That first song is really 🔥🔥🔥 Mind sharing the style prompt? Do you have a tutorial how to edit lyrics to a generated song without having to make a new song?
Thanks! I made the song public on Suno so you can reuse the prompt. Here's the link...suno.com/song/7e0a1204-8749-435e-9bff-d7e35175a2db Click the 3 dots, highlight "Create" and you'll see the "Reuse Prompt" option so you can use it for your own songs. I don't have a tutorial exclusive to only editing lyrics, but I'll give you the instructions here... 1. Find the song you want to edit in your library 2. Click the 3 dots 3. Hover over "Edit" 4. Click on "Song Details" This will show a pop up where you can edit the lyrics without creating a new song.
@@MasterAIFast I really appreciate you for sharing the link to your created track, also for the helpful suno tips! Looking forward to more of your content covering ai.
I guess it seems cool, but I don’t really see how it’s an improvement over the “extend” function. When I upload a song with vocals and I prompt a cover version that’s in the same style and genre all it does is remix my audio most of the time. It samples the vocalist and then recreates a vocal track that sounds almost exactly like the original. I thought that if it’s a “cover”, then it should basically sound like a different band playing the same song. The majority of the time it sounds like the same song with minor changes and a bunch of pitch corrections.
I've got it to recreate my own songs (where I'm singing) as other artists almost exactly like me, but I've also been able to totally change song structure, add verses, bridges, force styles, etc... Also, you'll find that it doesn't always recreate stuff as well, if it's like a noisy recording. So, on some I actually had to prompt (or negative prompt) in order to get it to more faithfully recreate my song. Other times, yeah it just nails it with no prompting. The negative prompting is so great, especially using it with the cover feature. But in general, I've wanted it for a long time. I've taken recordings of my voice that are mixed with very noisy bad quality instruments, separated the vocal with a free website, and only uploaded a weird noisy vocal with hints of instruments... and had it make it sound like Freddy Mercury, but with my vocal trills and melody... it's crazy. I love this update. The extend function is very different, actually. It's like extend tries to get the style of the song. The cover feature is more like, getting the essence/spirit of the song, and it can translate between styles. Extend is very useful when doing Covers though, because it generally loses it's coherence with the melody and stuff between 2:30 and like 3:20. And, they always stop al like 3:21, generally with the last 30 seconds or so all messed up. This is true whether you cover a song you made with Suno or if you uploaded 1 minute. The cover engine obviously just doesn't have the temporal range and falls apart at the end, losing melody and repeating the sam melody again and again a lot.
@@reallybigname I guess that makes sense. I wouldn’t really be trying for a much different style though. For example, I wouldn’t take a metal song and try to get a country cover version of it. In my case, I took some demos that my band made 25 years ago. The recordings are not great and the tuning and the singer’s pitch isn’t that of a professional studio recording. When I use the “extend” it tends to “sample” and retain the original sound quality and tones of the instruments, rather than sounding like a whole new recording session. With “cover” I would expect brand new instruments and a new voice and a whole new level of sound quality, even if I’m not asking for a style change. But I literally hear the original track mixed with the output. Most of my results are clearly more of a remix than a cover. In the real world, two bands that play the exact same style are still going to play very different-sounding versions of the same song. I’ll experiment with it more but so far it’s not very useful if every cover version literally contains the original recording in it.
You can add meta tags like [Crowd Sings] or [Harmony] to create this effect, but it's not perfect. Requires a few attempts to get it right. I created a video on how to add these kinds of tags to improve song creations. If you're interested here's a link to the video...ua-cam.com/video/40Pg2iNnWR4/v-deo.htmlsi=gii0dvuzBdn4XTQd
I agree, it is hard to get (especially male voices) voices that don't sound that way. You have to try for unusual combinations on prompt to get something that sounds different enough so that type of sound you are referring to doesn't sound so obvious. Udio does a better job on voices to my ear.
@@MasterAIFast AIVA can come up with some interesting results based on existing music you upload if you need some orchestral type music for your AI movie soundtracks. It's kinda hit and miss and I don't think it's had an update for a year now though. Still it made 3 nice pieces of music for me....out of 50 attempts :)
That first song is really 🔥🔥🔥 Mind sharing the style prompt? Do you have a tutorial how to edit lyrics to a generated song without having to make a new song?
Thanks! I made the song public on Suno so you can reuse the prompt. Here's the link...suno.com/song/7e0a1204-8749-435e-9bff-d7e35175a2db
Click the 3 dots, highlight "Create" and you'll see the "Reuse Prompt" option so you can use it for your own songs.
I don't have a tutorial exclusive to only editing lyrics, but I'll give you the instructions here...
1. Find the song you want to edit in your library
2. Click the 3 dots
3. Hover over "Edit"
4. Click on "Song Details"
This will show a pop up where you can edit the lyrics without creating a new song.
@@MasterAIFast I really appreciate you for sharing the link to your created track, also for the helpful suno tips! Looking forward to more of your content covering ai.
@@gamelife44 You're welcome :) Thank you for watching!
I guess it seems cool, but I don’t really see how it’s an improvement over the “extend” function. When I upload a song with vocals and I prompt a cover version that’s in the same style and genre all it does is remix my audio most of the time. It samples the vocalist and then recreates a vocal track that sounds almost exactly like the original. I thought that if it’s a “cover”, then it should basically sound like a different band playing the same song. The majority of the time it sounds like the same song with minor changes and a bunch of pitch corrections.
I've got it to recreate my own songs (where I'm singing) as other artists almost exactly like me, but I've also been able to totally change song structure, add verses, bridges, force styles, etc... Also, you'll find that it doesn't always recreate stuff as well, if it's like a noisy recording. So, on some I actually had to prompt (or negative prompt) in order to get it to more faithfully recreate my song. Other times, yeah it just nails it with no prompting. The negative prompting is so great, especially using it with the cover feature. But in general, I've wanted it for a long time.
I've taken recordings of my voice that are mixed with very noisy bad quality instruments, separated the vocal with a free website, and only uploaded a weird noisy vocal with hints of instruments... and had it make it sound like Freddy Mercury, but with my vocal trills and melody... it's crazy. I love this update.
The extend function is very different, actually. It's like extend tries to get the style of the song. The cover feature is more like, getting the essence/spirit of the song, and it can translate between styles. Extend is very useful when doing Covers though, because it generally loses it's coherence with the melody and stuff between 2:30 and like 3:20. And, they always stop al like 3:21, generally with the last 30 seconds or so all messed up. This is true whether you cover a song you made with Suno or if you uploaded 1 minute. The cover engine obviously just doesn't have the temporal range and falls apart at the end, losing melody and repeating the sam melody again and again a lot.
@@reallybigname I guess that makes sense. I wouldn’t really be trying for a much different style though. For example, I wouldn’t take a metal song and try to get a country cover version of it. In my case, I took some demos that my band made 25 years ago. The recordings are not great and the tuning and the singer’s pitch isn’t that of a professional studio recording. When I use the “extend” it tends to “sample” and retain the original sound quality and tones of the instruments, rather than sounding like a whole new recording session. With “cover” I would expect brand new instruments and a new voice and a whole new level of sound quality, even if I’m not asking for a style change. But I literally hear the original track mixed with the output. Most of my results are clearly more of a remix than a cover. In the real world, two bands that play the exact same style are still going to play very different-sounding versions of the same song. I’ll experiment with it more but so far it’s not very useful if every cover version literally contains the original recording in it.
Interesting that in your experience it retains the original sound quality and tones. Thanks for sharing.
Is it possible to add backing vocals, for example female for 2-3 voices?
You can add meta tags like [Crowd Sings] or [Harmony] to create this effect, but it's not perfect. Requires a few attempts to get it right. I created a video on how to add these kinds of tags to improve song creations. If you're interested here's a link to the video...ua-cam.com/video/40Pg2iNnWR4/v-deo.htmlsi=gii0dvuzBdn4XTQd
It's a great feature, but they only allow 60 seconds of music to be uploaded, which totally hampers the maximum potential of this feature.
I have a feeling Suno will increase this limit in the future if other AI song generators do the same.
like 2nd Cover Song...
Thank you. I made the song public. If you want to listen to it follow this link...suno.com/song/48120f5f-6445-4cd9-a6ff-d9810d9762df
They all sound autotuned (or pitch-corrected). They don't sound natural.
I agree, it is hard to get (especially male voices) voices that don't sound that way. You have to try for unusual combinations on prompt to get something that sounds different enough so that type of sound you are referring to doesn't sound so obvious. Udio does a better job on voices to my ear.
And the beats are always very generic and robotic .
Must check this out, AIVA didn't really do a great job 90% of the time even with instrumental/orchestral tunes I tried.
Thanks for watching. First time hearing about AIVA. Even though your experience was ideal I'm going to check it out.
@@MasterAIFast AIVA can come up with some interesting results based on existing music you upload if you need some orchestral type music for your AI movie soundtracks. It's kinda hit and miss and I don't think it's had an update for a year now though. Still it made 3 nice pieces of music for me....out of 50 attempts :)
@@retrotronics1845 Good to know. Thanks!
AIVA is GREAT if you want more control, since you can literally change the notes on their editor