Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
SnappyCam was based-upon an ARM NEON implementation, novel for its day.

It captured 8 Mpx photos @ 20-30 fps, novel for its day.

Apple's Burst mode is strictly a CPU-based implementation, which does NOT use NEON.
The CPU in iPhones at the time was the Apple designed A7. The A7 provided performance improvements for custom NEON implementations. So, the ARM NEON could run on Apple’s CPU and Apple’s Burst mode runs on the CPU.

So, your thinking is they hired this guy, specifically known for his image handling prowess, and didn’t patent or use any of his inventions in the time he was at Apple.
 
I have friends who are musicians. REAL musicians. I have a relative that is a professional composer for TV, film, video games and more. This kind of s*** will eventually put them out of work. Someday the music in films, the radio, and TV shows will suck, because some suit somewhere will have decided that $10 for an AI soundtrack is cheaper than a real person with talent.

F*** this.
We will adapt ?‍♂️
 
I didn’t read that at all. I read it as they had an archive of already made songs that it selected from. So it takes music that no one has heard before and then blends it together into an incompressible mess?
Right, although hopefully with some degree of success to where it’s not incomprehensible :) Here’s another article that focuses more on the royalty free dynamic part of the report.

Again, it’s nothing that’s particularly new audibly. Anyone who’s seen a movie, opera, or other media has experienced at one time or another the music shifting subtly (or not so subtly) from one feel to another. The “new” part is is that the AI is composing as it goes and is able to shift the tempo/instruments/play style based on information provided by the user (either in an application use situation or as feedback to physical telemetry).
 
The CPU in iPhones at the time was the Apple designed A7. The A7 provided performance improvements for custom NEON implementations. So, the ARM NEON could run on Apple’s CPU and Apple’s Burst mode runs on the CPU.

So, your thinking is they hired this guy, specifically known for his image handling prowess, and didn’t patent or use any of his inventions in the time he was at Apple.

Companies acquire other companies ALL the time simply to keep their Tech out of the hands of their competition !

Also, your logic is wrong WRT ARM NEON; it's simply the vector processor of the CPU, NO more, NO less.

Any software program can use the vector processor if they want, & many Image Processing tasks do.

I suspect Apple decided NOT to use SnappyCam's implementation simply because of the support headaches that could result if/when the guy left !

A good part of his implementation was written in NEON Assembly Code !

Who wants to take over responsibility for someone elses' NEON Assembly Code ?
 
  • Like
Reactions: mdriftmeyer
This is interesting, so it’s almost coded algorithm based, depending on your specifics and overall what you’re participating in, it will generate music based on that type of activity.

I think the only way this would really be successful though, is how accurate the algorithm is and does it actually match specific music that the user would be interested in.
 
Listening to real music is so passé. Anyway, until Apple can generate dynamic podcasts, I'm sticking with spotify. Better podcasting than rogan and better music than young. AI is the future!
 
Music AI be like

1. Generate tune
2. Check against Content ID
3. Oh no tune already copyrighted, go back to 1.

BLUE SCREEN OF RECURSIVE LOOP DEATH
 
Last edited:
  • Haha
Reactions: VulchR
I'm having flashbacks to the hype around "Collaborative Filtering" back in the 90s. Or worse, Clippy the Paper Clip from MS Office (I'm running to catch a plane. :apple:Music MyCurator buzzes to say, "It looks like you're exercising, should I turn on a :apple:Curated Work Out playlist?"). Ugh.
Most recommendation systems in e-commerce or video streaming home pages are a variation of "collaborative filtering" btw.
 
Is this practical or more of a gimmick???
Considering Apple probably paid a lot of money, it’s just too easy to label things a ‘gimmick’ on the Internet from random members who have no experience with this yet. So why don’t we just give this a chance to see what the potential is, am I right?

But if I had to make a guesstimate, it really depends on how accurate the algorithm would be based on the users preference of music/what activity they are doing.
 
Oh my goodness. I guess in this age of AI/ML everything, this should be worth a buck or two for $AAPL, likely to be written off as bad investment in a couple of years. The world is going crazy. It's like we've solved all the hard problems so now we need this.

We need someone to show us how AI/ML can deliver a v2.0 of this.
 
  • Like
Reactions: turbineseaplane
Companies acquire other companies ALL the time simply to keep their Tech out of the hands of their competition !
This was a one man company, so “acquiring his company” was “hiring him.” :) And the one way to keep their tech out of the hands of their competition is to patent any methods they acquire. But, you’re saying they wouldn’t have used his tech for anything, even to patent his methods? I guess it could be that he made a really mean batch of TexMex chili that made him the darling of any potlucks. (Actually according to his website, he joined the Apple Camera Software team in Cupertino. So, he thought he was hired at Apple to work on camera software, even if others think he wasn’t)

Also, your logic is wrong WRT ARM NEON; it's simply the vector processor of the CPU, NO more, NO less.
The vector processor…………… of the CPU. No, I think I understand it fine. :) Unless you’re saying the A7 wasn’t a CPU.

I suspect Apple decided NOT to use SnappyCam's implementation simply because of the support headaches that could result if/when the guy left !
Apparently they didn’t use it because by the time they hired him, the A7 was already capable of the same thing.
 
This is interesting, so it’s almost coded algorithm based, depending on your specifics and overall what you’re participating in, it will generate music based on that type of activity.

I think the only way this would really be successful though, is how accurate the algorithm is and does it actually match specific music that the user would be interested in.
It could, but as individuals may want different types of music for different activities, they may let the user make the connection and it plays what you want how you want it based on your decisions.
 
I may be wrong, but this makes it seem like it could share a persistent connection to your “current mood” with advertisers, so ad Companies could theoretically select specific ads for you, based on your different personalized “current moods.”

that seems like it’s worth much more than some fancy soundtrack maker.

maybe, if you’re depressed it will automatically show you cute cat videos and a helpline number to call.

Either that, or it will just play “rage against the machine“ after a rough day at work. ?‍♂️

did I mention that I may be wrong?
 
Waste of money they already have AI driven playlists.

AND nothing beats a real DJ to listen to and play your requests. I use to Be one. Would still be DJing if not for COVID
 
I may be wrong, but this makes it seem like it could share a persistent connection to your “current mood” with advertisers, so ad Companies could theoretically select specific ads for you, based on your different personalized “current moods.”

that seems like it’s worth much more than some fancy soundtrack maker.

maybe, if you’re depressed it will automatically show you cute cat videos and a helpline number to call.

Either that, or it will just play “rage against the machine“ after a rough day at work. ?‍♂️

did I mention that I may be wrong?

Are you aware of how invasively dystopian that all sounds?

"Share my current mood with advertisers?"

Are you serious?
Not in a thousand years -- and not even then

How disgustingly invasive do we need to let capitalist ideology get?
I don't want "advertisers" in my every waking thought, moment or experience..

Good lord people.
Must all of human existence be about turning a money making wheel?

What a sad sad reality we are inhabiting if that's all it's about anymore.
 
Waste of money they already have AI driven playlists.

AND nothing beats a real DJ to listen to and play your requests. I use to Be one. Would still be DJing if not for COVID
true of course. but it is a bit different thing i guess
 
Are you aware of how invasively dystopian that all sounds?

"Share my current mood with advertisers?"

Are you serious?
Not in a thousand years -- and not even then

How disgustingly invasive do we need to let capitalist ideology get?
I don't want "advertisers" in my every waking thought, moment or experience..

Good lord people.
Must all of human existence be about turning a money making wheel?

What a sad sad reality we are inhabiting if that's all it's about anymore.

It’s just a thought based on this line from the article:
AI Music described it as a feature that could allow advertisers to create audio that would match user context, like mood.
And the fact that Ad Companies would love to know our personal “current moods.”

This may sound dystopian, but it kind of is a next logical step given current technology/advertising trends.

*(There is nothing wrong with Capitalism itself. Your issue is with greed, which exists in all social economic systems.)
 
Are you aware of how invasively dystopian that all sounds?

"Share my current mood with advertisers?"

Are you serious?
Not in a thousand years -- and not even then
Well, considering Facebook is one of the largest advertisers and millions of folks provide enough information to them for them to infer current mood pretty much continuously, a thousand years came pretty quick :)
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.