jaflo

Yes, it is! I wanted a short domain for side projects, and then this one came across and I registered it. I tried to give an homage to the Pokemon in the favicon in the root domain :) https://loud.red/favicon-32x32.png

And yes, I had this issue while testing also! A lot of players added a small gap when looping which was a bit annoying to test with...

Glad you like it! Sadly this version is not, but this project is a rewrite from an old version I created that is open-source! https://github.com/jaflo/snipsnip

If you are interested in audio analysis in general, I would also recommend checking out librosa: https://librosa.org/doc/latest/index.html

Whoops, thank you, will update!

Fearing copyright lawyers, I implemented this message so you cannot share audio by just sending a link anymore, but have to import the song yourself to be able to use it :(

You can use this example though if you just want to try: https://mofi.loud.red/edit/8bd3fdf780f8c3927e41029f3b957f8a7...

This project brought together a lot of different pieces of engineering and was a larger personal project for me, so there were a lot of parts and making those work together was a challenge at times.

For the backend, I found it helpful to implement monitoring proactively to be able to detect and debug issues without having to bug users which has been really helpful so far and allowed me to make some performance improvements too!

For the frontend it was interesting to optimize the result searching using the Web Worker + WebAssembly approach. I found some interesting ways to speed up things by parallelizing and, interestingly, manually collecting garbage in the WASM runtime.

And of course learning about the entire audio analysis portion and reading on how to find similar sections of audio!

I think it could be possible (backend is in Python), but for most users a web service has a lower barrier of entry. Could you share your use case? I am curious.

I had forgotten about this check actually! I just removed it and deployed the changes. You can try again but am unsure if it will break something (it will spin indefinitely, but I would give it up to 10 minutes). Good luck!

I plan to implement paid features later with a basic free tier. I am still trying to figure that out. I've updated the terms to be more clear about that!

This is a good point, I have updated the site to only allow you to view content you submitted and also made it clear that you need to have the rights to the audio to edit it.

Whoops, looks like there was a bug, just fixed it, could you try again?

This project actually builds on the same analysis techniques as the Infinite/Eternal Jukebox! I really liked it when the original was up. I think they also had a version just for Gangnam Style along with a matching infinite video!

There is this version now, but it does not allow custom uploads anymore: https://eternalboxmirror.xyz/jukebox_index.html

The Echo Nest was really cool... they got bought my Spotify and some of the Echo Nest APIs made it into Spotify I believe: https://developer.spotify.com/documentation/web-api/referenc...

It definitely works better on pop songs or music with repeating choruses or patterns as those can be matched to each other more easily. What kind of music did you try it with?

There is no enforced maximum length for either input or output right now. In practice, the remote machine times out trying to analyze a song after 10 or so minutes. For output, your machine will take forever to generate really long results.

Looks like the machine ran out of disk space, I bumped it and it should work. Could you try again?

It does not create new music from scratch using one of the techniques you mentioned, but just reuses parts from the song, blending the transitions between them. That is a cool idea though!

Correct, this is it! It's basically a blend of screen recording and AE. The recording is made and then the software smoothes cursor movements and creates the zoom effects.

+1 on banana.dev, I used it for a side project and deployed some custom code and it was a good experience! I liked the pricing model (lack of minimums and pay for usage instead of a "plan") and how you can package up whatever code you want.

What’s wrong with text?

How else would you pay?

I really appreciate the “free up to a cap” and reasonable pricing. Makes me much more likely to try this for a side project unlike competitors who want me to commit to paying right off the bat.

One thing I noticed is that the link to themes (which I was hoping to see a demo at) is broken: https://supertokens.com/docs/emailpassword/common-customizat... under pricing

How are the prompts/captions generated? Also I love that you're using Svelte!

CapCut has a desktop version! I have not tried it myself so no idea how good it is: https://www.capcut.com/

Reminds me of the Bloomberg B-Unit authentication (https://www.bloomberg.com/professional/support/b-unit/, step 5) where it shows as flashing light to enroll a token generator for their terminal.

I am still trying to figure out the exact use cases for this project so it is a bit generic, but the features on Wanderlog look quite fleshed out! I have not used it myself, but it looks like it does not have the calendar view which I personally find helpful. I think that being able to use items from other places is really helpful though.

That sounds pretty neat. I am thinking maybe anonymous accounts that can then be upgraded to a proper account. And there are actually public view URLs you can set up! They can be enabled in the Share menu.

And yes, it is in a pretty early stage right now, but I should probably add ToS and a privacy policy to cover my bases.

I am interesting to hear more! I wasn't able to find your email address in your profile, but feel free to send me an email.

Those are really neat ideas! A prototype I had built at first had a map, but I didn't go with it because the cost to for location data searches and map renders was prohibitively expensive (I was considering Mapbox and their API got pretty pricey, Google's isn't much better and the remaining ones had licensing issues).

Thank you! I actually kind of cheated and didn't have to implement any of that myself because Firebase took care of it. But when I was considering how to build Calenday, those were on my mind. I had read about CRDTs on Figma's blog (https://www.figma.com/blog/how-figmas-multiplayer-technology...) and was considering using something like this to ease the burden on implementation: https://github.com/automerge/automerge