Beat Your Music is rather disappointing..

[Oh My Goth!]

One of the main reasons I decided to give this game a shot is that it allows you to use your own music files in a level. But I have to say, the feature as implimented is quite underwhelming. I'm not concerned with it only allowing .mp3 files (after all, it is the universally favored format for audio files), or that there is a 6 minute cutoff on song length--both are annoying, but not crippling, drawbacks.

However, I do have a few issues/suggestions that deal with the game mode as a whole, rather than its implementation:

- I was disappointed to see the game ask the user to input the tempo manually, because this tells me that it most likely doesn't have an algorithm to determine tempo itself. Also, given that there is only an overall BPM displayed after entering this information, it would indicate that only an overall BPM value is assigned. So what does that mean for songs with tempo changes (not exactly a rare occurrence in music)? I've tested several, and the results are not good, to say the least. I've seen everything from the enemies keeping the original speed to just gradually accelerating from the beginning tempo to the end tempo (which kind of debunks my 'one tempo overall' theory, but not in a good way).

- As far as I can tell, there is little-to-no response to the volume or intensity of the music, just the overall beat. A loud part with lots of fast notes often generates the same enemy formations as a quiet part with few notes of the same tempo. Audiosurf, and to a lesser extent Beat Hazard, have an algorithm that, while certainly not perfect, manages to do this reasonably well.

- It would be nice to have a level editor (like Soundodger+ has), or even a partial editor, that the computer then extrapolates from. When I do pick a mono-tempo song, the waves of enemies are pretty generic and don't really respond to the flow of the music, just the beat. This could be easily remedied by implimenting even a basic level creater.

[himmatsj]

1) Game already says use song with a constant beat. But yeah I agree manual input sucks.

2) I agree. My complaint as well. It's just flat.

3) I suggested as well.

Looks like we agree this BYM mode isn't worth it. Though the actual game itself is amazing. My thread on this subject matter: http://steamcommunity.com/app/255370/discussions/0/648817377880156755/

[Neil Sorens]

No beat detection algorithm is going to be accurate enough to use in a beat-matching game. Even the top engineers in the world working on nothing but that wouldn't be able to deliver a satisfactory result. Go ahead, try Traktor or any of the professional tools out there. And if it does give an accurate result, great! Now you can just type that into the appropriate place in Kickbeat without having to use the included tool. But most of the time it will not be anywhere near accurate enough.

With this implementation, it can take a little work, but the results are (or at least can be) better. Check out Dance Factory for a beat-matching game that lets you import music. Their algorithm is even patented. But all the reviews say it's not good enough. Plus, this way, once the numbers are correctly calculated once, they can be shared and manually input with no extra work.

Yes, songs with tempo changes won't work well, unless it's exactly halved or doubled.

There is definitely response to intensity of the music. It does some normalization/dyamic compression to avoid giant gaps in enemies, but you will see differences, especially at higher difficulty levels, which can have denser enemies and thus more contrast. You'll also notice that the slo-mos are usually associated with changes in the song (e.g., the "drop" in electronic music). Is there room for improvement? Definitely. Audiosurf and Beat Hazard are far less dependent on establishing/matching the beat, since they aren't beat-matching games, and thus have much more freedom in the scripts they can generate.

I agree that it would be nice to have a scripting tool that players can use, along with Workshop support. I think it just comes down to how much it costs to do that. Obviously, if the game sells well, it's a lot easier to justify adding it, whereas if people think the core gameplay is not that great or whatever, then building on top of that vs. putting those resources into making better products will make less sense.

[Oh My Goth!]

Originally posted by Neil Sorens:

There is definitely response to intensity of the music. It does some normalization/dyamic compression to avoid giant gaps in enemies, but you will see differences, especially at higher difficulty levels, which can have denser enemies and thus more contrast. You'll also notice that the slo-mos are usually associated with changes in the song (e.g., the "drop" in electronic music). Is there room for improvement? Definitely.

I didn't notice much of one, but I'll play around with some more songs with lots of contrasting bits and see what happens.

[himmatsj]

Neil I don't think there is response to intensity. I've played 20 tracks and I never felt it. At least not as much as within the main game. This was to me the biggest letdown of the mode.

PS: There are times when the enemies get intense during the chorus and matched the beat well, but I felt this was just a matter of luck and coincidence.

Hopefully the sequel delivers a full-on BYM mode.

[microlabs]

+1 to v0id19

[DeadAiM]

just played about 7 tracks , that keep the same'ish beat , worked great for me. slo-mo hits came out perfectly when the song dies down a bit / kicks in etc
its prob. your song choice tbh.
I've done a few that is high tempo at start , slows in middle , high tempo at end ... & aye , dosnt go to well.

[Drama[Theurgist]]

I would definitely buy a song editor if it comes as a DLC.

[Coffee]

Originally posted by Neil Sorens:

No beat detection algorithm is going to be accurate enough to use in a beat-matching game. Even the top engineers in the world working on nothing but that wouldn't be able to deliver a satisfactory result. Go ahead, try Traktor or any of the professional tools out there. And if it does give an accurate result, great! Now you can just type that into the appropriate place in Kickbeat without having to use the included tool. But most of the time it will not be anywhere near accurate enough.

With this implementation, it can take a little work, but the results are (or at least can be) better. Check out Dance Factory for a beat-matching game that lets you import music. Their algorithm is even patented. But all the reviews say it's not good enough. Plus, this way, once the numbers are correctly calculated once, they can be shared and manually input with no extra work.

Yes, songs with tempo changes won't work well, unless it's exactly halved or doubled.

There is definitely response to intensity of the music. It does some normalization/dyamic compression to avoid giant gaps in enemies, but you will see differences, especially at higher difficulty levels, which can have denser enemies and thus more contrast. You'll also notice that the slo-mos are usually associated with changes in the song (e.g., the "drop" in electronic music). Is there room for improvement? Definitely. Audiosurf and Beat Hazard are far less dependent on establishing/matching the beat, since they aren't beat-matching games, and thus have much more freedom in the scripts they can generate.

I agree that it would be nice to have a scripting tool that players can use, along with Workshop support. I think it just comes down to how much it costs to do that. Obviously, if the game sells well, it's a lot easier to justify adding it, whereas if people think the core gameplay is not that great or whatever, then building on top of that vs. putting those resources into making better products will make less sense.

I doubt its explain the 6mins limit on songs ... if it's because of that, just add more select the beat thing when a song ia longer.

[...]

Originally posted by Neil Sorens:

No beat detection algorithm is going to be accurate enough to use in a beat-matching game. Even the top engineers in the world working on nothing but that wouldn't be able to deliver a satisfactory result. Go ahead, try Traktor or any of the professional tools out there. And if it does give an accurate result, great! Now you can just type that into the appropriate place in Kickbeat without having to use the included tool. But most of the time it will not be anywhere near accurate enough.

With this implementation, it can take a little work, but the results are (or at least can be) better. Check out Dance Factory for a beat-matching game that lets you import music. Their algorithm is even patented. But all the reviews say it's not good enough. Plus, this way, once the numbers are correctly calculated once, they can be shared and manually input with no extra work.

Yes, songs with tempo changes won't work well, unless it's exactly halved or doubled.

There is definitely response to intensity of the music. It does some normalization/dyamic compression to avoid giant gaps in enemies, but you will see differences, especially at higher difficulty levels, which can have denser enemies and thus more contrast. You'll also notice that the slo-mos are usually associated with changes in the song (e.g., the "drop" in electronic music). Is there room for improvement? Definitely. Audiosurf and Beat Hazard are far less dependent on establishing/matching the beat, since they aren't beat-matching games, and thus have much more freedom in the scripts they can generate.

I agree that it would be nice to have a scripting tool that players can use, along with Workshop support. I think it just comes down to how much it costs to do that. Obviously, if the game sells well, it's a lot easier to justify adding it, whereas if people think the core gameplay is not that great or whatever, then building on top of that vs. putting those resources into making better products will make less sense.

Boy, I am so glad I found out about this before buying this game.

What about Audiosurf? It's not perfect, but at least it measures wavelength and amplitude, in DnB songs, the more intense sections in songs such as Showdown by Pendulum, it turns the track red and becomes a lot more fast paced, whereas in the calm areas of the song, the game matches it. Kickbeat could learn a lot from Audiosurf, and it only takes audiosurf a couple seconds to scan it. Audiosurf detects the beats using the data in the song, so it doesn't need to rely on things such as BPM or a static time signature, which is why Audiosurf also works with Dubstep tracks, or erratic time signatures and time signature changes. Again it's not perfect, but it sure as heck seems to work a lot better.

[Neil Sorens]

Originally posted by Ruruko:

What about Audiosurf? It's not perfect, but at least it measures wavelength and amplitude, in DnB songs, the more intense sections in songs such as Showdown by Pendulum, it turns the track red and becomes a lot more fast paced, whereas in the calm areas of the song, the game matches it. Kickbeat could learn a lot from Audiosurf, and it only takes audiosurf a couple seconds to scan it. Audiosurf detects the beats using the data in the song, so it doesn't need to rely on things such as BPM or a static time signature, which is why Audiosurf also works with Dubstep tracks, or erratic time signatures and time signature changes. Again it's not perfect, but it sure as heck seems to work a lot better.

It works in Audiosurf because Audiosurf is a not a beat-matching game. You don't press buttons to the beat of the song in Audiosurf, so it doesn't matter that Audiosurf can't accurately detect beats. In KickBeat, you do press buttons to the beat, so the beat detection has to be exact. It can't just be the "general mood of the song" stuff. It has to START with the beat and then build on that. And that's what KickBeat does. Check out the videos I've posted of BYM stuff. There are more enemies during the more intense parts. And it's perfectly on beat.

[2ndForm]

I think looking at the possibility of letting people make custom levels and uploading them via workshop is well worth a look. Could really turn this into a long running giant on Steam.

I get the copyright issues, but people dont have to upload the songs themselves. If they own the album, they can add the MP3 themselves, and if the one used to make the level is legit it will mesh fine.
Only point you would have CR issues is if people WERE uploading the songs.

Otherwise its just kind of assuming that people are criminals before they've committed the crime, and preventing pepole from growing your game for you.

If its got more to do with making the development of tools for players to make tracks(cost/time/this project has been done for a while), sure I can understand that.

Otherwise, I just really dont buy the whole copyright issue, unless a part of your contract with the licensees you currently have is to not allow this sort of thing to expose said tracks.
If thats the case, fair enough again, but I would be first in line to pick up DLC.

[Neil Sorens]

It's not a copyright problem. Although I don't work at Zen, I imagine that they're not seeing sales that would justify a major addition like a modding tool and full Workshop support. I'm actually surprised they brought the game to Steam at all, given its low sales elsewhere, and even more surprised that it's a pretty full-featured port, with extra content, various Steam features, design improvements, etc.

[Oh My Goth!]

So I've put a fair amount of thought into this since my original post, and I think the best solution is maybe not a full-fledged level editor (because that is a bit work intensive for the casual player), but more of a 'level setup" where the computer extrapolates from a few key inputs.

I get the whole beat matching thing prevents algorithms like what Audiosurf has from being viable, but I also feel like if the player were able to give a semi-in-depth model of a song's structure, it wouldn't be too hard to create something for the game to generate a coherent level from that information.

[Melodia]

Yeah I was going to get this game because of this feature, but it's not worth it. Especially given it's locked behind three hidious songs, it doesn't even allow the entire title of the tracks, and of course that it can't even use VBR Mp3s (according to another forum post)....to say nothing of FLAC support.
Even if all that were fixed, the fact that there's not even an attempt to match anything up means it's pretty useless.

[himmatsj]

Originally posted by Melodia:

Yeah I was going to get this game because of this feature, but it's not worth it. Especially given it's locked behind three hidious songs, it doesn't even allow the entire title of the tracks, and of course that it can't even use VBR Mp3s (according to another forum post)....to say nothing of FLAC support.
Even if all that were fixed, the fact that there's not even an attempt to match anything up means it's pretty useless.

When you listen to them like that, the songs may be "hideous" (especially if you're a mainstream kinda guy). But within the game itself, they work just fine. Actually, really well.