SUNO ai V3 Hacks, Tips And Tricks
ฝัง
- เผยแพร่เมื่อ 6 ก.ย. 2024
- In this video, I'm diving deep into the world of music generation with SUNO AI V3. 🎵 Get ready for a masterclass in unlocking the absolute best AI-generated songs possible! 🎶
Join me as I share hacks, tips, and tricks to harness the full potential of SUNO AI V3. From refining melodies to perfecting harmonies, we'll explore every aspect to ensure your AI compositions hit all the right notes. 🎼
Whether you're a seasoned musician or a curious creator, this video is your key to unleashing the power of AI music generation like never before. 🚀 Don't miss out on the ultimate guide to crafting unforgettable tunes with SUNO AI V3!
🔑 Unlock the future of music with Chris Unlocks AI! 🔑
#ChrisUnlocksAI #AIComposer #SUNOAI #MusicGeneration #AIHacks #MusicTips #CreativeTech #FutureOfMusic #ArtificialIntelligence #MusicProductionHacks #CreativeProcess #aimusic #artificialintelligenceinmusic #aimusictools #sunoai #ai #creativeai #
Put in your styles: Anthem or Anthem rock for huge driving choruses.
Great suggestion. Thank you!
[Instrumental Intro] At the beginning and [Instrumental Outro] at the end works best for me, but I have to say that it doesn't always make an intro or outro, unfortunately it's currently more gambling to get the intro or outro.
I have pretty good luck with it. Although it's not perfect.
Sometimes it will start with someone singing. I can work with that as long as it's not starting mid word.
Thanks for the great comment and watching.
[End] always works well for me
@@Verint97 Okay, thanks for the info, I'll definitely try it. ✌️💪
Thank you Chris.. you are really helping me to unlock AI music. I am a guitarist composer vocalist (kind of). I am looking for folks like yourself and others to collaborate .I will post examples..the last couple days I am experimenting with instrumentals to give myself backing music and scores for my video productions..more on that later..again thanks so much for your help..sorry for rant
Heck yeah. I personally think this could be a fantastic tool for guitar backing tracks. And video music. Let's collaborate!
The " kind of " explains the need for AI
Just curious, What genre music are working on?
I've the impression the Ai tends to follow the brackets more if they're correctly capitalized. When is written in all lowercase tends to follow them less...
I agree. And the brackets help a ton. Thanks for sharing.
In my experience, it always treats the first lyrics as an intro, so if you entended any of those it would start to implement your tags more in the way you expected in the next Chorus.
Great tip, thank you for sharing.
IT does NOT work consistently. [intro] in my cases is like 98% ignored. Is skips lyrics in 80% ...
That sucks. I've not been having that issue at all. Sometimes it doesn't work for me, however most cases it does. Have you tried [instrumental Intro] ?
Let me know what you are prompting other than just intro. I will try it on my end.
Thanks for the comment and hopefully it will start giving you better results.
@@ChrisUnlocksAI I tried [intro, synth solo], [instrumental], [melodic intro], [percussion intro], [bass intro], [instrumental intro], [music intro], ... what's not. Prompt HOUSE, TECHNO, DANCE, ... where it is obvious no song starts with lyrics. In like most cases (but not FLAMENCO Gypsy Guitar CATALAN, BALKAN BRASS BAND, DOO-WOP), songs starts with lyrics. I figured that giving full song worth of text into it, it will try to make whole song in 2 minutes (it will skip as much of verses it need and just glue last bit of text to finish song - I admit I used [END] at those times. If I just used portion of text, like [verse], [Verse1] and [chorus], than that would be squeezed at the start, last minute will be instrumental. -> So I started with instrumental songs, than extend at certain point. It comes out better. Now I make sure that I give [bridge], [continue] or something at end of verse, so it does not try to end song there.
SKIPPING of lyrics. It might do with the fact it is not English. IDK. But often it skips whole verse, or just portion of it. At that time (I spend some credits), I gradually add "mmm aaa, ooo" before my wanted lyrics and eventually it picks it at right spot. I did 20+ repeats before adding "mmm's" and not once picked right spot. In few times if it did, the voice was garbage then. I also learned that it is rigid how many lines it expects. If I happen to guess it, it'll be great. If not, It does not adapt the song length but rather skip lines out or sing last few lines of verse as [chorus] already. Than chorus got sliced and it starts verse ... let say I write 2x verse by 4 lines and than 4 lines of chorus. But song structure expects 3 lines long verses ... a mess. Again I learned it is better to just try to make one verse per extension. Fiddly.
SUNO try to make everything modernized ... DANCE, or TECHNO, REGGAE, POP, ROCK, or most prompts will be contemporary arranged. Today songs does start with lyrics sometimes. It is mostly impossible to recreate 1990s techno or trance. If I add [1990s] prompt it will produce garbage that is not 1990s, not good. Voice will be garbage too. All sorts of problems. Little sampling in AI learning as it seems, so no variation in song sounds nor structure. So it is best leaving as it is, with minimal prompt. Les it better. Often I aim for "this style" and get something completely else. If it is good, OK. Leave it.
ON the other hands, UDIO is very good at recreating proper time stales. 1990 and 1998 is different. If I wanted to recreate "pizzicato" (1st known from trance song "Insomnia") SUNO gave me nice orchestral song without any trace of trance or drum kick. UDIO give the "pizzicato" by it self when I just added "1998" with "trance", without "pizzicato" prompt. Also I find less fiddly to make music with UDIO. Also it has "backward" extension. BUT (huge deal!) it does not have TIME settings where to extent.
ALL in ALL SUNO is more refined in sound (where prompt is simple) and no specific era style is expected. UDIO is better at recreating exact styles. UDIO had better voices than SUNO, more clear and less sunken in the music, and [clean voice] is clean, solo voice, not some harsh resonant choir. It is hard to get "rapped verse" in SUNO. I abandoned idea ... or to get song where woman AND male has duet with each sing one verse.
This AI feature does not yet have a publicly available "seed". So I think, to create a consistency without seeds is something that is difficult to do.
@@ChrisUnlocksAI [Percussion INTRO], [Instrumental Intro], [Music intro], Music, Percussion, beat, bass intro, what's not. Also SUNO like 99% starts with vocal in DANCE, trance, pop, EDM, ... like 90% od styles starts with vocal, sonsuming verse 1 for it. I adaoted lyrics with "pre-verse". When I rather start with instrumental 1st part and than extend with lyrics I mostly stuck in 2nd or 3th extension. It rarely goes well, consuming ton of time (and credits).
Meanwhile I use UDIO with more joy. It is not the same model and exact same results of music can not be created, but as for production is FAR more satisfying. Extension can go both ways + special "intro" and "outro" extension which guarantee good ending or start. Instead of creating song from start and hoping it'll go well with verse and chorus UDIO gives you straight verse or chorus and IF result is good than can be added parts BEFORE and INTRO. They added three structure to see how song parts are related, they added CROP function to select just the portion of liking ... there is far less lyrics skipping and ignoring than SUNO. Hallucinating happens sometimes, but crop function can help here. So in the end song creation is faster with more predictable result. While SUNO creates more sophisticated music, UDIO is more joyful to use.-
Back to SUNO: I use it for 6 months, and bought subscription and extensions. I realised I use 500-800 credits to get song - no problem, I am prepared to pay, but it is frustrating to fiddle with lyrics - adding gibberish in front of real lyrics, to force it to pick right line, when he decides to skip lines or verses. It is frustrating also for voice not being proper ... I see V2 is much more consistent and predictable. V3 sounds better, but the best result is given when song is two minutes without extension. If lyrics and whole structure is able to fit into 2 minutes. I can't enjoy in song when words are pronounced all wrong and weird. If English lyrics it is somewhat better.
If you put fewer words at the beginning of the lyrics, you often get an intro, but sometimes an intro that's too long.
Little FYI, when you type something, read it out. If anyone’s watching on a phone, they can’t read tiny text on the screen.
Great advice, thank you and will be more conscious with that.
I'm watching it on the phone and you can use your fingers to increase the size
I'm watching it on the phone and you can use your fingers to increase the size
I did like suno but once their censorship kicked in I refuse to support them. I want to be able to make a song about what I want to without someone telling me what words I can and can't use.
I have had it not do a prompt when it thought I was saying something wrong in the lyrics. That can be frustrating. You can always try ace studio, drop in whatever lyrics you want with it.
Hope this helps, thanks for watching.
@@ChrisUnlocksAI I'll take a look at it. Thanks
I've just discovered Suno. What a great way to put my novels and FRPG to music! Thank you for giving some tips. I was getting a bit frustrated, but you made some things clear and my lastest is posted. I owe you pizza :)
Is there a list somewhere of what commands you can put in Brackets that the AI will understand? Or is it just trial and error
Unfortunately trial and error. Sometimes it ignores the commands. Hope this helps
It's not easy to create a good song, even AI Tool. Thank you Chris.
Thank you.
Thanks Chris go on with Suno !
Thank you!!
i wanna achive an Nirvana 90s type rock with it but it doesn't work,can you give me some tips on how i could do that?
Try 90s grunge, Seattle Rock, or pop rock. Male vocals.
I would keep trying to generate at least 5 times, then switch it up. Hope this helps.
Also, UDIO does 90s rock pretty good and its fee at the moment. You may check it out.
@@ChrisUnlocksAI Oh alright,thank you :D
Write arpeggio in style. It will😄give nice arpeggios in your song. I am from India. I made nice hindi, Malayalam, tamil song using this
I can't wait for V4 I just need to have the songs go 3 minutes and it will be perfect.
I'm excited for V4 as well. You can make your songs longer than 3 minutes as long as you keep extending it then once you have gotten all the parts you like click on get whole song. Hope that helps, thanks for watching.
22:48 min: Spotify already called and bought his song! 🤑
hahahaha
Hahaha yeah they did!
I have a iPhone 7 and the apps aid I can use unless software is in the 17
I think that will work.
To be fair, if you start a song 'theres a boy with a guitar strumming along' it'll never be good 😂
😂 solid point!!
maybe becaust you put (... beast so stong....) in stead of beats so strong.... just a thought.
Thank you for the video . This is a great example of how fickle and uncooperative AI can be with Suno or any AI generator, music, images, or video. I did learn some nuances though.
Thank for watching.
I bet you could do a parody of one of those compilation album commercials with this, kinda like “Song Styles” on WLIIA, but with fully produced AI songs.
I have been in the parody mindset since I first got into ai music. We are on the same page!!
@@ChrisUnlocksAI i got it to switch styles kinda th-cam.com/video/kl4pCIItZlo/w-d-xo.html
I like that idea. Would love to hear what you come up with. This isn’t the same thing, but I took the lyrics to a public domain hymn, and created in the style of soulful blues… it didn’t have the same melody as the original… but man! It sounded sooo cool
Thanks Chris, I wonder if you have a solution avoiding the harsh cut at the end of the music ?
I have had great success by generating the end, you can use [Outro] or [instrumental Outro] or even [fade out]
It may take a couple generations to get what you are wanting.
Also you can drop it into a Daw and simply fade the ending.
Hope this helps, and thanks for watching.
Omg, you gave me so may ideás with both Suno and Udio,
with this viseo, thak ou!
Awesome. I'm glad to have helped. Thanks for watching.
after chorus we got noisy audio quality, only good at intro and verse in the beginning
I've been having some issues with that as well. Not always but here lately. Hope it gets better.
Thanks for watching
[Choir]
Awesome!!! Where do we learn more about these tags?
I will try to put together a list to share.
I am subscriber. Look forward to future posts like this. Thanks for posting this
Anyone having luck with keeping the same vocalist and band style across multiple songs yet? It would be great to make an album using the same AI artist. I've found that i find a great sounding vocal but can't use it in another song.
You may want to look into ace studio. Perhaps separate the stems then replace the vocal track. Hope this helps.
Thanks for watching.
Cool cheers from portugal
Thank you, and Cheers to you from the US.
Is there any way to build the song in a specific key?
I've not had any luck with getting a specific key. Thanks for watching.
Maybe load your track into a DAW like Mixcraft and change the key there?
you my friend, have no idea how to input proper in this program. Big Chorus haha. Wasted my life watching this.
You my friend have no idea what I have created using "big chorus". I have been very successful using this terminology. You can move on, instead of being a negative person try finding peace in yourself my friend.