@@adissabovic If we gonna play the what if game let me get on complaining already on AI for taking away my (not yet achieved nor invented) 2037 "70s alternative rock analog mixing and mastering full time prompter and researcher" title from the online from the "Daniel Butterscotch" fully autonomous AI driven University
3:06 So wait, am I reading into this wrong? It sounds like your view is that newcomers ("bottom-feeders") should be completely gate-kept away and only those already in the industry for years should be worthy enough to survive the AI fallout? I really do hope I heard that wrong, because that's how an entire industry slowly collapses as retirees don't get replaced.
My fault for not being explicit enough. I'm referring to bottom-feeders as those that advertise services that misalign the expectations of customers from the actual product they deliver; "Industry standard mix" "Pro-Level Mixes" "Chart-Topping sound" $100 a mix.
@@panorama_mastering Hey no problem. Thanks for clarifying! I absolutely do understand what you're getting at now. Basically the scammers. I can totally see your point about AI gutting out the space where scammers thrive, yes.
Maybe AI will be integrated into the daw itself so each track has an automated mixer and channel strip. Then prompt for sound/style. Just wondering how it's going to look.
@@thank_you_thank_you I know, when you're focused on production, something to get most of the heavy listening done in the mix stage would be brilliant.
So what about people just starting or trying to learn? It'd be unreasonable to expect new engineers to be able to charge more than what an ai service could. Also, just like AI mastering, AI mixing services will over promise on what they can deliver. Artists will get screwed just the same but instead of having someone possibly grow from mixing things and getting compensated for their time somone will be passively collecting income off of those artists getting screwed while no-one improves.
Not everyone has the luxury of gifting free labor. I'm comfortable charging for the work I do. It may not be perfect but it's an improvement on the demos worth more than what I've been charging. I still think it's good to empathize with people in different situations.
AI may well be able to analyse objectively and cheaply master recordings but it has an inherent weakness and deficiency that can never aspire to in being a human mastering engineer be they cheaper or more expensive. I'm surprised more people havn't figured this out🤔
@@CraigScottFrost AI can be very useful as you say. It can analyse, compute etc to our advantage but it has a major weakness: it can't feel. Again, as you've mentioned, a well mixed track means the mastering will be subtle which also means the ability to feel the track/music along with analysis. AI doesn't have a heart. People listen to music with their heart. To me a great sound engineer is one who has both the technical skill and the feel, a heart and mind - recording, mixing and mastering is therefore both an art and a science, the combination of which both AI and robots can never aspire to.
with apple having the ai chips or whatever its called... I really think Logic Pro should be on this ai thing.. like every channel has ai integrated and it can analyse all your channels all your busses etc and it can be part of siri obviously all optional.. but imagine saying hey siri.. analyse my track and flag any masking problems and give suggestions how too fix it. or like auto mix auto side chain .. listen to a reference track make it sound like this.. I mean its kinda endless the opportunities and then you can have it do purely technical non subjective artistic mixing like cutting frequencies cleaning audio etc etc etc .. sorry for the rant
THIS IS COOL way for AI to leverage efficient workflows that bolster mixing engineers skillsets and decision making for better results! Thanks for sharing!
Machine learning is only ever as good as its input data, and even then it will only ever give you the best "average" result based on its dataset. While over time, these tools may become proficient with popular genres, creative mixing styles and uncommon music genres will always have their place with humans.
i think you are being a little naively optimistic assuming that AI is not going to be as good as a human, we are just at the birth of AI and its already amazing at creative tasks, of course its going to be better than mixing engineers, its going to destroy mixing engineers. And that human touch or that human creative decisions are just going to be as simple as writing a prompt like "make it more compressed" or "put that maraca down a little bit" or "make it sound more like depeche mode". Mixing and mastering is going to be extremely easy, no need for engineers anymore. And this is around the corner.
Can't disagree and I have been an engineer for over 30 year. We can argue the toss about how good it is now, but 5-10 years? Forget it. Your scenario will be accurate. And it will go further like 'turn the piano into a hammond organ', 'revocal as Sting'. Whatever I'm sure it will all come down the line.
I'm afraid AI services will generate generic results without an emotional connection from the listener. Weeell, let's see how how much numb we can get 🙂
Unless you are in the top 0.1% as an artist, mixing is overrated in 2024 for so many reasons. I spent 30 years honing in my ability to make a perfect mix and now most people listen to mp3s streamed from their phone speakers or earbuds and could care less about the meticulous details us mixing engineers toil over. Its against my own interest but I would encourage most people these days to mix themselves or use AI and put their money towards rent/food/gas etc. Recordings are cool and fun to make but the venue is too oversaturated with diminishing ROI. Focus on the live experience if you want to survive. I know it sounds harsh but its just being adaptable. Cheers
I know plenty of artists making a full time living not in the top 0.1% who can justify the investment in mixing their records. The artists who make it work leverage their music IP as a long term assett, have publishers behind them, synch opportunities, and tour. It's certainly not majority of artists, but it's much greater than 0.1%....
I for one know the AI wave is coming for us, and the only way is to adapt and implement. But In general this I think is a bit yeah-nah optimistic and missing the macro supply/demand outlook in the coming decade. (hello from west footscray) You say that AI has put upward pressure on engineers to do better and raise the default bar for these services, cutting 'the chaff from the wheat,'. But it’s just shriking the overall job market for engineers altogether (this is already being seen in other industries - see midjourney and artists, website design, photography, film/audio editing - Heck, I've been told first hand projects at audiobook companies working with elevenlabs eventually phasing out narrators/producers). Yes, the higher tier services are now highly valued, but the idea of AI forcing better quality backfires by simply reducing opportunities for new and mid tier engineers to practice and grow - you effectively cut the supply side for newcomers as theyre immediately priced out. And in regards to 'bottom-feeders'; no one wants to market themselves as 'decent mix for $100' because you intentionally diminish the perception your own product. And those who promote themselves as "quality" or "professional" for cheap and don't deliver are (through natural market forces) weeded out - the value add you attach to your price is a huge mismatch, and as a result no one is going to recommend you or go to you. If AI is raising the bar for what the default standard of mixing services should be (separate good engineers from mediocre ones for the value they demand) then it logically follows that as AI continues to improve (and it definitely will be), it will eventually reach a point where even pro services may be rendered obsolete from the equation, as AI services that even achieve 80% of pro result for 1/10th of the price will be the industry norm, effectively eliminating a huge supply side of new engineers to replace old. You've now eliminated the years of experience between clients and the techniques these engineers would learn. You also don't have to look far for evidence of this occuring in an industry. One of the most notable historical examples of automation leading to a shrinking market and increasing disparity between bespoke producers and entry-level services is industrial textile production during the Industrial Revolution (my partner makes wedding dresses and learning about this was insane). The disparity between mass-produced, affordable goods and high-quality, bespoke textiles increased. While factory-made goods were ubiquitous, high-end, handmade textiles became a niche market, catering only to the wealthiest clients who could afford to pay a premium for artisanal crafts. Tangentially, AI pushing out subpar engineers will eventually lead to homogenisation. If the paradigm now shifts to most artists utilizing their $100 on AI tools for mixing, it would intensely homogenize the overall output, detracting from the uniqueness and individuality engineers bring and the connections they have helping to transduce an artists vision. Creativity lies within diversity, and AI's reliance on patterns will diminish existing sonic uniqueness. You also fall into an 'AI incest problem' where the data set it learns from is itself, because the productivity output of AI tools dwarfs that of humans, and as such the data pool it will eventually learn from will be itself. To connect this to a broader point, there’s a saying: "Show me the incentives, and I’ll show you the outcome." The incentive structure here is clear: if AI mixing tools are cheaper, faster, and "good enough," and only improving, then more clients will naturally gravitate toward them, regardless of their limitations. The outcome of this incentive is that fewer engineers, particularly newcomers, will have the opportunity to develop their skills or break into the market. Established professionals will thrive, but the industry will suffer from a lack of fresh talent and creative diversity. The outcome isn't a market that fosters better engineers-it's one that drives out aspiring talent and makes the industry more homogenized overall, and drastically impacts the supply demand cycle between bespoke engineers and newer entrants to the market - the data is there to support it.
I was transparent in saying I paid for it but am yet to review it. There’s no links to the software in the description. That’s no recommendation or call to action.
In 5 - 10 years time AI will be able to Mix and Master the same as any successful commercial song to this day (or that day when it happens), it will Mix and Master better and more consistent than any human can in a matter of seconds, not hours or days, seconds. There will be no need for humans.
Many people will always prefer a human working on their stuff, cuz of their principles. There will forever be a market. Think of this: Pop music dominates the charts, but it doesnt mean country music cant exist alongside of it
There's a lot of context that's an algorithm misses. The whole communication channel, exploration of ideas and ideation of potential mixes that could result through an engagement with a proffesional happens on almost every project. AI assumes there's a predefined solution. There isn't.
@@panorama_mastering Thats AI right now, just look how AI video has improved in just one year. Over the course of the next decade not only will the AI's get more training data, also the method of training, level of understanding and AI's level of adaptability will improve as well. Just in terms of communication why wouldn't I be able to write a prompt to my AI mix engineer to make artistic and creative adjustments I want just like a regular human engineer. It will inevitably become indistinguishable from any avg engineer and possibly top tier engineer especially over the coming decades.
Hmm, why would you waste time separating the processes when you can prompt release ready songs? Just like video you prompt and get a finished scene, you don't go into a seperate process to do the color grading, framing, etc.
@BRamen. There's literally a published paper that denies everything you're saying. More data would not mean better ai. We've actually hit a bottleneck. Unless a entirely new system is developed, you won't be seeing better ai.
With Ai, there's no future in this. Mixing or mastering, it will do it better than any human. Is just a matter of time. And I am not talking about all these crappy "supposed AI" mastering services that are online now. I am talking about what's gonna come soon. We need to Adapt
There's genuinely not enough training data for that. AI will start training on itself and generation loss will compound. Meanwhile there will be a gap in engineers from the time AI was marginally useful because new engineers will not be getting the experience they'll need. Ultimately it'll hurt the entire industry and long term I expect a sharp decrease in the quality of music coming out.
AI can’t steal my job if I don’t have one to begin with
Could it be that you don't have one because the AI already stole it? :)
@@adissabovic If we gonna play the what if game let me get on complaining already on AI for taking away my (not yet achieved nor invented) 2037 "70s alternative rock analog mixing and mastering full time prompter and researcher" title from the online from the "Daniel Butterscotch" fully autonomous AI driven University
@@user-rd6jt3op8p Permission granted. 😆
Yikes 😂
I can give you a job
3:06 So wait, am I reading into this wrong? It sounds like your view is that newcomers ("bottom-feeders") should be completely gate-kept away and only those already in the industry for years should be worthy enough to survive the AI fallout? I really do hope I heard that wrong, because that's how an entire industry slowly collapses as retirees don't get replaced.
My fault for not being explicit enough.
I'm referring to bottom-feeders as those that advertise services that misalign the expectations of customers from the actual product they deliver;
"Industry standard mix" "Pro-Level Mixes" "Chart-Topping sound" $100 a mix.
It will inevitably hurt both
@@panorama_mastering Hey no problem. Thanks for clarifying! I absolutely do understand what you're getting at now. Basically the scammers. I can totally see your point about AI gutting out the space where scammers thrive, yes.
I have you on facebook for years. I really enjoyed the part that you said about client vision. Reminded me why I subbed to your content.
Maybe AI will be integrated into the daw itself so each track has an automated mixer and channel strip. Then prompt for sound/style. Just wondering how it's going to look.
thats exactly where its headed, no doubt.
@@thank_you_thank_you I know, when you're focused on production, something to get most of the heavy listening done in the mix stage would be brilliant.
No Windows yet for OSMix BTW so be forewarned...
what?
@cvaderx At the time of this video there was no Windows support.
So what about people just starting or trying to learn? It'd be unreasonable to expect new engineers to be able to charge more than what an ai service could. Also, just like AI mastering, AI mixing services will over promise on what they can deliver. Artists will get screwed just the same but instead of having someone possibly grow from mixing things and getting compensated for their time somone will be passively collecting income off of those artists getting screwed while no-one improves.
Don’t charge until you feel comfortable taking peoples money, that’s what I did, mixed hundreds of songs for free to gain access xperience
Not everyone has the luxury of gifting free labor. I'm comfortable charging for the work I do. It may not be perfect but it's an improvement on the demos worth more than what I've been charging. I still think it's good to empathize with people in different situations.
I'm guessing OSMix can't be used for things like EDM where there are 30 or 40+ tracks?
You can put higher track counts in; I put 33 tracks in it yesterday.
AI may well be able to analyse objectively and cheaply master recordings but it has an inherent weakness and deficiency that can never aspire to in being a human mastering engineer be they cheaper or more expensive. I'm surprised more people havn't figured this out🤔
Bang on.
@@CraigScottFrost AI can be very useful as you say. It can analyse, compute etc to our advantage but it has a major weakness: it can't feel. Again, as you've mentioned, a well mixed track means the mastering will be subtle which also means the ability to feel the track/music along with analysis. AI doesn't have a heart. People listen to music with their heart. To me a great sound engineer is one who has both the technical skill and the feel, a heart and mind - recording, mixing and mastering is therefore both an art and a science, the combination of which both AI and robots can never aspire to.
Is there anywhere I can buy one of those shirts?
prolly have to take the mastering course (on successful completion only )
with apple having the ai chips or whatever its called... I really think Logic Pro should be on this ai thing.. like every channel has ai integrated and it can analyse all your channels all your busses etc and it can be part of siri obviously all optional.. but imagine saying hey siri.. analyse my track and flag any masking problems and give suggestions how too fix it. or like auto mix auto side chain .. listen to a reference track make it sound like this.. I mean its kinda endless the opportunities and then you can have it do purely technical non subjective artistic mixing like cutting frequencies cleaning audio etc etc etc .. sorry for the rant
Taking notes, very insightful!!!
THIS IS COOL way for AI to leverage efficient workflows that bolster mixing engineers skillsets and decision making for better results! Thanks for sharing!
Machine learning is only ever as good as its input data, and even then it will only ever give you the best "average" result based on its dataset. While over time, these tools may become proficient with popular genres, creative mixing styles and uncommon music genres will always have their place with humans.
I've been looking for something like this! I'm convinced AI will have an big impact on us, so I'd better be prepared
i think you are being a little naively optimistic assuming that AI is not going to be as good as a human, we are just at the birth of AI and its already amazing at creative tasks, of course its going to be better than mixing engineers, its going to destroy mixing engineers. And that human touch or that human creative decisions are just going to be as simple as writing a prompt like "make it more compressed" or "put that maraca down a little bit" or "make it sound more like depeche mode". Mixing and mastering is going to be extremely easy, no need for engineers anymore. And this is around the corner.
Can't disagree and I have been an engineer for over 30 year. We can argue the toss about how good it is now, but 5-10 years? Forget it. Your scenario will be accurate. And it will go further like 'turn the piano into a hammond organ', 'revocal as Sting'. Whatever I'm sure it will all come down the line.
T-Shirt is 🔥🔥🔥
Thanks m808
interesting perspective, well said 👍
Oh man... so many thoughts. Not sure if I should leave a comment or make a video. haha.
DO BOTH ! :) Big
I'm afraid AI services will generate generic results without an emotional connection from the listener. Weeell, let's see how how much numb we can get 🙂
That's the beauty of it. It will weed out mixers who do the same thing... they're there!
most of the music is already copy n paste, AI will just do it more efficiently.
Have you heard pop anything these days? It all sounds the same.
Unless you are in the top 0.1% as an artist, mixing is overrated in 2024 for so many reasons. I spent 30 years honing in my ability to make a perfect mix and now most people listen to mp3s streamed from their phone speakers or earbuds and could care less about the meticulous details us mixing engineers toil over. Its against my own interest but I would encourage most people these days to mix themselves or use AI and put their money towards rent/food/gas etc. Recordings are cool and fun to make but the venue is too oversaturated with diminishing ROI. Focus on the live experience if you want to survive. I know it sounds harsh but its just being adaptable. Cheers
I know plenty of artists making a full time living not in the top 0.1% who can justify the investment in mixing their records.
The artists who make it work leverage their music IP as a long term assett, have publishers behind them, synch opportunities, and tour. It's certainly not majority of artists, but it's much greater than 0.1%....
I for one know the AI wave is coming for us, and the only way is to adapt and implement. But In general this I think is a bit yeah-nah optimistic and missing the macro supply/demand outlook in the coming decade. (hello from west footscray)
You say that AI has put upward pressure on engineers to do better and raise the default bar for these services, cutting 'the chaff from the wheat,'. But it’s just shriking the overall job market for engineers altogether (this is already being seen in other industries - see midjourney and artists, website design, photography, film/audio editing - Heck, I've been told first hand projects at audiobook companies working with elevenlabs eventually phasing out narrators/producers). Yes, the higher tier services are now highly valued, but the idea of AI forcing better quality backfires by simply reducing opportunities for new and mid tier engineers to practice and grow - you effectively cut the supply side for newcomers as theyre immediately priced out. And in regards to 'bottom-feeders'; no one wants to market themselves as 'decent mix for $100' because you intentionally diminish the perception your own product. And those who promote themselves as "quality" or "professional" for cheap and don't deliver are (through natural market forces) weeded out - the value add you attach to your price is a huge mismatch, and as a result no one is going to recommend you or go to you.
If AI is raising the bar for what the default standard of mixing services should be (separate good engineers from mediocre ones for the value they demand) then it logically follows that as AI continues to improve (and it definitely will be), it will eventually reach a point where even pro services may be rendered obsolete from the equation, as AI services that even achieve 80% of pro result for 1/10th of the price will be the industry norm, effectively eliminating a huge supply side of new engineers to replace old. You've now eliminated the years of experience between clients and the techniques these engineers would learn.
You also don't have to look far for evidence of this occuring in an industry. One of the most notable historical examples of automation leading to a shrinking market and increasing disparity between bespoke producers and entry-level services is industrial textile production during the Industrial Revolution (my partner makes wedding dresses and learning about this was insane). The disparity between mass-produced, affordable goods and high-quality, bespoke textiles increased. While factory-made goods were ubiquitous, high-end, handmade textiles became a niche market, catering only to the wealthiest clients who could afford to pay a premium for artisanal crafts.
Tangentially, AI pushing out subpar engineers will eventually lead to homogenisation. If the paradigm now shifts to most artists utilizing their $100 on AI tools for mixing, it would intensely homogenize the overall output, detracting from the uniqueness and individuality engineers bring and the connections they have helping to transduce an artists vision. Creativity lies within diversity, and AI's reliance on patterns will diminish existing sonic uniqueness. You also fall into an 'AI incest problem' where the data set it learns from is itself, because the productivity output of AI tools dwarfs that of humans, and as such the data pool it will eventually learn from will be itself.
To connect this to a broader point, there’s a saying: "Show me the incentives, and I’ll show you the outcome." The incentive structure here is clear: if AI mixing tools are cheaper, faster, and "good enough," and only improving, then more clients will naturally gravitate toward them, regardless of their limitations. The outcome of this incentive is that fewer engineers, particularly newcomers, will have the opportunity to develop their skills or break into the market. Established professionals will thrive, but the industry will suffer from a lack of fresh talent and creative diversity. The outcome isn't a market that fosters better engineers-it's one that drives out aspiring talent and makes the industry more homogenized overall, and drastically impacts the supply demand cycle between bespoke engineers and newer entrants to the market - the data is there to support it.
It's like a really shitty version of band lab lol
thanks for making me waste 40 bucks buying it. I'm going to message the devs and hopefully try and get my money back.
I was transparent in saying I paid for it but am yet to review it. There’s no links to the software in the description.
That’s no recommendation or call to action.
That plugin sucks Bonkey Dalls! Don't buy it. And it isn't AI, either.
Video makes no sense.
In 5 - 10 years time AI will be able to Mix and Master the same as any successful commercial song to this day (or that day when it happens), it will Mix and Master better and more consistent than any human can in a matter of seconds, not hours or days, seconds. There will be no need for humans.
Many people will always prefer a human working on their stuff, cuz of their principles. There will forever be a market.
Think of this:
Pop music dominates the charts, but it doesnt mean country music cant exist alongside of it
There's a lot of context that's an algorithm misses.
The whole communication channel, exploration of ideas and ideation of potential mixes that could result through an engagement with a proffesional happens on almost every project.
AI assumes there's a predefined solution.
There isn't.
@@panorama_mastering Thats AI right now, just look how AI video has improved in just one year. Over the course of the next decade not only will the AI's get more training data, also the method of training, level of understanding and AI's level of adaptability will improve as well. Just in terms of communication why wouldn't I be able to write a prompt to my AI mix engineer to make artistic and creative adjustments I want just like a regular human engineer. It will inevitably become indistinguishable from any avg engineer and possibly top tier engineer especially over the coming decades.
Hmm, why would you waste time separating the processes when you can prompt release ready songs? Just like video you prompt and get a finished scene, you don't go into a seperate process to do the color grading, framing, etc.
@BRamen. There's literally a published paper that denies everything you're saying. More data would not mean better ai. We've actually hit a bottleneck. Unless a entirely new system is developed, you won't be seeing better ai.
With Ai, there's no future in this. Mixing or mastering, it will do it better than any human. Is just a matter of time.
And I am not talking about all these crappy "supposed AI" mastering services that are online now. I am talking about what's gonna come soon.
We need to Adapt
I don' think so I think labels are ready to pay for experience and someone who has a feel for things
Every track mix/master you ever loved has been made by a skilled human with ears, emotions, taste and a vision
Yep, it is going to happen.
There's genuinely not enough training data for that. AI will start training on itself and generation loss will compound. Meanwhile there will be a gap in engineers from the time AI was marginally useful because new engineers will not be getting the experience they'll need. Ultimately it'll hurt the entire industry and long term I expect a sharp decrease in the quality of music coming out.
@@JazzyFizzleDrummers Yeah but to actually use A.I. is and will be a choice.