Would you be interested in a style creation competition?

StylePlayer

Here is an idea I had for a long time, let me know if it's something you would be interested in.

What if we organize a friendly style creation competition? We pick a popular song (by voting), then during let's say 2 weeks anyone can submit their own styles for this specific song, and at the end we vote for the winner whose style sounds the best and is closest to the original.

Down the line we can even come up with a prize pool if you like.

Maybe at the end we could also review the submissions on our youtube channel so you get the exposure and bragging rights.

There are a few open questions that need to be figured out, like:

How would people who don't have Genos2 be able to evaluate those styles, should we ask for an accompanying audio recording sample? If yes, how can we make it so the recording only showcases the style quality and not playing skills? Perhaps have people play through a series of chords without the right hand?
Do we simply judge the styles overall or try to evaluate them in different categories (note accuracy, voice accuracy, mastering)
Would it be fair for Genos2 made styles to go against Tyros made styles. I wouldn't want to make artificial restrictions on voices to make Genos2 styles sound worse than they could be, but don't know if Genos2 would always win otherwise.
Do we judge the style alone or the recorded performance. Perhaps we can create a separate competition for the performances based on the submitted styles, this way people who are not into style creation but into playing can participate and show off their skills.
We can also create a parallel lead sheet+chords competition for people who are into sheet music.

So at the end we get a bunch of great styles, a bunch of sheet music to go along, and performance examples for reference. I think this would be epic and super fun.

I am currently busy with a few site features that I want to wrap up in the following weeks, but we can brainstorm this idea meanwhile and come up with a good set of rules. Also let's get a feel for how many people would be interested in participating in something like this.

Thank you.

FabiOh

Hi,

yes, I like the idea! I have a few thoughts about this: Should "is closest to the original" really be the benchmark? Sounds the best should be sufficient enough for rating, so that also alternative instrumentations come into play (no pun intended!) which may sound different to the original, but also good in a musical way.

I would say, also to address your first open question regarding Genos 2: What if we do it by Keyboard model(s) (e.g. SX600, S670, etc.; another group might be SX700/720/900/920, then GENOS 1&2 etc.), which serves as a kind of "league" so to say? -> So that each player plays in his natural habitat, which he is most familiar with and has no drawback due to lack of certain features in his keyboard when competing against each other. Later on, one can convert these styles between the different keyboards/leagues for sharing purposes, if one would like to share.

StylePlayer

FabiOh Thank you for the feedback!

I was thinking of using "closest to the original" criteria as it the only somewhat objective criteria, not a matter of a taste. Also it evaluates a style itself, not a performer. A virtuoso performer can take any style and do all the magic in right and left hand parts so it will sound the best, but it doesn't mean this style is any good, it is only good in its creator hands.

Sure it would be fun to listen to a bunch of great performances, but besides providing temporary entertainment it doesn't make any long lasting impact. I wanted to focus more on a style creation part for several reasons:

It will help popularize the style creation, as currently the majority of people never tried to create their own styles, this will give them a chance to try it.
We involve more people in the style creation process, who will produce more styles, and those styles in turn will attract more people to the arranger keyboards, so we create a positive feedback loop
A style quality can be judged independently from a performer, and if we use "closest to the original" criteria it can be somewhat objective evaluation where everyone can agree if it sounds close or not close
It will let people who don't have a performer gift to also participate
At the end we produce a bunch of quality styles that can be now enjoyed by everyone for years to come. A performer will just give a one time great performance and that's all, the community as a whole won't gain anything from it besides an aesthetic pleasure.

It's not that I am looking at it only from the material perspective, I would enjoy a good performance as much as anyone else. I just want to focus on something that makes a long lasting positive impact on as many people as possible, not just provide a temporary entertainment for selected few.

I know that listening to performances is probably more entertaining for most people and bystanders than analyzing how accurate a style is, so I think it would be best to separate a style creation part from the performance part, so we have two different parallel competitions with own winners. This way we have something for everyone.

Regarding own leagues for different keyboards. That would work great if we had hundreds of participants, I think would be great if for starters we got just 5 people, so far looks like the idea is not really resonating, I think arranger owners are just not creating their own styles, and that's what I want to ultimately change.

RudyVriendts

StylePlayer if you use the MIDI to Style App you can create very good styles that you could use, is that in your opinion allowed when you make a competition ?

StylePlayer

RudyVriendts Sure, I don't see why not. I don't think it would be that easy to create a quality style with it, or any other software.

Though this touches on another difficult question that I don't really know the answer to: What is a good style really?

You can take a midi file, turn it into a huge Intro A section and call it a day. Would it be a good style? If you want to ever play a single specific song with it then it's hard to find arguments why this is not the best possible style.

There is this fine balance between a style being too song specific and being too generic. It is especially unclear when a style was made specifically to play a single song, what kind of variations are expected to be supported by such style, if any? Is it possible to come up with some style requirements that would differentiate a style from a backing track? For example one such requirement could be that you are supposed to be able to loop a singe verse or chorus infinitely.

Also what about including the right hand portion, how much should you play and how much a style should play? You can make a style to play everything and your responsibility is just to press the "start" button (there are lots of "styles" like that on youtube).

How to define the requirements for a style that make it a (good) style?