Hi Steve, feel free to share the Suno link here if you like, so everyone can listen 🎶🎶

@@CB-AIthx

Hi @SiOfRelief, thank you for watching!

Thanks @TKNorway, I hope they'll post a fix soon, too. I'm of the opinion that many casual listeners are listening with an untrained ear and probably won't be bothered by the v3.5 production if the songwriting, music, and composition are high quality.

@@CB-AI I have songs up that I can clearly hear shimmer but most of my friends or family and even random people won't hear it. It's quite alarming how people can not hear it in my opinion. That is how I spot AI songs on spotify now, the "shimmer/lasering effect" is so distinct to AI generated instruments. Even Udio which is objectively higher sound quality even struggles with this. I will say V4 of SUNO did massive leaps in sound quality. I can not wait till full blown AI DAW's arrive, it will shatter and rewrite the music scene forever, and people who never had the possibility to make music will now be able to do that.

Hey Michael, thanks for the compliment! If you want to troubleshoot replacing the word using only Suno features, I have a video on that that might help (check the description for the chapters to help save time): kzbin.info/www/bejne/l2ScZGNtibyppaM Or, you might be able to use the Get Stems feature in Suno, use a DAW/Audacity to remove the word from the vocal track only, then combine the tracks in the DAW/Audacity. I haven't tried this myself, though, just an idea.

@@CB-AI Yeah i have not tried stems either because once I take it off of suno I cant put it back on... I just like writing lyrics so for now this site works. Easy to share the "Videos" with the lyrics or get ideas when I was stuck. I will check out that video for sure, Thank you!

You might have more luck trying out external tools to get the stems. I haven't tried these myself but there's a discussion around it on this Reddit page: www.reddit.com/r/SunoAI/comments/1eaqwn9/you_can_now_get_separate_stems_instruments_vocals/

Dude just do it.

Personally I like the EQ of the voice in the 2nd version more; it sounds clearer and more up front in the mix. I wonder if we are referring to the same thing here regarding the 8k frequencies... I remember trying to fix the v4 quality issues in Audacity and I was trying to reduce certain frequencies in the mix. Can't remember off the top of my head what the range was. And I couldn't find a way to reduce the frequencies without a significant negative impact to the overall sound; it made it worse than keeping it as is.

@@CB-AI I disagree whole heartedly. Maybe you need a new set of headphones. I don't hear it as much in just desktop speakers, but in headphones it is GROSSLY obvious. And...there is something weird with background vocals too...mostly choir. But clearly, they did something to the EQ in V4 that has lessened the quality QUITE a bit.

Hey man! I am looking into it and will let you know if I find something that's a fix. My guess is there's probably a different software other than Suno that does a better job of splitting the stems out. I was hoping it was possible to use the Cover feature on the Instrumental stem, but it doesn't give it as an option in Suno.

There's a discussion about this on Reddit and some people have mentioned different tools they use instead of the Get Stems feature in Suno. I hope this helps! www.reddit.com/r/SunoAI/comments/1eaqwn9/you_can_now_get_separate_stems_instruments_vocals/

@@CB-AI awesome, thank you. Im looking into👍

My pleasure; thanks for watching! 📺

Hi @SimpleTechAI, my pleasure. Thank you for watching!

Hi Ivan, the difference is easier to hear in these Suno links to the songs. I think KZbin and/or my production software compresses the sound so it's hard to hear the difference in the video. Hoping this helps. EXAMPLE 1 (BAD SHIMMER): suno.com/song/7f001a83-6304-4b23-a32a-52fad8c27521 EXAMPLE 2 (REDUCED SHIMMER): suno.com/song/e669a7d6-4f65-46c6-b524-15b01bf4e802

@@PrashayaShrestha Thanks Prashaya 🙋‍♂️ What genres are you into these days?

@@Mikemuchboy Mike's spittin rhymes ta pass the time well spent, so thanks for the comment, Keep it real in 2025 and letcha Suno flow vibe for ya tribe that's heaven sent

Hi Steve, what types of topics would you like me to cover?

@ i want See what Effekts with * functionally, Suno Wiki Write about this.

Hi Kiko, you might be able to fix it by using the Extend feature. Go to the track >> ⋮ >> Create >> Extend. Then Extend at a timestamp you like from before the "shhhhh" sound takes over. Copy-paste over any lyrics and tags that happen after the timestamp, and keep your style input the same as the original song. That should Extend it for you without the "shhh" effect and you can try a few times to get the extension you like most. Then, go to the extension track and click ⋮ >> Create >> Get Whole Song.

Interesting, yes possibly it's some EQ/reverb problem. I'm less experienced with the audio production and engineering side of things and haven't figured out a fix yet using a DAW (just using Audacity at the moment). Have you found any tricks to reduce it in a DAW?

How tf did they get reverb this wrong 😂😂 Nobody would ever desire this effect in a professional track

I agree with you, this is a reverb that is set to max peak and when it keeps building multiple times in a row it gets worse. It can be fixed with insanely expensive DAW plugins. Not worth it when Suno will fix it again with the next push. I've tried to apply prompts for "no reverb" in Suno sometimes it works and it reduces the shimmer drastically but also makes your track flat if you are going for a ethereal feel, you can potentially get the stems and apply your own reverb in a DAW.

@@CB-AI If you really want to get rid of shimmers investing into iZotope RX 11 is almost a must, it's studio grade audio editor with every denoiser you can imagine. But it's 439 Euros for the base version. You could probably do something in DAWs but nothing on the levels you can with RX 11. Better to just wait for suno to mask it better again like they did in 3.5

Try using the Extend feature instead. It's better than Replace Section for changing the end of the song. Go to the "⋮" next to the song you want to change >> Create >> Extend.

@@CB-AI thanks

I believe this does not work on free account

Hi Steve, you can hear the difference more in the Suno links. I think KZbin and/or my production software compresses the sound so it's kind of hard to hear the difference (sigh). EXAMPLE 1 (BAD SHIMMER): suno.com/song/7f001a83-6304-4b23-a32a-52fad8c27521 EXAMPLE 2 (REDUCED SHIMMER): suno.com/song/e669a7d6-4f65-46c6-b524-15b01bf4e802

​@@CB-AI i think songs with guitar would be a better example. it's super noticable with guitar honestly lol

@@raysyndrome7209 Hey Ray, thanks for the suggestion. You are definitely right about the guitars.

Possibly. According my understanding, it's normal for tech startups to focus on releasing a Minimum Viable Product (MVP) that works enough to use, and then they tweak and improve it based on their user's feedback, because they don't want to waste time/resources building features that customers won't use. They say v4 is in Beta mode, but IMHO they probably over-hyped it and made it look like it was a non-Beta feature. I did find a better way to reduce the shimmer that lets you keep the song in v4, I made a part 2 video for it here: kzbin.info/www/bejne/e3aQiWp9nd2BfdEsi=3NRw7WrPjJbS7ehN This video at 10:38 talks about the "releasing a crappy product to help certain customers" thing that tech startups do: kzbin.info/www/bejne/h4O9kJ9tmLutsLcsi=pXw5ijeqh9JpemFd&t=638

That's great to hear. Glad it finally produced a good ending that completed your song.

Nice! Do you think it works as well as the Extend feature? That's the one I usually use for Outros.

Yes, good tip. I do that too with the lyrics when I want the voice to pronounce the words a certain way.

@@TUNEScorner-k7o My pleasure! Thank you for watching!

Thank you! I threw a subscribe your way. If you have any challenges come up in Suno let me know!

@@CB-AI Thanks! Got you in my abo's too!

Thank you!

Exciting? or absolutely worthless )

@@Lyudtaru Hey there! I have another video on troubleshooting Replace Section common problems that might help. It's pretty long but if you check the Description there are timestamps with descriptions so you can go to the section that applies to you. kzbin.info/www/bejne/l2ScZGNtibyppaMsi=micquMQtmbI8cIY4 I hope this helps!

My pleasure; thanks for watching!

Since v4 came out, I've noticed the shimmer a little in v3.5, although it's so much less unpronounced I didn't notice it 'til comparing with v4. Thanks for the tip about spacing the lyrics. Do you mean it helps to literally add more space (like more space in the text prompt), or are you referring to spacing the lyrics by adding instrumental breaks in between?

@@TheRazeluXE You might be able to get rid of it by using the Extend feature instead of Replace Lyrics

@CB-AI no I mean every songs I generate has their vocals gradually decays towards the end. Usually pass 2 minutes mark that the problem start arise.

@@TheRazeluXE Hmmm, that's interesting. I haven't noticed that with my songs but they are mostly in English. Something I've found that can make the sound quality better is to create the song, then use Cover and keep the song the exact same, and keep the Style prompt the same or change it slightly (ex. changing RnB to 2000s RnB, 2010s RnB, etc.). Sometimes the Cover songs have better vocals than the original

Try using Remaster, as well. That's another trick that sometimes improves the quality

Hi Helen, is the extra word at the beginning, middle or end of the section?

@@CB-AIThanks so much for the reply! There is one duplicate word in the last line of the very first section and another in the second line of the very last section of the song. So, I can’t correct them in one try. I was able to remove the word from the first section and saved the song. But once I try to change the last section, it completely changes the ending. It’s quite frustrating, as I’ve used up many points trying to correct it.

@@helenb.5337 Sorry to hear it's not working well on the ending section! You could experiment with these tips on the last section (it will use up more credits, though): - Make sure when you select the section that it's only 10 seconds long. I assume you're already doing this but wanted to share in case not. - Try choosing the 10 seconds "in front of" the word, include the word, and keep all the lyrics the same except that last word. The singer will sing the line a little differently but hopefully not enough that it's no longer to your liking. You might even try adding a "." to end the phrase where you want to suggest to the AI that the line stops at the period. - Try choosing the 10 seconds "behind" the word, starting with the last 2-3 words and have the section go from there - Instead of using the Replace Section feature, try using the Extend feature, choosing a timestamp roughly at the last lyrical line. Include the corrected final lyrical line, as well as tags that describe as close as possible what you want the ending to sound like. (Ex. [Gentle Instrumental Break] [Fadeout] [End], etc.). This, in theory, would help guide the AI towards re-doing the end close to or the same as the way it is in the original. The above steps are what I would try if it was me. 🙂 Hopefully you'll find a way soon without having to use many more credits. Let me know if it works or doesn't work, I'd like to hear how it goes! -Cam

@@CB-AI​​⁠Thank you so much, Cam, for such a detailed response. I am a total newbie to Suno, so it’s trial and error and the credits get eaten up aggressively.)) I will definitely try your suggestions. Thank you again.🙏🏻

It can be pretty tough to get stuff to come out right with it, for sure. What issues have you been experiencing with it?

Hi Jake, my pleasure. Thank you, as well, for the tip about Remaster. I tried it on a few tracks and didn't like the results, but I didn't try many times on the same one. It's good to know it will work after enough tries.

I've also tried remaster, but only once. And because it was not helpful I did not try more than once. But next time I will....thanks to both for the topic!

Sorry to hear that. What was the original lyrics line and what was the replaced one?

@@CB-AI It's in Polish so I doubt you would understand. I tried to edit the vocal audio stem - but it does not highlight the lyrics and when I hit the "recreate section" it translates the lyrics to russian langue :/

Bummer. A lot of people are having problems with it, so you are not alone! I have an idea you could try but I've only tried in English, so I'm not sure if it would work. For the Polish lyrics instead of putting it how it's spelled, try writing it out phonetically in English. Ex: Instead of "Karol Wojtyła" put "KA-dole VOY-tee-wah." OR try changing the spelling and capitalization to "push" it in the direction you want it to go. It's worth a try.

@@SteveM45Thank you, greetIngs to you as well!

Hi Steve, thank you for the feedback. I agree that there can be a reduction in production quality when going to to v3.5. I've been testing both out quite a bit and think they both have pro's and cons. I like v3.5 for getting the production sound from a particular era whereas v4 has a more modern, clear distinction between the instruments and vocals, and sounds more "up front" in the mix. But the shimmer effect is so bad sometimes I switch to the more flat sounding output of v3.5. It also seems to depend on the genre which version works better. Hopefully Suno will patch an update for v4 soon. I've tried putting the shimmery tracks in Audacity to try to fix it there, but haven't had any luck yet. (Could be because I just lack the audio engineering knowledge, though. I haven't found any audio production videos yet that talk about this specifically)

@@CB-AI I've also tried filtering out the shimmer in Audacity, but the lowpass filters aren't very sharp. The frequency of the shimmer is about the same as fast drum tracks, so you'd lose both. The folks at Suno, I hope, understand the engineering clues that a) the spread spectrum of the shimmer is always the same, and b) the frequency is always the same regardless of the tempo of the song. I've sent in complaints to them in a steady stream, but I presume they are working hard to figure it out (thus it is still labeled "beta").

​@@stevewalton3151 Hopefully they'll release the fix for it soon. I've noticed there's a lot of threads on other sites with people complaining about it.

Have you found that splitting the stems and recombining in a DAW reduces the problem? I've tried testing the mp3 and Wav files in Audacity to remove the effect but couldn't do it without a substantial negative impact on the EQ and mix quality.

@@CB-AI It just seems like an option that doesn't always work. I read it in a forum. Unfortunately it is not a reliable method. I hope SUNO solves the problem.

The shimmer effect is a kind of background noise.

I believe it's a feedback effect. I don't know the mechanics of their algorithms, but somewhere it's picking up some feed-forward noise. The shimmer is always the same frequency, a hallmark of resonance modes. I really hope they get it fixed soon, because I'm having real trouble getting clean versions that minimize it.

It's that slotmachine/phone ring/laser/whatever you wanna call it in the backgrounds of any song made in v4

Hi there, when you try to regenerate the lyrics, does it change them in the recording, or do they stay the same?

Hello, thank you for your reply. I followed your instructions, and I found that the English lyrics can be modified, but the Chinese lyrics cannot be modified. I don't know why. I look forward to your help. Thank you again! You are great 🎉

@@victoryu8653 My pleasure 🙂 Is the whole song in Chinese, or is it a mix of Chinese and English? One thing I've found helps is when you create a new song, make sure that all the lyrics are there, even if it is repeated (like if the Chorus repeats 2 or 3 times), that way the AI can "tag" each word and so it will show up in Replace Lyrics when you highlight them. For an existing song that won't show up on the left, I wonder if the following steps would work (I am not sure if this will work but you could give it a try and let me know): - Go to your song you want to replace lyrics in - Use the Cover Song feature - Put in ALL the lyrics as they sound on the original recording, even if there are lyrics in the recording that the AI added (like an extra chorus) - keep the Style prompt the same as the original recording - Generate the Cover song - Pick one of the Covers that sounds closest to the original - Try Replace Lyrics in the Cover song and see if the section you want to replace shows up on the left now and if you are able to replace them

Hello, thank you very much for your patience and detailed reply! My lyrics are in full Chinese, and then I tried to use the full English lyrics to partially replace the lyrics. This song works, but for the Chinese lyrics, this operation is unsuccessful, and the lyrics to be modified are not displayed on the left. Why is this?

​@@victoryu8653Hmmm, I'm not sure why that is. If I'm understanding correctly, there are 2 songs: 1 in English and 1 in Chinese. The English lyrics can be replaced, but the Chinese ones cannot. Have you tried replacing lyrics in any other full Chinese songs? Did it work for those or did all the Chinese songs not work?

@@RookySteed I agree that it's a limited tool. The minimum time you can replace is 10 seconds I think. Hopefully they'll improve it soon.

@@unLeiShed Thank you!

@@Adam_Kobyshev "With Personas, you can save the essence of a song-its vocals, style, etc., and recall it for new songs whenever you need it!" It saves the vibe/essence of the song. It does not create the same song as the original. help.suno.com/en/articles/3484161

You might be able to use Extend instead, although the song may end up changing a little. It'd be like this: >> Go to your song in the Library >> Extend button >> Choose the timestamp for where the changes begin >> add all your updated lyrics, plus all the unchanged lyrics that follow after >> Extend a few times and see if any Extended versions come out that you like Let me know: A) if it works, or B) if not, did it get closer to what you were going for even if it wasn't an exact fit?