I tried the Weights voice training model, and paid for 1 month subscription to get their ‘better quality’ models but these free ones from Replay sound way better! Thank you for making this Bob. I asked you to make a video about this very Replay feature a week or two ago, and you delivered! Subscribed! ⭐️⭐️⭐️⭐️⭐️
I had a cunning plan: I started the update near the beginning of the video... "Replay" was already by the end... thanks Udio for the slow generation time!
Can someone tell me how I can choose the trained model. Or where to find the files. I trained a model in Replay, it said task completed but the model doesn't show up in the list of models.
When i attached my own voice : 2x error, tried to capture my voice, 30 min waiting and did nothing.. again tried with attaching and same , any soulition ? :D always get this error message : 'ModerData' object has no attribute 'mdx_dim_f_set'
What plug-in allows a dj to sing live with ai upgrade to wow voice of my awful voice? Similiar a bit but wow like autotune live but better not tune sounding. But natural like audimee and me mixed but live broadcast?
the last seconds of the video scared me 😝🤣, and I decided not to mess with you and leave a subscription, thanks for the video, I was looking for something like that,
Before Replay updates, can't you go into the app to the models folder and save the models to a backup folder so they don't get deleted? Then, after the update, copy them back to the models folder and also copy the new ones to the backup folder.
Gave up trying to download. I’ve tried to download on both Pc and Mac (both older computers) and can’t get it downloaded. So bummed. Thanks for the video. Great stuff
Thanks so much Bob for your great videos and information. Is there any new AI that is similar to Replay where we can use our own voice for songs? I've had lots of issues with Replay. In fact, it has never worked and kept loading etc.... any other good AI App having the same features? thank you!!!
Hi Bob, Thanks for your great video, I got this error every time I try to use this tool, "Error creating song: CUDA error: the launch timed out and was terminated Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.", Could you please let me know how to fix it and what is the issue? Thank you in advance
@BobDoyleMedia Great video Two questions: - does Replay do longer audios with music in the background - so like a 20-30 or even a 60 minute file? - what other softwares do this that are free and are paid for? thank you for your content
The problem I've had so far, at least with suno tracks, is that the replay slurs the words and sounds drunk. It's been a minute since I've tried it. But I would love to have a reliable model of my voice for music
hello Bob, i recorded a 17 min of me on the mic rapping and talking expoprted as a wav. When i try to start train model, unfortunately a error comes up like this.... Error; Request failed with Status code 422.. what am doing wrong?
@@KANIVERSETV I wish I had an answer for you. I don’t really know what the code numbers mean because I’m not really involved with replay at that level. Have you tried reaching out to them? They have a Discord.
Everything in your voice in the " narration" file was just you talking correct, but you were still able to make a good quality vocal. Sparks my interest wonder if I use a CPU what kind of quality it would return? What did you read in your narration file?
Thank you so much for this video. I’m in the process of finishing my song I’m about to record it but I can’t sing lol so I’ll give AI try to sound like me to sing the song again thank you
I got my song written and everything is beautiful except for AI when it sounds like me sings a little high in the course and it gives you the AI feel when I want to sound like me not sure why
@@BobDoyleMedia Done! I used the female choir from this song th-cam.com/video/defVOgHwxvU/w-d-xo.html. I isolated that and pre stemed and un-reverb etc. It worked way better than I was expecting! A little ugly sometimes and I LOVE that! :D
@@Saltovka_Scientist if you look into anything about how this is done, you’ll know that it’s not being harvested. Everything is running locally and that voice data isn’t going anywhere unless you put it there.
at the day today, they have sent through discord this message: We removed free training because it was too expensive. If we kept free training, weights would go, is i true?
@@hugotom9591 It looks like this only pertains to the training at weights.gg - NOT training on your own machine within Replay, which is using RVC locally and would not use any of their resources. This seems to only be if you're using their servers.
ISSUE SOLVED - TY!!! [ solved issue] after installing is impossible to download trained models 'cause it always comes with an error and an endless downloading process. Thank you for the try. I'm going to uninstall it [Solution] sounds weird but needed to download it three times until it works and guess what? It's working now!!!
@@BobDoyleMedia after three aptempts its now working... for three times it failed to download the model but now it's all downloaded. It started the server and I am now starting to train my first model. One question: is it possible to make Replay run in GPU?
@@BobDoyleMedia do you have a team of collaborators? I'm experienced in ML (prototype and training) working with algorithms in Python and C++! I'm testing more your software now, creating my first model.
That has not been my experience. Of course we may have very different graphic cards. But on my 3090. The model trained pretty quickly. Way less than an hour.
If I create a model with my best attempt at a singing voice, will the voice generations sing in tune? and if so how long should my singing voice sample be to create the model?
Have you heard whether people get silenced or shadow-banned on Soundcloud for posting Udio generated songs? After having published more than 70 songs since Udio was launched and reaching 31K plays in a month, all of a sudden, my insight stats for plays fallen to zero in a day!
I have not personally heard about that happening, but I guess I wouldn't be surprised. Probably worth looking into. That would suck if they did that without any kind of warning or heads up, if that's the case.
@@BobDoyleMedia I am not so sure anymore about this. Apparently during the day they reported zero plays, but as we moved to today at midnight there appeared quite a few plays for the previous day, above 100 plays... whereas again for today they are showing zero plays for the time being. In other words, does this mean that they only show individual track plays during the day and only show totals the following day? Strange anyways...IMHO
I feel so too , since the latest update ! The quality of the generated songs seems ot as good as more as before. And I am not talking about audio wise , more melody and auto generated lyrics seem quality down. Maybe they changed the something in the algorithm!
I'm not sure I'll be able to help, but at what point are you getting that? Perhaps I can reach out to my contacts and get an answer, but I need some context. Thanks!
@@Hasmatix i’m sorry that’s your experience, but that doesn’t inherently mean that Replay sucks. I have great success with it. But these AI programs can run very differently on different systems.
It happens a lot when fiddling with AI for the right stuff. You just wish you could beat him hard. In this video i had to use a different app for a statue part. th-cam.com/video/I_7Rhh9l0l0/w-d-xo.html
LOL - seems like you and your regulars aren't into big downloads eg, gaming - try waiting for a full download of Fallout 76 or similar, or maybe the Samplitude player with it's gazillions of GB's of sampled sounds.
Hard No! You flushed it down the toilet but recommended it? I don't know, I think you told the truth -- it sucks but you have a high tolerance to torture.
@@BobDoyleMedia You implied that it will make you pull your hair out. Many of the comments say they ran into serious problems. I believe you, when it works it's good but it doesn't always work. Maybe it will work out its bugs.
@@hannahbackward oh, right. Yes, it can be frustrating for a lot of people. A lot of people just give up. All this AI stuff is just so new and there are so many configurations of computers that it’s going to be hard to get something to work for absolutely everyone for a while. many of these things come from small teams that don’t have the resources of huge companies so we just have to deal with the growing pain if we want to play with the toys.
Maybe it’s just me but suno makes far better songs for me. I’ve tried and tried with udio but I always get a song that just doesn’t feel anything close to what I’m looking for. And it takes a long time to generate. I keep trying udio because they have better voice quality but there song structure and overall flow is much worse than suno. In the end I always end up going back to suno for making full length songs.
I tried the Weights voice training model, and paid for 1 month subscription to get their ‘better quality’ models but these free ones from Replay sound way better!
Thank you for making this Bob. I asked you to make a video about this very Replay feature a week or two ago, and you delivered!
Subscribed!
⭐️⭐️⭐️⭐️⭐️
Funny because weights and Replay are the same people, but I had the same experience!
Your videos never get boring!!! Thanks
Man you aren’t kidding. Downloading replay can be frustrating
I learned a lot over the last 2 days from your channel, great thanks. Subscribed!!!
Glad to hear it, and thanks so much!
x2 Thanks a lot!!
I had a cunning plan: I started the update near the beginning of the video... "Replay" was already by the end... thanks Udio for the slow generation time!
Hello! Thank you for the video!
Why I don’t have a “Train a model” button on top?
Hi Have you tried the new version 6.2.1 Replay it failed in the javascript
Hi Bob,I've downloaded the latest version of replay already but it doesn't seem to have the "train a model" button beside the " create new AI cover"
Can someone tell me how I can choose the trained model. Or where to find the files. I trained a model in Replay, it said task completed but the model doesn't show up in the list of models.
How can i find dutch speaking or singing models? super videos Bob!
Awesome tutorial! Thanks for showing people how to use it :D
When i attached my own voice : 2x error, tried to capture my voice, 30 min waiting and did nothing.. again tried with attaching and same , any soulition ? :D always get this error message : 'ModerData' object has no attribute 'mdx_dim_f_set'
Me too
Thank you for videos. Which model will be better and more realistic? Created with RVC or in Replay?
7:26 how is it working for you? I get errno 2 no such file or directory
same here...........any lucky so far
Udio seems to be limited on genres compared to Sudo. I just wish Sudo would upgrade their mp3 quality from 192 kbps to 320 kbps like Udio.
Yip, completely agree. Tried Udio for a month and the genres are dated and bland. I went right back to Suno
What genres that udio cant d0?
@@1brooklyn718 mostly modern genres.
Tried nth times.. just does complete the download… stops at the fag end at 3.5 gb. Should i wait looooonger or keep restarting 😢
Where u get the rap videos you’re showing? How? Scrape TH-cam or? How you get those singers who look like ai?
What plug-in allows a dj to sing live with ai upgrade to wow voice of my awful voice? Similiar a bit but wow like autotune live but better not tune sounding. But natural like audimee and me mixed but live broadcast?
the last seconds of the video scared me 😝🤣, and I decided not to mess with you and leave a subscription, thanks for the video, I was looking for something like that,
Before Replay updates, can't you go into the app to the models folder and save the models to a backup folder so they don't get deleted? Then, after the update, copy them back to the models folder and also copy the new ones to the backup folder.
You R the "Digital A.I. Audio Whisperer"... You RoCk.
Thanks so much.
@@BobDoyleMedia Don't thank someone for kissing your ass. Disgusting.
I wish Udio would generate the instrumental separated from the vocals
Udio have that option, you are able to create instrumentals
@@Elchengo I want the instrumentals from the full music that was generated with the voice so I can record my own vocals following the original.
@@nightmisterioBandlab has free ai track splitter. Works decent for Udio songs.
@@nightmisterio u can use ultimatevocalremover, to cut off the vocal part. today is a one way of it
Error message on Ventura 13.5 : Error creating model: cannot cache function '...
you need to update to the latest version of mac
Gave up trying to download. I’ve tried to download on both Pc and Mac (both older computers) and can’t get it downloaded. So bummed. Thanks for the video. Great stuff
When I try to train a model it happens after a minute
Error creating model : 'ModelData' object has no attribute 'mdx_dim_f_set'
Same here 😢
this is amazing it’s working for me!
I had been trying, but would always end up in error. something related to 'unable to find engine', any one knows how to overcome this?
I have the same issue it say error when I want to train my model ☹️
Have you experimented with pre-stemmed vs stemmed vocals? If so, is there a big difference in quality?
For me it says no error: found files in gt.wave_dir replay" when i try to create from a wave file
Thanks so much Bob for your great videos and information. Is there any new AI that is similar to Replay where we can use our own voice for songs? I've had lots of issues with Replay. In fact, it has never worked and kept loading etc.... any other good AI App having the same features? thank you!!!
Hi Bob, Thanks for your great video, I got this error every time I try to use this tool, "Error creating song: CUDA error: the launch timed out and was terminated Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.", Could you please let me know how to fix it and what is the issue? Thank you in advance
it stucks at epoch 1 when i train 😢
same, any luck yet?
can you share A TUTORIAL about How to create a clone duet voice using Replay AI
i tried replay, 4 hours later after trying to train a model,,,, it just keep turning the gears. Meanwhile using all my CPU. 1star product
Can this program convert audio to a musical instrument? Is there a program that does this?
@BobDoyleMedia Great video
Two questions:
- does Replay do longer audios with music in the background - so like a 20-30 or even a 60 minute file?
- what other softwares do this that are free and are paid for?
thank you for your content
Download is not frustrating. Training it is.
Would be awsome to have these in pinokio to have local but not take up much space
Not one time I saw did he upload the audio track of his voice into udio....
Did you use a particular program for the 17-minute recording that you uploaded for the model?
I used Adobe Audition. But truly anything would work as long as it records high quality audio.
The problem I've had so far, at least with suno tracks, is that the replay slurs the words and sounds drunk. It's been a minute since I've tried it. But I would love to have a reliable model of my voice for music
I have no gpu and when I try to train my model using cpu or mpu it is showing error. please help me troubleshoot.
Bro good job.but need effects for train model voice.what to do for effects for vocals?
I am subscribed
please could you compare this to SoundId VoiceAI?
Bob you da bomb bro! Thanks again!
hello Bob, i recorded a 17 min of me on the mic rapping and talking expoprted as a wav.
When i try to start train model, unfortunately a error comes up like this.... Error; Request failed with Status code 422..
what am doing wrong?
@@KANIVERSETV I wish I had an answer for you. I don’t really know what the code numbers mean because I’m not really involved with replay at that level. Have you tried reaching out to them? They have a Discord.
@@BobDoyleMedia Thanks Bob, ill try
Everything in your voice in the " narration" file was just you talking correct, but you were still able to make a good quality vocal. Sparks my interest wonder if I use a CPU what kind of quality it would return? What did you read in your narration file?
What is a "jiff"
Wish Replay was available on IOS, great Liam outro!
replay ai vs Audimee Which one you chose excpt the Price
What’s the best app or program to generate lyrics to songs with your own ai tuned voice
ISSUE: Replay is not detecting my GPU card (it's AMD Radeon 220 under Windows 10 Pro). Can someone help me, please?
Sorry, but only NVIDIA GPUs are supported.
Thank you so much for this video. I’m in the process of finishing my song I’m about to record it but I can’t sing lol so I’ll give AI try to sound like me to sing the song again thank you
I got my song written and everything is beautiful except for AI when it sounds like me sings a little high in the course and it gives you the AI feel when I want to sound like me not sure why
I can't use Replay.Is giving error all the time and I can't enter to Discord to see which is the problem
@@elmusicantedormido I’m still waiting for a resolution myself!
i need how to instal and use it in mobil please bob
Thanks much appreciated, Pretty awesome, Subscribed
I was fine with the installation process, but the training is taking FOREVER!! and it's not using my GPU Idk why...
hope it works with a choir, i'm gonna test it
Not sure how that would work, but I'd love to know if it does!
@@BobDoyleMedia Done! I used the female choir from this song th-cam.com/video/defVOgHwxvU/w-d-xo.html. I isolated that and pre stemed and un-reverb etc. It worked way better than I was expecting! A little ugly sometimes and I LOVE that! :D
Does it harvest voices of users? I don't believe that it doesn't
@@Saltovka_Scientist if you look into anything about how this is done, you’ll know that it’s not being harvested. Everything is running locally and that voice data isn’t going anywhere unless you put it there.
Thank you so much for incredible and very helpfull video.
Bob thank u friend u make my day 💓 keep up the realness
at the day today, they have sent through discord this message:
We removed free training because it was too expensive. If we kept free training, weights would go, is i true?
@@hugotom9591 It looks like this only pertains to the training at weights.gg - NOT training on your own machine within Replay, which is using RVC locally and would not use any of their resources. This seems to only be if you're using their servers.
Yay always bringing the best lit read my mind 🎉 n I love ur topics lol
ISSUE SOLVED - TY!!! [ solved issue] after installing is impossible to download trained models 'cause it always comes with an error and an endless downloading process. Thank you for the try. I'm going to uninstall it [Solution] sounds weird but needed to download it three times until it works and guess what? It's working now!!!
Can you be more specific with the error messenger getting?
@@BobDoyleMedia after three aptempts its now working... for three times it failed to download the model but now it's all downloaded. It started the server and I am now starting to train my first model. One question: is it possible to make Replay run in GPU?
@@BobDoyleMedia do you have a team of collaborators? I'm experienced in ML (prototype and training) working with algorithms in Python and C++! I'm testing more your software now, creating my first model.
Having no valid GPU, my model took 24 hours to complete. No biggie I was either sleeping or at work.
That's annoyingly insane, though....frustrating. Do you have a Mac?
CPU?
this doesnt work i keep getting an error saying there's no path whenever i try to record my own voice
@@immortalhybrid7570 Have you reached out to their support?
the latest update gives errors, and clone your own voice takes hours and hours, so i went back to the first version
That has not been my experience. Of course we may have very different graphic cards. But on my 3090. The model trained pretty quickly. Way less than an hour.
If I create a model with my best attempt at a singing voice, will the voice generations sing in tune?
and if so how long should my singing voice sample be to create the model?
Have you heard whether people get silenced or shadow-banned on Soundcloud for posting Udio generated songs? After having published more than 70 songs since Udio was launched and reaching 31K plays in a month, all of a sudden, my insight stats for plays fallen to zero in a day!
I have not personally heard about that happening, but I guess I wouldn't be surprised. Probably worth looking into. That would suck if they did that without any kind of warning or heads up, if that's the case.
@@BobDoyleMedia I am not so sure anymore about this. Apparently during the day they reported zero plays, but as we moved to today at midnight there appeared quite a few plays for the previous day, above 100 plays... whereas again for today they are showing zero plays for the time being. In other words, does this mean that they only show individual track plays during the day and only show totals the following day? Strange anyways...IMHO
Is replay totally free?
Thanks Buddy' Great Job
Maybe it's just me but udio kinda went downhill lately. Too slow as well. Suno this month, udio next? Always choose the monthly plan.
Well, all I can say is that it was not being friendly with me for this video...
There's a slider on advanced setting. You can adjust the generation time on udio. Better result will be longer to generate.
I feel so too , since the latest update ! The quality of the generated songs seems ot as good as more as before. And I am not talking about audio wise , more melody and auto generated lyrics seem quality down. Maybe they changed the something in the algorithm!
THank u for this
Try “Jazz inspired 1950’s rock-n-roll” which gave me Frank and big band
Thank you! I will definitely do that.
Hi Bob how are you ?
I get error Code 422 you have any idea why ?
I'm not sure I'll be able to help, but at what point are you getting that? Perhaps I can reach out to my contacts and get an answer, but I need some context. Thanks!
@@BobDoyleMedia HI Bob Thank you so much, wen im trying to create a model I upload the audio file wen I hit create , it comes out with that message
Error request failed with status code 422
Same here, did you sorted? 😢
Send my regards to NY sounded like Elmer Fudd lol
😄
Sir.. you are one professional F up.
The current version 6.5.3: The training part is gone!
It's back again with a new download I just did
And it's gone again
Cool!
6:27 15:26
put some damn noise supression on your mic
Training epoch took forever
The best way is to let the software download and do something else on the computer , if nothing watch TH-cam
Bro replay sucks. I tried it and it the voice keep glitching and sounding like a robot
@@Hasmatix i’m sorry that’s your experience, but that doesn’t inherently mean that Replay sucks. I have great success with it. But these AI programs can run very differently on different systems.
@@BobDoyleMedia is there a prompt in sudo to get the vocal without reverb ans effects i feel like that makes the whole process worse
In the future, you will be able to sing like Freddie Mercury.
I get an error :(
can someone pls help me process a song with my own voice? i want to surprise my future husband in our wedding.
What error are you getting exactly? What GPU are you using?
It happens a lot when fiddling with AI for the right stuff. You just wish you could beat him hard.
In this video i had to use a different app for a statue part.
th-cam.com/video/I_7Rhh9l0l0/w-d-xo.html
Yeah, AI is going to turn you into Supertramp......😂😂😂😂😂
LOL - seems like you and your regulars aren't into big downloads eg, gaming - try waiting for a full download of Fallout 76 or similar, or maybe the Samplitude player with it's gazillions of GB's of sampled sounds.
That’s how it was to download Band in a Box. I’m used it.
Hard No! You flushed it down the toilet but recommended it? I don't know, I think you told the truth -- it sucks but you have a high tolerance to torture.
What do you mean I flushed it down the toilet? I love Replay! Maybe I’m just not understanding your comment.
@@BobDoyleMedia You implied that it will make you pull your hair out. Many of the comments say they ran into serious problems. I believe you, when it works it's good but it doesn't always work. Maybe it will work out its bugs.
@@hannahbackward oh, right. Yes, it can be frustrating for a lot of people. A lot of people just give up. All this AI stuff is just so new and there are so many configurations of computers that it’s going to be hard to get something to work for absolutely everyone for a while. many of these things come from small teams that don’t have the resources of huge companies so we just have to deal with the growing pain if we want to play with the toys.
good info but torture hearing your musical selections lol
🤷🏻♂️
Maybe it’s just me but suno makes far better songs for me. I’ve tried and tried with udio but I always get a song that just doesn’t feel anything close to what I’m looking for. And it takes a long time to generate. I keep trying udio because they have better voice quality but there song structure and overall flow is much worse than suno. In the end I always end up going back to suno for making full length songs.
Better results with 30 second then extending before or after.
Seems like a whole lot of trouble for karaoke.
Thompson John Harris Scott Rodriguez Jennifer
what glasses. again again javascript error for everyone
Scammer
Sorry, I'm out as soon as you said rap, bye
Rap is only a small part of the video, it’s worth watching. He later tried to create a 40s crooner song.
Bye😂
🤣🤣🤣
Bye unc
Good we don’t need close minded people. It’s okay to not like genres but making someone “suffer” for it is a weird flex
replay download is stupid
They've definitely had issues with that for a while. It's actually been awhile since I've updated it. Not sure what it's current state is.
TOO MUCH BLAH BLAH BLAH, GOODBYE