00:20:07.400
to license an authentic-sounding indie Retromaster over a major label
412
00:20:07.420 --> 00:20:11.019
catalog track. simply because the indie artist can immediately hand
413
00:20:11.079 --> 00:20:13.460
over full legal stem control for the dialogue mix.
414
00:20:14.069 --> 00:20:16.150
It perfectly illustrates the modern ecosystem.
415
00:20:16.650 --> 00:20:19.329
You have to provide the initial creative spark to get noticed,
416
00:20:19.450 --> 00:20:22.529
but you must reinforce it with a frictionless legal and
417
00:20:22.650 --> 00:20:24.930
technical workflow to actually secure the bag.
418
00:20:25.470 --> 00:20:27.960
So what does this all mean for you listening? It
419
00:20:27.980 --> 00:20:31.680
means success in the 2026 sync world requires mastering a
420
00:20:31.740 --> 00:20:33.420
truly fascinating duality.
421
00:20:33.660 --> 00:20:35.099
It's two completely different mindsets.
422
00:20:35.500 --> 00:20:39.079
It is. On the creative side, you need the uncompromising
423
00:20:39.170 --> 00:20:42.740
weirdness of a Phoebe Bridgers or a Debbie Harry. You
424
00:20:42.759 --> 00:20:46.200
have to write the habanero-infused tart that cuts through the noise.
425
00:20:46.640 --> 00:20:49.680
But the moment the creative pitch is accepted, you have
426
00:20:49.700 --> 00:20:50.839
to pivot instantly.
427
00:20:50.940 --> 00:20:52.960
Turn off the artist brain, turn on the admin brain.
428
00:20:53.259 --> 00:20:56.990
Exactly. You need the rigorous, highly organized professionalism of a
429
00:20:57.049 --> 00:21:00.529
corporate accountant when delivering your stems, ensuring every file is
430
00:21:00.589 --> 00:21:04.309
perfectly labeled and legally cleared. The barrier to entry in
431
00:21:04.349 --> 00:21:07.569
making music has never been lower. But the barrier to
432
00:21:07.660 --> 00:21:12.230
enduring success requires mastering both the art and the administration.
433
00:21:13.009 --> 00:21:14.750
And as we wrap up this deep dive, I want
434
00:21:14.769 --> 00:21:16.970
to leave you with one final thought to ponder regarding
435
00:21:16.990 --> 00:21:18.670
the future of this ecosystem.
436
00:21:18.970 --> 00:21:19.289
All right.
437
00:21:20.250 --> 00:21:24.450
As AI models inevitably continue to perfect stem separation, approaching
438
00:21:24.490 --> 00:21:29.359
a point where an AI-extracted vocal is literally mathematically indistinguishable
439
00:21:29.420 --> 00:21:33.200
from the original studio master, will our rigid legal frameworks
440
00:21:33.299 --> 00:21:34.660
eventually be forced to adapt?
441
00:21:34.799 --> 00:21:36.559
That's a great question. Right?
442
00:21:36.960 --> 00:21:39.680
Will we see a future where sync licenses evolve to
443
00:21:40.079 --> 00:21:44.079
automatically grant on-the-fly remixing rights to film editors simply because
444
00:21:44.119 --> 00:21:47.640
the technology is too seamless to ignore? Or will the
445
00:21:47.680 --> 00:21:52.440
strict historical definition of a derivative work force music supervisors
446
00:21:52.509 --> 00:21:56.250
to rely exclusively on legally cleared, human-delivered stems forever?
447
00:21:56.869 --> 00:21:59.609
It's a massive gray area just waiting to be tested
448
00:21:59.650 --> 00:22:00.049
in court.
449
00:22:00.549 --> 00:22:01.069
It really is.
450
00:22:01.950 --> 00:22:04.130
Think about that the next time you hear a perfectly
451
00:22:04.230 --> 00:22:07.980
placed track in a tense season finale. and wondered just
452
00:22:08.180 --> 00:22:11.599
how many technical, administrative, and legal hurdles had to be
453
00:22:11.680 --> 00:22:14.279
cleared to get that single vocal line to fade out
454
00:22:14.759 --> 00:22:18.400
right as the actor begins to speak. Thank you so
455
00:22:18.420 --> 00:22:21.019
much for joining us on this deep dive. Keep writing,
456
00:22:21.319 --> 00:22:24.609
keep organizing those folders, and keep learning. We'll catch you
457
00:22:24.619 --> 00:22:24.920
next time.
1
00:00:00.230 --> 00:00:02.270
You know, usually when we talk about a film or
2
00:00:02.310 --> 00:00:06.580
TV scene just perfectly matching a piece of music, there's
3
00:00:07.219 --> 00:00:10.619
this expectation of pure, effortless magic. Oh, totally.
4
00:00:10.660 --> 00:00:12.380
Like it just fell out of the sky. Right.
5
00:00:12.759 --> 00:00:14.859
Like you just take the visual, you drop the song
6
00:00:14.919 --> 00:00:17.850
on top and boom, you know, cinematic gold. Yeah.
7
00:00:17.969 --> 00:00:21.410
I mean, it feels completely intuitive to you, the viewer,
8
00:00:21.469 --> 00:00:24.230
sitting on your couch. The assumption is simply that a director,
9
00:00:24.870 --> 00:00:27.410
I don't know, heard a cool track on Spotify, clicked
10
00:00:27.429 --> 00:00:29.410
a button and just dropped it into the timeline.
11
00:00:30.050 --> 00:00:32.859
But then you actually step into the reality of audio
12
00:00:32.880 --> 00:00:37.539
post-production and sync licensing. And suddenly that effortless magic looks
13
00:00:37.740 --> 00:00:39.859
a lot more like a really high-stakes puzzle.
14
00:00:40.259 --> 00:00:41.979
Oh, absolutely. It's a massive puzzle.
15
00:00:42.520 --> 00:00:47.130
You're suddenly staring at this landscape of clashing audio frequencies,
16
00:00:47.530 --> 00:00:52.570
missing digital files, and legal tripwires that is genuinely murky.
17
00:00:53.049 --> 00:00:57.460
Yeah. The modern sync placement is basically the absolute definition
18
00:00:57.539 --> 00:01:01.859
of a creative and administrative minefield. Winning requires navigating both
19
00:01:01.920 --> 00:01:03.340
sides flawlessly.
20
00:01:03.640 --> 00:01:06.060
Which is exactly why our mission for this deep dive
21
00:01:06.519 --> 00:01:08.620
is to guide you through that exact minefield.
22
00:01:08.640 --> 00:01:09.439
Bring your hard hats.
23
00:01:10.060 --> 00:01:14.859
Seriously, we are unpacking the wild, rapidly evolving ecosystem of
24
00:01:14.939 --> 00:01:19.010
TV and film music syncs. And we're tailoring this specifically
25
00:01:19.049 --> 00:01:22.489
for all you indie publishers, songwriters, and producers trying to
26
00:01:22.530 --> 00:01:24.269
keep up with the 2026 landscape.
27
00:01:24.590 --> 00:01:26.290
It is a lot to keep up with right now.
28
00:01:26.590 --> 00:01:30.260
It really is. So we're synthesizing insights from across the
29
00:01:30.299 --> 00:01:33.780
industry today. We've got industry guides from CD Baby and
30
00:01:33.859 --> 00:01:37.959
Clint Music, technical manuals on AI stem extraction from AI
31
00:01:38.019 --> 00:01:39.239
Magix and the Postflow.
32
00:01:39.400 --> 00:01:40.980
Some really fascinating data in there, yeah.
33
00:01:41.280 --> 00:01:44.079
And fresh data from sound charts and Billboard from the
34
00:01:44.120 --> 00:01:48.489
latest Prestige TV finales. Okay, let's unpack this. Because before
35
00:01:48.530 --> 00:01:52.349
we even worry about contracts and technical asset delivery, we
36
00:01:52.370 --> 00:01:54.849
have to start with the creative DNA of a sync.
37
00:01:55.430 --> 00:01:56.769
Right, the actual music.
38
00:01:57.150 --> 00:02:00.260
Exactly. What actually makes a piece of music attractive to
39
00:02:00.280 --> 00:02:01.739
a music supervisor today?
40
00:02:02.079 --> 00:02:06.480
Well, that underlying philosophy has completely inverted over the last decade. Historically,
41
00:02:06.540 --> 00:02:09.840
there was this prevailing myth in the industry that sync
42
00:02:09.909 --> 00:02:12.770
music needed to be vague or universal.
43
00:02:13.360 --> 00:02:17.800
Right, like that standard mid-2000s indie folk or commercial pop
44
00:02:17.840 --> 00:02:19.300
playing over network TV dramas.
45
00:02:19.680 --> 00:02:23.900
Exactly. Those songs were literally engineered to broadly apply to
46
00:02:23.960 --> 00:02:26.659
anyone feeling a little bit sad or a little bit happy,
47
00:02:26.780 --> 00:02:27.020
you know?
48
00:02:27.259 --> 00:02:28.780
Just keeping it right down the middle.
49
00:02:29.680 --> 00:02:33.539
The goal was low-friction background music. But today, the sheer
50
00:02:33.580 --> 00:02:37.030
volume of streaming media has exposed viewers to highly nuanced,
51
00:02:37.159 --> 00:02:40.699
diverse storytelling— Directors don't want background music anymore. They want
52
00:02:40.759 --> 00:02:44.960
a co-star in the scene. Music supervisors actively seek out quirky,
53
00:02:45.099 --> 00:02:46.219
highly specific songs.
54
00:02:46.539 --> 00:02:49.909
It's essentially like trying to bake a generic vanilla cake
55
00:02:50.409 --> 00:02:52.469
that you think everyone in the room is going to tolerate.
56
00:02:53.110 --> 00:02:56.530
But the music supervisor is actually frantically searching the city for, like,
57
00:02:57.050 --> 00:02:59.090
a habanero-infused chocolate tart.
58
00:02:59.689 --> 00:03:00.210
Exactly.
59
00:03:00.849 --> 00:03:01.729
You have to be the tart.
60
00:03:02.229 --> 00:03:02.810
Be the tart.
61
00:03:03.330 --> 00:03:03.750
I love that.
62
00:03:04.069 --> 00:03:06.210
And if you look at the current guidance being handed
63
00:03:06.250 --> 00:03:09.780
down to DIY musicians from places like CD Baby, the
64
00:03:09.819 --> 00:03:11.020
warning is explicit.
65
00:03:11.500 --> 00:03:12.199
What are they telling them?
66
00:03:12.520 --> 00:03:14.759
They're saying do not round off your creative edges to
67
00:03:14.780 --> 00:03:16.280
be more syncable. Really?
68
00:03:17.099 --> 00:03:18.020
So keep the weirdness?
69
00:03:18.500 --> 00:03:23.120
Yes. Removing a strange time signature, an unusual instrument, or
70
00:03:23.159 --> 00:03:25.580
a highly specific lyric just to play it safe will
71
00:03:25.620 --> 00:03:28.319
likely remove the very element a music supervisor would have
72
00:03:28.379 --> 00:03:29.319
used to anchor a scene.
73
00:03:29.340 --> 00:03:30.770
Oh, wow.
74
00:03:31.050 --> 00:03:31.250
Yeah.
75
00:03:31.900 --> 00:03:34.159
And a brilliant example of this just played out in
76
00:03:34.199 --> 00:03:36.699
the season three finale of Netflix's Heartstopper.
77
00:03:36.919 --> 00:03:37.659
Oh, that was huge.
78
00:03:37.979 --> 00:03:41.099
It was. Looking at the latest weekly TV Music Sync
79
00:03:41.159 --> 00:03:44.300
research report from October 3rd, they closed out the season
80
00:03:44.319 --> 00:03:47.419
with this incredibly emotional needle drop using I Know the
81
00:03:47.520 --> 00:03:48.680
End by Phoebe Bridgers.
82
00:03:49.000 --> 00:03:50.479
Such a great track. It is.
83
00:03:50.879 --> 00:03:54.370
And this track from her 2020 album Punisher is the
84
00:03:54.560 --> 00:03:57.469
exact antithesis of a standard vague pop song.
85
00:03:57.629 --> 00:03:58.090
Completely.
86
00:03:58.110 --> 00:04:00.889
I mean, it's a nearly six minute chamber pop suite.
87
00:04:01.469 --> 00:04:05.349
It starts as this hushed, intimate, acoustic ballad about depression
88
00:04:05.469 --> 00:04:07.050
and driving up the California coast.
89
00:04:07.150 --> 00:04:08.969
And then it just shifts gears entirely.
90
00:04:08.990 --> 00:04:14.500
It literally detonates into a massive, metal-inspired, apocalyptic screaming match.
91
00:04:15.159 --> 00:04:18.420
We are talking chaotic horns, a massive choir, and a
92
00:04:18.500 --> 00:04:20.879
blistering lead guitar part by Nick Zinner of the Yeah
93
00:04:20.899 --> 00:04:23.459
Yeah Yes. It is wonderfully weird.
94
00:04:23.600 --> 00:04:24.240
It really is.
95
00:04:24.259 --> 00:04:27.420
I mean, Bridgers used to perform this on NPR's tiny
96
00:04:27.459 --> 00:04:30.769
desk from inside a fake Oval Office. And she smashed
97
00:04:30.829 --> 00:04:32.269
a monitor playing it on SNL.
98
00:04:32.569 --> 00:04:35.899
See, it's an uncompromising creative vision, which is exactly why
99
00:04:36.019 --> 00:04:37.959
it works so brilliantly on screen. Yeah.
100
00:04:38.259 --> 00:04:39.680
A generic track wouldn't have done that.
101
00:04:39.860 --> 00:04:43.660
Not at all. That screaming brass and guitar finale layered
102
00:04:43.720 --> 00:04:47.250
over a pivotal crossroad moment for the main characters in Heartstopper.
103
00:04:47.329 --> 00:04:50.209
It just created an overwhelming cinematic payoff.
104
00:04:50.269 --> 00:04:50.449
Right.
105
00:04:50.730 --> 00:04:54.410
A generic sweeping orchestral cue would have felt entirely unearned there.
106
00:04:54.879 --> 00:04:57.620
The specificity of the chaos in the track mirrors the
107
00:04:57.800 --> 00:05:00.180
internal emotional chaos of the characters.
108
00:05:00.600 --> 00:05:04.350
And we are seeing this demand for hyper-specific friction everywhere
109
00:05:04.379 --> 00:05:07.170
right now. Like, as we move into the fall streaming
110
00:05:07.209 --> 00:05:11.490
season for horror and psychological thrillers across platforms like Shudder
111
00:05:11.529 --> 00:05:12.009
and Hulu.
112
00:05:12.149 --> 00:05:14.029
Oh, the briefs for those are fascinating right now.
113
00:05:14.290 --> 00:05:19.379
Right. Music supervisors are actively issuing briefs for Gothic Americana
114
00:05:19.819 --> 00:05:21.439
and Minimalist Dark Folk.
115
00:05:22.100 --> 00:05:25.680
Yeah, they are actively hunting for high friction acoustic arrangements.
116
00:05:26.399 --> 00:05:31.980
They want like solitary acoustic guitars, bowed bass, haunting low
117
00:05:32.019 --> 00:05:36.290
register vocals, and really dark, non-romantic lyrics.
118
00:05:36.769 --> 00:05:40.990
Exactly. They want all those weird atmospheric imperfections left in
119
00:05:41.009 --> 00:05:43.829
the track to give the sound design room to breathe. Yeah.
120
00:05:44.250 --> 00:05:48.420
And what's fascinating here is that this desire for uncompromising,
121
00:05:48.660 --> 00:05:52.139
specific artistic vision isn't actually a new phenomenon.
122
00:05:52.319 --> 00:05:52.540
No.
123
00:05:52.839 --> 00:05:56.060
No, it's just becoming the mandatory standard. If we connect
124
00:05:56.079 --> 00:06:00.339
this to historical sync perfection, Blondie's 1980 hit Call Me
125
00:06:00.620 --> 00:06:02.040
serves as the ultimate blueprint.
126
00:06:02.379 --> 00:06:03.939
Oh, from American Gigolo.
127
00:06:04.199 --> 00:06:08.990
Yes. The legendary Italian disco producer Giorgio Marotta was scoring
128
00:06:09.029 --> 00:06:12.120
that neo-noir film. And he possessed an instrumental track called
129
00:06:12.379 --> 00:06:14.850
Man Machine that he initially pitched to Stevie Nicks.
130
00:06:15.189 --> 00:06:16.209
Oh, I didn't know that. Yeah.
131
00:06:16.750 --> 00:06:19.709
But contract conflicts prevented her from taking it. So he
132
00:06:19.750 --> 00:06:22.350
brings the instrumental to Debbie Harry. He asks her to
133
00:06:22.389 --> 00:06:24.930
write the lyrics. And rather than penning a vague pop
134
00:06:25.009 --> 00:06:28.240
anthem about love, she writes lyrics specifically tailored to the
135
00:06:28.279 --> 00:06:31.839
psychology of the film's main character, Richard Gere's male prostitute,
136
00:06:31.879 --> 00:06:32.500
Julian Kay.
137
00:06:32.879 --> 00:06:35.439
So she put herself completely inside the head of the
138
00:06:35.500 --> 00:06:37.199
character on screen. Exactly. Exactly.
139
00:06:37.720 --> 00:06:40.560
She pictured the opening visual, which is Julian driving his
140
00:06:40.620 --> 00:06:44.279
Mercedes convertible along the coast of California, and matched the
141
00:06:44.339 --> 00:06:46.160
psychological energy flawlessly.
142
00:06:46.360 --> 00:06:47.699
That's brilliant.
143
00:06:48.019 --> 00:06:50.310
The track is rooted in D minor, driven by a
144
00:06:50.350 --> 00:06:54.629
blistering 143 beats per minute dance rock tempo. It physically
145
00:06:54.750 --> 00:06:56.870
pushes the visual momentum of the car forward.
146
00:06:57.269 --> 00:07:00.029
And it paid off. I mean, it spent six consecutive
147
00:07:00.069 --> 00:07:01.790
weeks at number one on the Billboard Hot 100 in 1980. Right.
148
00:07:03.430 --> 00:07:06.790
Both Bridgers and Blondie prove that tailoring music to specific
149
00:07:06.850 --> 00:07:10.420
psychological states yields the most legendary cinematic payoffs.
150
00:07:11.569 --> 00:07:15.000
So let's follow that thread forward. Say you have captured
151
00:07:15.019 --> 00:07:19.100
that magic. You've embraced the weirdness. You've written the habanero tart.
152
00:07:19.759 --> 00:07:22.339
And the supervisor places it in a massive streaming show.
153
00:07:22.879 --> 00:07:24.300
You get the yes. Right.
154
00:07:24.519 --> 00:07:28.189
You get the yes. What actually happens in the aftermath? Obviously,
155
00:07:28.209 --> 00:07:31.050
you collect the upfront licensing fee, but What is the
156
00:07:31.310 --> 00:07:35.569
ongoing real-world value to the artist in a fragmented digital landscape?
157
00:07:36.089 --> 00:07:38.149
Well, we really have to examine the Shazam effect.
158
00:07:38.810 --> 00:07:40.110
Okay, the Shazam effect.
159
00:07:40.569 --> 00:07:44.009
Yeah, market data from places like sound charts consistently shows
160
00:07:44.050 --> 00:07:47.470
that Shazam is the ultimate early indicator of a cultural hit.
161
00:07:48.069 --> 00:07:50.339
Why is that? Just because people are curious?
162
00:07:50.680 --> 00:07:53.319
Because it captures active, high-intent discovery.
163
00:07:53.740 --> 00:07:54.019
Okay.
164
00:07:54.439 --> 00:07:56.500
When a viewer is watching a show and pulls out
165
00:07:56.540 --> 00:07:59.139
their phone to identify a track, they are crossing a
166
00:07:59.399 --> 00:08:00.839
massive behavioral threshold.
167
00:08:01.019 --> 00:08:01.959
Yeah, they're taking action.
168
00:08:02.160 --> 00:08:05.220
Right. They are no longer passively consuming background noise. They
169
00:08:05.240 --> 00:08:08.199
are demanding to know the identity of the artist. Spikes
170
00:08:08.259 --> 00:08:12.069
and Shazam counts almost universally precede streaming and radio growth
171
00:08:12.110 --> 00:08:13.230
by days or weeks.
172
00:08:13.750 --> 00:08:17.089
It's the purest signal that a track is rapidly gaining traction.
173
00:08:17.189 --> 00:08:17.670
Exactly.
174
00:08:17.990 --> 00:08:19.970
But let me push back on the scope of that
175
00:08:20.009 --> 00:08:23.069
for a second. Sure. If a music supervisor uses a
176
00:08:23.129 --> 00:08:27.240
song by an independent, completely unknown artist— A Shazam spike
177
00:08:27.379 --> 00:08:31.980
obviously translates to massive discovery. But supervisors frequently use tracks
178
00:08:32.039 --> 00:08:35.980
by massive, instantly recognizable legacy artists. Yeah, they do. Like,
179
00:08:36.139 --> 00:08:38.440
does a Shazam spike actually move the needle for a
180
00:08:38.480 --> 00:08:43.639
song everyone ostensibly already knows? Aren't discovery tools inherently for
181
00:08:43.679 --> 00:08:44.379
the undiscovered?
182
00:08:44.730 --> 00:08:47.570
It is a logical assumption, but it overlooks the mechanics
183
00:08:47.610 --> 00:08:49.149
of catalog revitalization.
184
00:08:49.169 --> 00:08:50.169
Okay, explain that.
185
00:08:50.490 --> 00:08:53.870
Syncs are the engine of back-end performance royalties. Every time
186
00:08:53.889 --> 00:08:57.059
that show airs in international markets... moves to a secondary
187
00:08:57.100 --> 00:09:00.360
streaming service, or is syndicated, it generates publishing royalties.
188
00:09:00.539 --> 00:09:01.279
Right, that makes sense.
189
00:09:01.539 --> 00:09:05.090
But more importantly, Async introduces legacy tracks to a younger
190
00:09:05.129 --> 00:09:08.029
demographic who might not possess any cultural memory of a
191
00:09:08.549 --> 00:09:08.850
1980s hit.
192
00:09:09.330 --> 00:09:10.710
Oh, that's true. Right? They weren't there.
193
00:09:10.970 --> 00:09:13.929
Right. So when those younger viewers Shazam the track, it
194
00:09:14.009 --> 00:09:18.269
feeds a cascading algorithm. The Shazam tags trigger an influx
195
00:09:18.340 --> 00:09:21.919
of Spotify searches, which pushes the song into algorithmic playlists
196
00:09:21.980 --> 00:09:25.830
like Discover Weekly, triggering a massive compounding effect across the
197
00:09:25.929 --> 00:09:27.409
entire streaming ecosystem.
198
00:09:27.830 --> 00:09:31.090
We actually have a flawless real-world example of this algorithmic
199
00:09:31.110 --> 00:09:34.450
cascade happening right now with HBO's The Penguin.
200
00:09:34.889 --> 00:09:36.740
Oh yeah, the needle drops on that show are wild.
201
00:09:37.379 --> 00:09:40.440
They really are. The music supervisors on that project, Jen
202
00:09:40.460 --> 00:09:44.059
Malone and Whitney Piltzer, are conducting an absolute masterclass in
203
00:09:44.139 --> 00:09:46.120
ironic high-contrast needle drops.
204
00:09:46.519 --> 00:09:51.019
Yeah, they are placing highly recognizable tracks underneath ink incredibly tense,
205
00:09:51.019 --> 00:09:52.480
gritty crime scenes.
206
00:09:52.480 --> 00:09:55.500
Like in episode one, Oz and Victor get into a
207
00:09:55.500 --> 00:09:58.509
car and Tolly Parton's 9 to 5 plays on the radio.
208
00:09:58.820 --> 00:09:59.169
So good.
209
00:09:59.500 --> 00:10:02.779
And in episode two, Bobby Darin's Call Me Irresponsible scores
210
00:10:02.799 --> 00:10:06.139
a brutal shootout. They are pulling in Gil Scott Heron's
211
00:10:06.179 --> 00:10:08.500
Me and the Devil and The Cures of Forest.
212
00:10:09.200 --> 00:10:13.960
By juxtaposing those evocative, familiar tracks against high-stakes mob violence,
213
00:10:14.480 --> 00:10:17.220
they instantly drive viewers to their phones to re-engage with
214
00:10:17.240 --> 00:10:17.580
the music.
215
00:10:18.139 --> 00:10:20.950
And the weekly sync data really illustrates the sheer power
216
00:10:20.970 --> 00:10:24.070
of that juxtaposition. Like in episode three of The Penguin,
217
00:10:24.450 --> 00:10:27.769
during a fast-paced vehicle chase sequence through Gotham's industrial district,
218
00:10:28.210 --> 00:10:29.769
they featured Blondie's Call Me.
219
00:10:30.169 --> 00:10:31.629
A track from 1980.
220
00:10:31.629 --> 00:10:35.250
Right, from 1980. And over the following seven days, Shazam
221
00:10:35.269 --> 00:10:37.889
recorded a top 10 global spike for the song.
222
00:10:38.409 --> 00:10:38.909
It's amazing.
223
00:10:39.269 --> 00:10:42.570
That isolated moment drove a massive wave of fresh streaming
224
00:10:42.610 --> 00:10:45.470
discovery among a younger demographic watching the crime saga.
225
00:10:45.919 --> 00:10:49.240
And the Heartstopper finale yielded an even more dramatic metric.
226
00:10:49.259 --> 00:10:50.220
Oh, yeah. What happened there?
227
00:10:50.860 --> 00:10:54.120
Following that sync, Phoebe Bridger's 2020 master of I Know
228
00:10:54.159 --> 00:10:59.259
the End experienced an immediate 410% surge in official U.S.
229
00:10:59.330 --> 00:11:00.610
Audio streams.
230
00:11:00.610 --> 00:11:00.690
410%.
231
00:11:00.690 --> 00:11:03.509
That's insane. That's insane. Yeah.
232
00:11:04.029 --> 00:11:06.529
Shooting straight to number one on the Billboard Top TV
233
00:11:06.570 --> 00:11:08.169
Songs chart for the week of October 3rd.
234
00:11:08.649 --> 00:11:12.289
Okay. Here's where it gets really interesting, though. Okay. You
235
00:11:12.309 --> 00:11:15.669
can write the brilliant, specific song. You can know exactly
236
00:11:15.730 --> 00:11:19.269
how massive that Shazam-driven streaming boost is going to be
237
00:11:19.309 --> 00:11:22.429
on the other side. But there is a heartbreaking reality
238
00:11:22.769 --> 00:11:24.480
hiding in the shadows of this business.
239
00:11:24.779 --> 00:11:27.100
Oh, boy. The administrative side.
240
00:11:27.299 --> 00:11:31.179
Exactly. You can win the creative lottery, get chosen by
241
00:11:31.220 --> 00:11:34.580
the music supervisor, and completely lose the deal on a
242
00:11:34.639 --> 00:11:39.649
purely administrative technicality. It's brutal. The biggest hurdle to actually
243
00:11:39.730 --> 00:11:42.929
getting your music on television isn't always writing a great hook.
244
00:11:43.750 --> 00:11:44.669
It's file management.
245
00:11:45.029 --> 00:11:48.350
It serves as the great unforgiving filter of the sync
246
00:11:48.389 --> 00:11:51.539
licensing world. Industry producers, like the folks at Clint Music,
247
00:11:51.919 --> 00:11:53.399
constantly warn about this bottleneck.
248
00:11:54.139 --> 00:11:56.860
Because supervisors are just too busy. Yes.
249
00:11:57.299 --> 00:12:01.710
Music supervisors are functioning under immense pressure. They are simultaneously
250
00:12:01.730 --> 00:12:06.129
juggling hundreds of audio cues, managing tight post-production deadlines, and
251
00:12:06.169 --> 00:12:09.509
appeasing demanding directors. So when a producer finally gets a
252
00:12:09.570 --> 00:12:13.389
yes on a creative pitch, the supervisor immediately requests the stems.
253
00:12:13.929 --> 00:12:16.350
And for those who haven't had to deliver a commercial project,
254
00:12:16.710 --> 00:12:19.929
stems are the individual isolated audio tracks that make up
255
00:12:19.960 --> 00:12:23.559
the final song. Right. So the isolated lead vocal, the
256
00:12:23.720 --> 00:12:27.899
isolated drum kit, the isolated bass guitar. The video editor
257
00:12:27.960 --> 00:12:31.320
needs those individual pieces so they can dip the volume
258
00:12:31.360 --> 00:12:33.929
of the vocals when the actors are speaking or perhaps
259
00:12:34.070 --> 00:12:37.049
use just the drum beat to build tension before cutting
260
00:12:37.070 --> 00:12:38.230
to commercial. Exactly.
261
00:12:38.570 --> 00:12:41.409
And the breakdown happens when a producer sends over a
262
00:12:41.549 --> 00:12:42.870
chaotic digital folder.
263
00:12:43.370 --> 00:12:45.220
Like what? What are they doing wrong?
264
00:12:45.460 --> 00:12:48.299
Well, the files might possess the wrong sample rates or
265
00:12:48.360 --> 00:12:51.639
bit depths, which is basically the audio equivalent of submitting
266
00:12:52.080 --> 00:12:56.080
a low-resolution, heavily pixelated photograph to a high-end magazine.
267
00:12:56.100 --> 00:12:56.799
Ooh, yikes.
268
00:12:57.139 --> 00:13:00.450
Or the files are lazily named things like, Final Drums V3,
269
00:13:00.870 --> 00:13:01.450
use this one.
270
00:13:01.889 --> 00:13:03.429
We've all named a file that way. We have.
271
00:13:04.120 --> 00:13:07.220
But if a music supervisor opens a folder and encounters
272
00:13:07.539 --> 00:13:11.500
an unorganized disaster, they do not have the time to
273
00:13:11.600 --> 00:13:12.500
act as your tech support.
274
00:13:12.990 --> 00:13:14.879
They won't just email you back and ask for the
275
00:13:14.919 --> 00:13:16.019
right ones? Nope.
276
00:13:16.970 --> 00:13:19.549
They will simply close the folder, delete the email, and
277
00:13:19.590 --> 00:13:21.529
move to the next producer on their list who has
278
00:13:21.570 --> 00:13:22.269
their act together.
279
00:13:22.570 --> 00:13:24.370
That is a brutal reality check.
280
00:13:24.850 --> 00:13:28.090
You pour your soul into the creative process, you secure
281
00:13:28.129 --> 00:13:31.090
the placement of a lifetime, and you lose the payout
282
00:13:31.149 --> 00:13:35.029
because your folder hierarchy was a mess. Professionalism in asset
283
00:13:35.070 --> 00:13:38.049
delivery carries the exact same weight as the songwriting itself.
284
00:13:38.509 --> 00:13:41.730
The music supervisors who consistently return to an independent publisher
285
00:13:41.870 --> 00:13:44.850
are doing so because that publisher is utterly frictionless to
286
00:13:44.889 --> 00:13:45.230
work with.
287
00:13:45.690 --> 00:13:46.570
That makes a lot of sense.
288
00:13:47.149 --> 00:13:51.830
If your stems are pristine, intuitively organized, and perfectly formatted,
289
00:13:52.350 --> 00:13:54.950
you transition from being just another artist to being a
290
00:13:55.190 --> 00:13:57.250
reliable problem solver for their production.
291
00:13:57.269 --> 00:14:01.990
Okay, I understand the demand for perfect organization. But I
292
00:14:02.029 --> 00:14:04.440
want to challenge this workflow for a second because we
293
00:14:04.539 --> 00:14:08.279
are living in 2026. Okay, hit me. Let's say I'm
294
00:14:08.320 --> 00:14:14.139
an indie publisher representing a brilliant, authentic 1960s analog soul recording. Sure.
295
00:14:14.559 --> 00:14:17.750
Or maybe I'm a modern bedroom producer who simply lost
296
00:14:17.759 --> 00:14:19.970
the original hard drive for a track I produced five
297
00:14:20.009 --> 00:14:23.389
years ago. I physically do not have the multi-track stems.
298
00:14:23.529 --> 00:14:24.009
It happens.
299
00:14:25.090 --> 00:14:28.250
Can't the film's video editor just run the stereo file
300
00:14:28.269 --> 00:14:31.019
through an AI tool and rip the vocals out? Like,
301
00:14:31.210 --> 00:14:32.870
separating a mix song used to be like trying to
302
00:14:33.009 --> 00:14:35.409
unstir a cup of coffee or unbake a cake, right?
303
00:14:35.669 --> 00:14:37.110
Yeah, that was the old analogy.
304
00:14:37.470 --> 00:14:41.690
But hasn't artificial intelligence completely solved the unbaking problem by now?
305
00:14:42.389 --> 00:14:44.490
Why am I losing the deal if the editor can
306
00:14:44.509 --> 00:14:47.129
just click a button and generate the stems themselves?
307
00:14:47.429 --> 00:14:50.090
To understand why relying on the editor to do the
308
00:14:50.149 --> 00:14:53.110
extraction is a trap, we have to unpack both the
309
00:14:53.230 --> 00:14:57.210
staggering technical capabilities of modern AI and the incredibly strict
310
00:14:57.309 --> 00:14:59.090
legal framework of audio post-production.
311
00:14:59.740 --> 00:15:00.960
Okay, so let's start with the tech.
312
00:15:01.419 --> 00:15:06.149
Technically, your assumption is valid. Artificial intelligence has made quantum
313
00:15:06.190 --> 00:15:10.769
leaps in stem separation. The data from AI magics is wild.
314
00:15:11.029 --> 00:15:14.909
We have moved entirely beyond the messy, artifact-heavy vocal extractors
315
00:15:14.929 --> 00:15:15.460
of the early 2020s.
316
00:15:15.450 --> 00:15:19.070
The ones that sounded like they were recorded underwater? Exactly.
317
00:15:19.490 --> 00:15:23.669
Today, neural networks are trained on unimaginably massive data sets
318
00:15:23.850 --> 00:15:27.750
consisting of fully mixed songs paired with their original isolated
319
00:15:27.870 --> 00:15:28.830
studio multitracks.
320
00:15:29.309 --> 00:15:32.549
So the software isn't just applying basic equalization to cut
321
00:15:32.590 --> 00:15:36.789
out the middle frequencies. It actually understands what an instrument
322
00:15:36.870 --> 00:15:37.409
sounds like.
323
00:15:37.769 --> 00:15:41.690
Yes. It learns the distinct spectral and temporal patterns of audio.
324
00:15:42.230 --> 00:15:45.269
When the AI analyzes a complex mix, it recognizes the
325
00:15:45.370 --> 00:15:49.370
unique harmonic structures of a human voice vibrating or the
326
00:15:49.429 --> 00:15:51.950
sharp transient attack of a drumstick hitting a snare.
327
00:15:52.370 --> 00:15:58.129
That's incredible. Cloud-based neural networks like lolol.ai... can now cleanly
328
00:15:58.210 --> 00:16:01.710
separate a single stereo track into up to 10 pristine
329
00:16:01.750 --> 00:16:05.980
individual stems. And tools like iZotope RX12 can do it
330
00:16:06.039 --> 00:16:06.899
locally on a machine.
331
00:16:07.240 --> 00:16:10.059
And the Sync Research Report actually highlights a massive trend
332
00:16:10.100 --> 00:16:11.039
here with that tech, right?
333
00:16:11.059 --> 00:16:11.559
It does.
334
00:16:11.580 --> 00:16:16.159
Supervisors are actively using this high-fidelity AI audio unbaking on
335
00:16:16.320 --> 00:16:17.460
original Vintage Masters.
336
00:16:17.960 --> 00:16:21.659
Right, because they take authentic 1950s and 60s jazz and
337
00:16:21.700 --> 00:16:26.009
blues recordings that literally never possessed multi-track stems. They were
338
00:16:26.149 --> 00:16:27.129
recorded straight to tape.
339
00:16:27.460 --> 00:16:27.620
Yeah.
340
00:16:28.039 --> 00:16:30.940
And they use AI to isolate the upright bass or
341
00:16:30.980 --> 00:16:35.289
the horn section. This allows editors to place authentic historical
342
00:16:35.409 --> 00:16:39.610
instrumental beds underneath dense dialogue scenes without the original singer's
343
00:16:39.669 --> 00:16:41.549
vocals colliding with the actor's voices.
344
00:16:42.029 --> 00:16:43.750
So the technology definitely works.
345
00:16:44.389 --> 00:16:48.019
The supervisors are already doing it with vintage tracks. But
346
00:16:48.299 --> 00:16:52.080
audio post-production editors working on major films require a very
347
00:16:52.120 --> 00:16:55.279
specific layout, don't they? They don't just want a folder
348
00:16:55.320 --> 00:16:56.460
of 10 different instruments.
349
00:16:56.919 --> 00:17:00.480
No, they don't. They require a highly specific configuration known
350
00:17:00.519 --> 00:17:02.820
in the industry as an M & E split.
351
00:17:03.539 --> 00:17:04.400
Which stands for?
352
00:17:04.900 --> 00:17:07.660
Dialogue, music, and effects. When a film is prepared for
353
00:17:07.880 --> 00:17:10.559
foreign distribution or sent to a trailer house to cut
354
00:17:10.579 --> 00:17:13.599
a promo, the dialogue, the musical score, and the sound
355
00:17:13.619 --> 00:17:17.160
effects must be isolated into three completely independent lanes.
356
00:17:17.880 --> 00:17:18.640
Ah, okay.
357
00:17:18.880 --> 00:17:21.500
This allows a studio in Germany, for example, to seamlessly
358
00:17:21.539 --> 00:17:24.740
drop in a German voice actor without erasing the explosion
359
00:17:24.779 --> 00:17:28.450
sound effect or the underlying music. Specialized audio software like
360
00:17:28.549 --> 00:17:32.990
iZotope RX12 Advanced or Moises now offers tools specifically designed
361
00:17:33.009 --> 00:17:35.950
to force a stereo track into this mandatory three-way split.
362
00:17:36.569 --> 00:17:40.470
Okay. So the technology is miraculous, and the editors possess
363
00:17:40.490 --> 00:17:43.789
the exact software needed to generate these M & E splits. Yep.
364
00:17:44.349 --> 00:17:47.150
So I return to my challenge. Why can't they just
365
00:17:47.210 --> 00:17:49.750
run my indie track through the AI if I forget
366
00:17:49.789 --> 00:17:50.470
to send the stems?
367
00:17:51.410 --> 00:17:55.430
Ah, this raises an important question. The brutal reality of
368
00:17:55.690 --> 00:17:56.940
copyright law. Ah.
369
00:17:57.289 --> 00:17:57.690
The law.
370
00:17:58.190 --> 00:18:02.089
Industry resources for video editors like the Postflow are incredibly
371
00:18:02.130 --> 00:18:05.549
clear on this boundary. Creating an isolated vocal, generating a
372
00:18:05.730 --> 00:18:10.049
drumless version, or taking an AI-extracted stem and relaying it
373
00:18:10.150 --> 00:18:15.089
under a different visual sequence without explicit contractual permission creates
374
00:18:15.130 --> 00:18:17.089
an unauthorized derivative work.
375
00:18:17.579 --> 00:18:17.880
Oh, wow.
376
00:18:18.039 --> 00:18:23.000
Yeah. Under U.S. copyright law, specifically 17 U.S.C. Section 101,
377
00:18:23.380 --> 00:18:26.859
The right to adapt, transform, or recast a copyrighted work
378
00:18:27.019 --> 00:18:29.640
belongs exclusively to the copyright holder.
379
00:18:30.059 --> 00:18:33.240
That is a massive distinction. So even if the production
380
00:18:33.259 --> 00:18:35.480
company is paid for the sync license to use the
381
00:18:35.519 --> 00:18:38.680
song in the scene, that license doesn't grant them the
382
00:18:38.720 --> 00:18:41.960
legal right to fundamentally alter the structural DNA of the
383
00:18:42.000 --> 00:18:43.809
master recording by extracting the vocals.
384
00:18:44.029 --> 00:18:47.390
Nope. The mechanism of extraction is irrelevant to the law.
385
00:18:47.910 --> 00:18:50.880
Whether you built that instrumental version with a razor blade
386
00:18:51.000 --> 00:18:55.460
splicing magnetic tape in 1975 or a neural network generated
387
00:18:55.500 --> 00:18:58.759
it this morning, you are violating the rights holder's exclusive
388
00:18:58.799 --> 00:19:00.240
control over adaptations.
389
00:19:00.670 --> 00:19:04.569
So it's legally radioactive for the editor to do it themselves.
390
00:19:04.910 --> 00:19:08.789
Completely. If an editor tries to slip an unauthorized AI
391
00:19:08.849 --> 00:19:12.859
instrumental into a broadcast, the platform's content ID systems which
392
00:19:12.940 --> 00:19:16.599
constantly scan audio against registered databases, will flag the copyright
393
00:19:16.640 --> 00:19:17.900
violation immediately.
394
00:19:17.920 --> 00:19:18.779
Oh, I bet.
395
00:19:19.119 --> 00:19:22.539
Major stock music libraries explicitly ban the registration of edited
396
00:19:22.579 --> 00:19:25.279
or derivative versions of their tracks for this exact reason.
397
00:19:25.920 --> 00:19:29.740
Film editors simply cannot legally unbake a licensed indie track
398
00:19:29.799 --> 00:19:30.740
merely out of convenience.
399
00:19:31.200 --> 00:19:34.299
Because a separated stem of a licensed cue can never
400
00:19:34.380 --> 00:19:38.529
exist as a standalone digital file without violating the underlying license.
401
00:19:38.970 --> 00:19:39.410
Exactly.
402
00:19:39.869 --> 00:19:42.549
Which brings us perfectly back to the warning about messy folders.
403
00:19:43.210 --> 00:19:46.519
The film editor is legally prohibited from unbaking the track
404
00:19:46.599 --> 00:19:49.970
for you. They must have the official authorized stems provided.
405
00:19:50.109 --> 00:19:50.910
directly by you.
406
00:19:51.130 --> 00:19:51.349
Right.
407
00:19:51.809 --> 00:19:54.369
And if you don't have them cleanly organized and ready
408
00:19:54.410 --> 00:19:58.039
to deliver the moment the supervisor asks, the production moves
409
00:19:58.119 --> 00:19:58.970
on to someone who does.
410
00:19:59.630 --> 00:20:03.099
The Sync Report explicitly notes that supervisors will actually choose
411
00:20:03.119 --> 00:20:07.400
to license an authentic-sounding indie Retromaster over a major label
412
00:20:07.420 --> 00:20:11.019
catalog track. simply because the indie artist can immediately hand
413
00:20:11.079 --> 00:20:13.460
over full legal stem control for the dialogue mix.
414
00:20:14.069 --> 00:20:16.150
It perfectly illustrates the modern ecosystem.
415
00:20:16.650 --> 00:20:19.329
You have to provide the initial creative spark to get noticed,
416
00:20:19.450 --> 00:20:22.529
but you must reinforce it with a frictionless legal and
417
00:20:22.650 --> 00:20:24.930
technical workflow to actually secure the bag.
418
00:20:25.470 --> 00:20:27.960
So what does this all mean for you listening? It
419
00:20:27.980 --> 00:20:31.680
means success in the 2026 sync world requires mastering a
420
00:20:31.740 --> 00:20:33.420
truly fascinating duality.
421
00:20:33.660 --> 00:20:35.099
It's two completely different mindsets.
422
00:20:35.500 --> 00:20:39.079
It is. On the creative side, you need the uncompromising
423
00:20:39.170 --> 00:20:42.740
weirdness of a Phoebe Bridgers or a Debbie Harry. You
424
00:20:42.759 --> 00:20:46.200
have to write the habanero-infused tart that cuts through the noise.
425
00:20:46.640 --> 00:20:49.680
But the moment the creative pitch is accepted, you have
426
00:20:49.700 --> 00:20:50.839
to pivot instantly.
427
00:20:50.940 --> 00:20:52.960
Turn off the artist brain, turn on the admin brain.
428
00:20:53.259 --> 00:20:56.990
Exactly. You need the rigorous, highly organized professionalism of a
429
00:20:57.049 --> 00:21:00.529
corporate accountant when delivering your stems, ensuring every file is
430
00:21:00.589 --> 00:21:04.309
perfectly labeled and legally cleared. The barrier to entry in
431
00:21:04.349 --> 00:21:07.569
making music has never been lower. But the barrier to
432
00:21:07.660 --> 00:21:12.230
enduring success requires mastering both the art and the administration.
433
00:21:13.009 --> 00:21:14.750
And as we wrap up this deep dive, I want
434
00:21:14.769 --> 00:21:16.970
to leave you with one final thought to ponder regarding
435
00:21:16.990 --> 00:21:18.670
the future of this ecosystem.
436
00:21:18.970 --> 00:21:19.289
All right.
437
00:21:20.250 --> 00:21:24.450
As AI models inevitably continue to perfect stem separation, approaching
438
00:21:24.490 --> 00:21:29.359
a point where an AI-extracted vocal is literally mathematically indistinguishable
439
00:21:29.420 --> 00:21:33.200
from the original studio master, will our rigid legal frameworks
440
00:21:33.299 --> 00:21:34.660
eventually be forced to adapt?
441
00:21:34.799 --> 00:21:36.559
That's a great question. Right?
442
00:21:36.960 --> 00:21:39.680
Will we see a future where sync licenses evolve to
443
00:21:40.079 --> 00:21:44.079
automatically grant on-the-fly remixing rights to film editors simply because
444
00:21:44.119 --> 00:21:47.640
the technology is too seamless to ignore? Or will the
445
00:21:47.680 --> 00:21:52.440
strict historical definition of a derivative work force music supervisors
446
00:21:52.509 --> 00:21:56.250
to rely exclusively on legally cleared, human-delivered stems forever?
447
00:21:56.869 --> 00:21:59.609
It's a massive gray area just waiting to be tested
448
00:21:59.650 --> 00:22:00.049
in court.
449
00:22:00.549 --> 00:22:01.069
It really is.
450
00:22:01.950 --> 00:22:04.130
Think about that the next time you hear a perfectly
451
00:22:04.230 --> 00:22:07.980
placed track in a tense season finale. and wondered just
452
00:22:08.180 --> 00:22:11.599
how many technical, administrative, and legal hurdles had to be
453
00:22:11.680 --> 00:22:14.279
cleared to get that single vocal line to fade out
454
00:22:14.759 --> 00:22:18.400
right as the actor begins to speak. Thank you so
455
00:22:18.420 --> 00:22:21.019
much for joining us on this deep dive. Keep writing,
456
00:22:21.319 --> 00:22:24.609
keep organizing those folders, and keep learning. We'll catch you
457
00:22:24.619 --> 00:22:24.920
next time.
