AI Music Revolution

QA: You Asked, I'll Be Straight With You

• Josh • Episode 27

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 20:49

Send us Fan Mail

This week I'm doing something different. I pulled the questions you've been sending me over the last couple of weeks, from email, from comments, from the community, and I'm answering them straight. We cover how much myoney you can actually make releasing AI music, whether mastering really matters, why I use Reaper, the best ways to pull stems outside of Suno, how to prompt for what you want instead of what you don't, and what's behind the rise in copyright notices Suno users are seeing. If you've got a question you want answered on a future episode, email me at josh@jgbeatslab.com and put PODCAST in the subject line.

The Unlock System is JG BeatsLab's methodology for serious musicians working with AI tools. Lane 2 work: human-authored, AI-assisted music creation.

Visit JG BeatsLab: https://www.jgbeatslab.com

Every Thursday morning, I send out a newsletter. One email a week. It's where I share what I'm working on in the studio, the research running in the lab, and the methodology I'm refining in real time. New books, new Blueprints, new findings from the Red Lab Protocol research. If that sounds useful, you can sign up at jgbeatslab.com/newsletter. Thanks for listening, and I'll see you next Friday.

AI Music Revolution is produced by JG BeatsLab LLC, an AI music education company building the methodology, research, and community for serious creators working in Lane 2.

Get more from JG BeatsLab LLC:

Connect:

Contact: josh@jgbeatslab.com

Stop gambling. Start directing.

SPEAKER_00

Hello and welcome to the AI Music Revolution. I am your host, Josh Gelloland, the founder of JG Beats Lab. I get questions all week long. Some come in by email, some show up in the comments on posts, some come from inside the community. A lot of them are really great questions, the kind that deserve real answers and not just, you know, one-line replies in a comment thread. I also get a lot of questions that I just don't have time to answer. As you can imagine, running the business, etc., you don't have time to get to every question that comes in. So what I want to do today is I'm going to do something a little different. I'm going to work through a batch of these questions, one at a time, and give you the honest version, the honest answer for each of them. If you have questions and you want me to take a turn of answering them on a future episode, there's an easy way to get in front of me. And I'll tell you how to do that at the end. But for now, let's get into some questions. Alright, first question. And it's the one almost everybody is thinking even when they don't actually ask it out loud. And that question is hey Josh, how much money can you actually make from AI music? Great question. Let me answer that truthfully. I'm gonna give you the answer that most people won't. The vast majority of people who release music, it doesn't matter if it's AI assisted or not, they do it for the love of music, the love of releasing music, the love of creating music. They love the process, they love the feeling of finishing something and putting it out into the world. That's the real engine here. Now, some of those people make a little money back along the way, and a few more make a little bit more than that. But the honest middle of that distribution is not what the get rich quick crowd is selling you. Let me give you a real example. So someone I work with told me that their first few months of releasing around $300 in streaming revenue. That's genuine money and it's and it's growing. But in that same window, they spent about $1,000 on promotion. So on paper, that's a net loss. And here's the thing they understood that I think people in this space miss. So let me be blunt, because I'd rather you hear it from me than learn it the expensive way. If you are in this purely to make money, you will most likely be disappointed. The math is very hard, the market is extremely crowded, and it's changing fast. But if you come into this with passion as the thing that's really driving you, and you make some money along the way, huh? That's an upside. And every now and then, someone will hit a song or a genre or a niche that people really gravitate towards, and the streams will spike and doors will open. A lot of things are possible in this industry. I have seen them happen. Just don't build your whole reason for being here on the thing that is absolutely the least in your control. All right. Next question. Is mastering my AI tracks really that important? Will listeners even notice if I skip it? I am not going to hedge this one. Mastering is not optional. It's a necessity. And I'm not saying this as a guy selling you a mastering book. I'm saying this as someone who sits on the other side of the desk. As you may know, I curate playlist on Submit Hub. I listen to submissions all day long. And I can tell you within a few seconds which tracks have not been mastered. They sound smaller, they sound duller, they sound quieter and less finished, less polished than everything else I'm listening to. And it's not subtle from where I'm sitting. And I'm exactly the kind of gatekeeper you're trying to get past. And here's why it matters even more with AI music specifically. The tools tend to hand you something that already sounds a little dull, maybe a little dark. I covered this in the Murika video recently. Tracks come out spectrally darker than a commercial reference, missing some air on the up top of the song. So if you skip mastering, you're not releasing a neutral track. You're releasing one that's already starting a step behind. And when a curator hears that, a listener will hear that too. Even if they can't name what's wrong, they know something's wrong. They know that this sounds off. And what do they do? They just move on to the next song. So if you want to release music and have people actually listen to it, mastering is the price of a mission. It's not a nice to have. The thing that gets your work taken seriously is your mastering. Okay, that leads straight into the next question. Because if you're going to master your tracks, you need something to master in. And one of you commented on the recent video I posted and asked, Why did I choose Reaper? Well, there's three primary reasons why I chose Reaper. And these are the same three I'd tell anybody. First is cost. Listen, like we just talked about, you don't make money in this. Well, you can, but very few do. So costs are important. Keep your cost down. Reaper is a one-time $60 license that covers both Windows and Mac users with free updates for years and years. There's no subscription, there's no featured tiers where the cheap version is crippled. Everybody gets the same full software. For people who are coming into this who don't want to sign up for another monthly bill, that matters. The finances matter here. Second is power. This isn't a beginner's toy that you outgrow. Reaper runs in commercial studios. It runs inside of broadcast. It runs in game audio. It runs in film production. The ceiling is high enough that you will never hit it doing AI music work. You can grow into it for years and years and years. Third, and this is the one I think people tend to underestimate. There's a full sixty-day evaluation. Free. Not a crippled trial. I'm talking about the whole thing. You can run the entire process, master real tracks, and decide if it's for you before you spend a dollar. By the way, Reaper is not paying me in any way, shape, or form. I'm not an affiliate for Reaper, so this is just a user experience with Reaper. Now, I will be honest with you about trade-offs. Reaper is built by a small team called Cocos. And they care about function over looks. It's not going to win a beauty contest against Logic or Ableton. The interface is plain, it's boring. But once you know where everything lives inside of Reaper, inside of the interface, that plainness works in your favor, especially for the kind of cleanup and mastering that AI tracks need. So that's why. That's why I chose Reaper. And I am extremely happy that I did. Okay, staying on tools, this one came in from a curator who said every stem extraction they do in Suno now produces artifacts. And they wanted to know what to use outside of Suno. So first, the good news is I've done controlled testing on exactly this under one of the Red Lab protocol reports that's available in the Red Lab access. So I can give you more than just an opinion. Here is the short version of what we found in this testing. Suno's built-in stem extractor is a convenient option. And convenient counts for something, right? But it does have limits. It has real limits. It caps the top end of your stems at around 15 kilohertz. So you lose air and detail. And it bleeds in genre-specific ways where it struggles most of the things like hip-hop metal and country. So when you go outside of Suno, there are two tools that stood out. La LaLal.ai is purpose built for separation. It preserves the full frequency range, and the isolation is noticeably cleaner, especially on genres where Suno does struggle. It has its own artifacts. They all do. It's not magic, but for a lot of content, it's a clear step up from Suno. And then there's Isotope RX. And it's it's the professional standard in this space. It's got a feature called music rebalance that lets you shift the level of a part without fully separating it, plus surgical tools for removing the exact artifacts AI vocals tend to have a lot of. The catch is it's a professional tool and it comes at a professional price. Now here's the part I really, really do want you to hear because it speaks to how fast this space moves. Since I ran that analysis, Suno rolled out an updated stem separator. And on the surface, that sounds like progress. Yes, more stem separating development is a good thing, right? Unfortunately, the update is just more stems. So more instruments you can pull apart. It is not, as far as I can tell through my own experimentation, better stems. The quality and the artifacts are still there where they were before. And more separation is not the same as cleaner separation. Which, let's be honest, that's really what we're striving for is cleaner separation, not more of it. So the takeaway is this match the tool to your content and your budget. Test it on your own track before you commit, because the right tool is content dependent. And don't get fooled into thinking a higher stem count means anything from a better result standpoint. That's a directing decision, not a button you press. All right. Speaking of directing, here's a great one from the comments. A viewer made the point that AI, like the human brain, doesn't really understand negatives. You tell yourself, don't think about the elephant, and what do you do? You think about the elephant. And yeah, there's real truth in that, at least in practice. And it shows up constantly in music tools. With Suno and with Murika, describing what you want will almost always steer you better than telling the model what to avoid. If you write no drums, there's a decent chance you're you're going to get drums. Why? Because the model heard the word drums. The fix is to flip it. Instead of no vocals, you say pure instrumental. Instead of no electronic sounds, you say organic acoustic instrumentation. Prompt for what is there, not for what's missing or you don't want there. I'll add one honest caveat here because I don't want to oversell it. These models handle negatives better than they used to. And both platforms that I talk about here, that I do most of my work with, Suno and Eureka, they've dedicated exclude fields that exist for exactly this reason. So it's not that the model is incapable of understanding. No, it's that as a working habit, painting the picture you want is the stronger move. And that habit is most of what directing a track actually is. You're not telling the AI to avoid a wrong answer, you're describing the right one clearly enough that it has somewhere to go. Okay, moving right along. The last question that I have for this episode. And it's the heaviest of the bunch. Several of you have asked about the rise in copyright notices Suno users are seeing lately. What the hell is going on? I have seen this myself. So let me separate what's confirmed from what's anecdotal because this is exactly the kind of topic where speculation runs well ahead of the facts. So here's what is confirmed. Suno is under serious legal pressure right now. No surprise. Sony is the last major label still fighting Suno in court. And there's a key hearing set for this summer. The labels recently moved to expand their claims, alleging that tens of thousands of additional songs were used in training without their permission. Warner has already settled, they settled late last year, and Suno is in the middle of transitioning to licensing models or licensed models, along with new download restrictions. Free tier songs are becoming play and share only, and paid users are getting monthly download caps. On top of that, Suno's own policy already says it can block a generation, flip a song to link only, or pull material down without warning. And it follows takedown requests from rights holders. A platform like Suno that is under this much legal pressure, in the middle of moving to a licensed model, tightening what it flags and removes, that's exactly what you would expect to see. So more notices landing in people's laps fits the moment that we are currently in. What I am not going to do is tell you Suno announce some specific crackdown because I've not seen that announcement anywhere. I'd rather give you the climate and let you draw the line than feed you a rumor dressed up as news. What I will give you is what to do about it, because that part is what is actually in your control. Keep your human authorship trail, your lyrics, your edits, your DAW work, the decisions only you made. Register what you legally can and never treat any platform's library as your only copy of your own work. Pull your files down, back them up, and truly own them. That predictive productive posture is the right one, no matter how the lawsuits land. And if you want the full breakdown of the rights and registration side of the business, that's exactly what the music rights and registration book walks you through step by step. Okay, that's the batch for today. Six real questions, six straight answers. I try to give straight answers, I try to not candy coat things, and I try not to just make up stuff along the way. So hopefully these landed and they were helpful. And here's the part I promise you at the top. If you have a question for me that you want me to take on on this episode or on this podcast, I want it. Send it to me at Josh at JGbeatsLab.com. But be sure to put the word podcast in the subject line so I can find it. I read everything, and the good ones we're going to end up right here in an episode just like this one. Everything I publish, every book, every research report, the blueprints, the community of people actually doing this work, all of it lives in one place at jgbeatslab.com. If today was useful, that's where you go to go deeper. Thank you for listening, and I will talk to you in the next one.