Tokyo Speirs is a multi-Platinum, Juno Award-winning producer, audio engineer and songwriter in pop rock band / YouTube sensation Walk Off the Earth.
SHARE
Hey Reader!
Now that the turkey coma has passed* , we're gonna jump right in to last week's strangest music industry news:
๐ธ FOMO INDEX๐ค
The music business, plotted on importance and vibes.
โ IMPORTANT โ
โ CRASH & BURN
๐ค
Bot farmer gets 18 months for an AI-song stream scam: 10,000 bot accounts siphoning royalties. His most loyal fans were 10,000 bots. The bots have been sentenced to buffering. Source
๐งฎ
Songwriters want a flat $0.0030 mechanical rate per stream in 2028. Songwriters brought a calculator to a knife fight. Daniel Ek will (probably) retaliate with a drone strike. Source
TO THE MOON โ
๐ฆ๐บ
Australia tells AI giants: no free ride on copyright. The AI companies said licensing was too complex. Australia replied: skill issue, pay up. Source
๐๏ธ
Spotify renews Joe Rogan in a multi-year deal, reportedly worth a cool $250M earnout. Spotify found the podcast couch money. Musicians found the couch. Source
โ๏ธ
Olivia Dean got sued over a chorus the comments section flagged first. YouTube commenters have been promoted to expert witnesses. Source
๐ค
Rihanna has at least 170 unreleased songs and zero release dates. Seventeen vaults of music. I liked her more in Battleship. Source
๐ฉธ
Slayer is selling blood-filled vinyl, because subtlety is dead. A liquid-filled zoetrope Reign in Blood for its 40th. Finally, a record that bleeds before your stylus does. Source
๐ฟ
Ottawa band Viscera sold 500 CDs in two weeks; US CD sales are up 16%. Your Discman is no longer a cry for help. It's an asset class. Source
โ WHO CARES โ
What I'm working on this week:
I've been (secretly) building an insane new compressor plugin.
You: "ANOTHER COMPRESSOR?"
Yes. Flattenstein:
a stitched-together mix of multiple different compressor types you can combine in weird, colorful ways.
VCA topology into Opto ballistics, feed-forward dual-stage release modes with variable auto-release tail memory
LUFS-based loudness matching with a two-stage clipper limiter out-stage.
A vintage pressure gauge and a backyard hose handle.
variable source-dependency for attack, release, and recover stages
I'm out here just trying stuff.
if you think this is overwhelming, it is.
Sponsored Segment
Two years ago I decided to model a full analog signal path: preamp, console, tape, summing, component by component. Real circuit modelling. Math, not vibes.
Everyone told me to just train a neural net on the hardware and call it a day. It feels like that's what everyone is doing now: feed the box some sine sweeps, ship the model, write "AI-powered" on the website, and call it a day.
My little ol' team of 1 did it the long way instead. Hand-built in collaboration with DSP pros Audioloom. Just me, some Blender tutorials and an unreasonable amount of time obsessing over my laptop. On trains, in hotels, anywhere I was able to steal even just five minutes.
I did NOT do any soldering. this picture is a joke.
Why? Because smarter people are already doing the neural-net stuff and I already knew exactly the behavior I was looking for. I was committing 5 different plugins directly to my source takes and instruments before editing or mixing and I wanted to streamline that step without losing my favorite aspects about analog recording: the way my favourite preamps lean into my preferred consoles, the way my favorite tape emulations compress when you hit em hard, the way the whole path interacts instead of five plugins fighting for DSP and my own indecison.
It's called RoundTrip. It's $99. There's a 14-day demo and I'd genuinely rather you demo it than take my word for any of this.