osu!stream 2020 release

published 26 Feb 2020

osu!stream was a project devised by me in 2012 to experiment with the idea of an osu! game on mobile platforms. The focus was on trying new things, so not only did I want to experiment with mobile game development, but also add elements using multi-touch. In addition, unlike the desktop release I wanted to have a go at curating all game content, including both licensing music and putting together a team of awesome mappers to create the beatmaps.

Before I delve into some history and tidbits on the project, if you’re just interested in giving it a shot, click here to visit the app store and grab yourself a copy!


osu!stream began development on May 8 2010 and was released on August 1 2011. During this time, I was concurrently working on the PC version of osu!, so osu!stream was somewhat a “side” project. Updates were released in the form of new features (such as the ability to preview maps before downloading) and song packs over the next 2 years or so.

It gained some attention over the internet, and was featured by Xamarin as a game made completely on the Xamarin.iOS framework (link now defunct, unfortunately).

The game did quite well from a sales perspective, but was hindered by the complexity of the beatmap creation process and limited song choice for mappers. Many of the mappers either lost interest or didn’t have enough spare time to keep a constant.. stream of beatmaps coming in.

At the same time, osu! on PC was constantly growing in size. It’s very hard to keep track of multiple projects, so I eventually made the decision to move my focus back completely to PC and let osu!stream sit without real updates for years. I don’t regret this decision; the game was left in a good state and has enough beatmaps to be a very fun and unique “osu!” experience. I still see it as the most polished osu! release to date – although this is soon to be surpassed.

Interestingly, a lot of the concepts we are using in osu!framework have some history of being present in osu!stream. The concept of hierarchical drawables/sprites began to appear, and while quite raw, was used to apply game-wide scale and rotation effects which would otherwise not have worked on limited mobile hardware.

Some statistics

There were a total of 3,839,438 scores submitted over osu!stream’s lifetime. This doesn’t include fails, unlike osu!, and only considers those that logged in via twitter (241,540 total). For a better idea of total audience, there were a total of over 1.5 million installations as of 2016. It’s safe to say a lot of people enjoyed the game!

2.6 million free beatmap packs were downloaded over this period. As for sales, there were a total of 24,943 beatmap pack sales (averaging US$1.99 a piece) by 13,635 users. After taking out apple’s cut and artist’s cut, this worked out to around $24,000 in revenue over the lifetime of the product. This was distributed between myself, the mappers, other devs who contributed time and the graphics design work.

The mapping team consisted of around 15 people, hand-picked based on their mapping style (I wanted to emulate the amazing feel of iNiS maps – the stuff that was in the Osu! Tatakae! Ouendan! / Elite Beat Agents series). We released a total of 37 beatmaps for it by the end. Beyond this though, there were over 100 beatmaps in some stage of creation.

A final release

Today marks the first update since September 2015. I was actually planning on making this update over a year ago, but ran into some serious issues with newer Xamarin versions where the game would freeze at random due to a native memory issue. I came back to this muiltiple times over the last year and while I have still not tracked down the underlying cause for it (likely a mono bug), I found a hacky workaround which allowed this release to happen.

A few things have changed in this release:

  • The game soundtrack has received a complete refresh thanks to superstar osu! composer nekodex! Experience a new take on the original osu!stream soundtrack in the game (pay attention to the main menu visualiser which is synced up to the new track). If you want to listen outside the game, you can also buy or stream the album (including the old version of each track) from your favourite store/streaming service .
  • Support has been added for new device aspect ratios and notch things (iPhone 10 / 11). The implementation of this was a bit hacky, but oh well.
  • All beatmap packs are now free. They are still available via the “store” screen as I wanted to retain the original game experience.
  • A few extra unreleased packs are also available as a bonus. These never made it out due to licensing or quality control issues, but I feel like there’s not much reason to hold back now.
  • Online ranking and login has been disabled. This used to be provided via a twitter login, but the API has changed and I don’t really feel like updating that or maintaining the scoreboard. Existing scores are still visible, as I want people to be able to experience as much of the game experience as it was when fully activated.

I will not be working on this project further due to time constraints. While this may disappoint many, please be rest assured that the osu!lazer project has mobile in its sights, including bringing back elements of osu!stream (like multitouch)! For those unaware, you can already join the testflight beta program and jump straight in to give that a try.

Of course, there’s one little thing I failed to mention: I am also releasing the source code to osu!stream, including all the beatmap creation tools and local multiplayer server (see the network branch).

While this is by no means a pristine codebase (try osu!lazer / osu!framework for that!), my hope is that some curious developers can see what goes in to making a game from scratch with a focus on releasing it rather than making the code perfect. You have to cut corners sometimes (and oh so many were cut) to get things released.

As for the future of this repository, I may provide code review and potential app updates if there are user contributions, although the criteria for these is quite specific so please make sure to read the README before diving in too deep.

While released under the all-permissive MIT license, please note that the osu! name is still covered by trademark law and you shouldn’t go distributing binaries of this app unless you first rebrand it. And I’d still rather you ask before doing so in such cases, out of courtesy.

Fun facts

A lot of cool things came from this project. Here’s some things you may now know about it:

  • The codename for this project was “osum”, as in “osumobile”. You might still see this term used around in the source code.
  • osu!stream was written almost completely from scratch, including rewrites of file parsing, hitobject display etc.
  • osu!stream was written with performance in mind. it could achieve a consistent 60fps on the iPhone 3G and original iPad, where most games were struggling to hit 30fps.
  • An Android release was planned. Test builds were sent out to many users and while it ran perfectly on a few devices, the number of issues and number of different devices to support, coupled with the fact that Android touch screens at that point in time has shocking response times meant I never released it.
  • All mappers were compensated for their time (between $100-200 per beatmap).
  • All music for osu!stream was licensed ahead of time of the game’s release. As I didn’t have huge up-front money to spend, artist were offered a percentage of game sales (30-50%).
  • I approached record labels to try and obtain correct licensing for larger artists, but was met with confusion over how to handle video game licensing. This situation has definitely improved over the last 8 years, as seen with featured artists.
  • Over halfway through development, and after finishing some amazing doujin music based beatmaps, licensing terms were changed beneath us and we were unable to release a whole heap of maps. They’ve finally been released with this release (I hope this doesn’t step on any toes, but it’s completely free so hopefully it’s fine).
  • Mapping for osu!stream is quite a challenge due to the stream-switches requiring alignment checks, consideration for hands hiding hitobjects and general multitouch playability concerns.
  • While some of the more arty UI components were provided by a community member (Koko Ban), the in-game skin is by me! and one of my favourite to-date.
  • We had some quite serious legal issues which caused the app to actually get removed from the app store after some douche made fraudulent claims that he owned copyright over the spinner texture.
  • osu!stream was used in the osu!arcade cabinets. This was a limited run prototype arcade setup constructed by my dad and I. osu!stream would run on a mini ITX board with lighting controlled over serial to an arduino. A LAN multiplayer server could allow two instances of osu!stream to compete in real-time head-to-head gameplay. One cool thing about this was you could see the other player’s touches on your screen due to the magic of no latency networking!
  • To increase the beatmaps available to osu!arcade, the beatmap tools were modified to do a best-attempt conversion of PC beatmaps. It doesn’t always work well and due to the tools being a completely different codebase, can be thrown off by all the zany things osu! maps do which are not supported. You can see a good assortment of beatmaps played here.
  • The game got pirated (of course) on jailbreak stores. They bought the packs using a specific UDID then hardcoded that into the app to continue to download the beatmaps from my server. Unfortunately for the poor souls that used this version, I quickly changed the server-side code to return 1gb of random data to these users, instead of the beatmaps they requested. This would waste their download quota and eventually crash their device.

A definitive osu!lazer FAQ

published 21 Oct 2019

With the osu!lazer project creeping closer and closer to being a feasible alternative to osu-stable, the number of questions I receive on a daily basis is growing. Let’s take a step back and answer some of the most frequency asked questions, sourced from you guys.

Development and Release Schedule

What is lazer / Is osu!lazer a new game or an update?

“lazer” is just a codename for the development project. When we’re done, lazer will become “osu!” and be released as an update for the existing client. It will migrate all existing user content and work on (basically) any system which osu-stable could run on.

Why isn’t lazer out yet?

Technically, it is already out (and has been since February 2017!). You can find the latest release available for Windows, macOS, iOS and android here. Linux users can self-compile (a binary will be available soon).

For now, it is released as a preview which will run alongside your existing osu! install.

Do note that as everything is stored in a new format, it will take up more disk space if you import your existing beatmap collection.

But I can’t submit scores! Until I can earn pp it’s not “out”

Fair call. We’re holding off adding this final step until everything else is ready, because we want to ensure the game experience is where we expect it to be first.

There is also a lot of final checking of compatibility which will need to happen before we can add score submission, to ensure the vast library of beatmaps, skins and otherwise are compatible and on fair grounds with historic scores.

Until then, you can enjoy the new multiplayer system which is exclusive to lazer, allowing local leaderboards to be created by anyone.

The release date was supposed to be 2015! I saw it in a teaser video!

The 2015 had a “?” and was meant to be a very optimistic outlook on development. The video description mentioned 2016 as being the goal.

Along the way, we ran into some serious blocking issues which resulted in the decision to rewrite osu!lazer as a new project from scratch, rather than building on top of the existing osu! codebase. This happened in the early months of 2016.

Since then we have been working extra-hard to build both a game framework and rebuild all the intricacies of osu!, while also fixing all the issues that existed in the old codebase to allow for implementing new features and extending on the game more than ever before.

We hope you can understand how excited we are to have this opportunity. The decision could only be made because we already have such a stable game client which (in general) satisfies the community’s needs quite well.

Why is it taking so long?

We are aiming for perfection.

That has meant, on several occasions, writing and rewriting systems until we are content that they will last well into the future. What we are working on is not a temporary solution, rather something to carry the osu! community through the rest of its (hopefully long) life.

We are also taking the time to involve the community via open source, allowing anyone to contribute code.

There are a few benefits of taking this time and diligence:

  • More people getting involved means there are less points of failure
  • High quality code and documentation makes it easier for new developers to jump in and help out.
  • We will hopefully not ever reach a similar fate as the old codebase, requiring a rewrite.

Why open source?

I have always wanted osu! to be open source. It is part of my belief that game development and operation should be very transparent to the end user. It may take a bit longer to do development in the open, but we are already seeing appreciation from other devs (and users) for making this move.

It also ensures that osu! can live on beyond me. If something were to happen, it is very feasible that members of the team and community can take over development into the foreseeable future.

Will there be a transition period where I can use stable?

In a similar way to stable and the stable-fallback release stream, the existing client will be available until it no longer has a function. For those of you with the mentality of “I want to use the current client forever, I don’t like lazer!” please bear with us. Our end goal is to make lazer better than the old client in every way, to the point there will be no reason not to use it.

New Features

What’s new in lazer?

While our primary goal is feature parity with stable, there are already some features that are exclusively available in lazer

Many new mods

We have already had several user-contribute mods, and welcome contributions of new and interesting takes on gameplay (we are sure many more will come!). Combine from the vast array of new mods to create unique custom challenges.

A new “timeshift” multiplayer mode (read more about this here).

This will pave the way for localised leaderboards that can be created and customised by the community. Of course, traditional real-time multiplayer will also be added in the future.

The ability to change settings anywhere and have them immediately take effect

Every single setting game-wide is made to take immediate effect. Hit ctrl-o anywhere to bring up the options (or ctrl-t to bring up the toolbar) and change a setting to get a real-time preview!

A more streamlined UI and a fresh design

The end goal is to have a user-first experience that feels like its own custom operating system. And will be nicely scalable across all screen sizes and resolutions.

Access to most of the web-only content from within the game (news, changelogs, user profiles, rankings, beatmap info)

The end goal is to make the game stand-alone from the web site.

A song density graph in gameplay

This is one of the first new gameplay UI additions. More will be coming, giving better insights into your gameplay and performance.

The ability to seek in replays

A long-standing feature request. The framework is also in place to allow rewind/replay support from point-in-time, paving the way for a “practice section” mode or similar.

The ability to hide difficulties

We’ve rewritten the beatmap database to allow for more flexibility and extensibility into the future.

Soft delete everywhere

If you make a mistake, you will be able to undo delete operations.

The ability to change the size and scale of the game with super-fine control.

This will allow players to remain in exclusive fullscreen (for lowest latency) while having full customisation of their UI and gameplay areas. You can choose whether to scale everything, just the menus, or just the gameplay for extremely fine control over your personal layout!

Will xxx feature be implemented in lazer?

We are primarily working on feature parity, but if the feature you’ve asked for has been something people have been asking for for quite a while, you can be 90% sure that it is one of the reasons we are working so hard to rewrite osu! in the first place!

Why is xxx feature being implemented in lazer but not stable?

As much as we’d like to bring new features to the stable client, it doesn’t make sense in many instances.

Generally, the implementation overhead for adding something in lazer is magnitudes lower than stable. By implementing it on both clients, the time of implementation would be more than double, delaying the release of lazer further.

For the last couple of years we have had a feature freeze on stable for this reason.

Will xxx feature be available on mobile?

Yes. Mobile and desktop will have feature parity.

Will new skin features be added?

Yes. While the existing skin/beatmap formats have been locked, we will be adding new formats (and in-game editors) which allow for huge amounts of flexibility.

Again, we’ve been planning everything with extensibility in mind. If you can think of it, it will likely be possible at some point in the future. That said, we are looking to make lazer have feature parity with stable before adding new options.

Will lazer have adjustable beatmap difficulty settings?

Yes, we are actively working on per-mod settings which will open a pathway for this.

Performance and Compatibility

Will it run on my PC?

If you can run stable, you will be able to run lazer. Note that users who can’t run stable for some reason (and are using stable-fallback) may or may not be able to run lazer. We are aiming for better compatibility than stable if anything.

Likewise, we also predict the system requirements will be equal or lower than stable. We are already seeing better performance, less frame drops and lower latency than stable in many scenarios (but not all, yet).

I’ve tried lazer but it runs worse than stable!

While lazer (usually) runs better than stable on most systems, there are always edge cases! We’ve still got plenty of work to do on optimisations (and already know all the remaining optimisations which will provide the largest benefits).

Make sure to check back regularly.

Will vulkan be supported?

While we are currently targetting OpenGL 4.0 and GLES (for mobile), we have been developing with graphics API portability in mind. We will be supporting vulkan and metal in the future.

Will my skin work?

Yes. We will be 100% supporting skins in gameplay, along with some menu-level interface elements (like the back button). We’re currently at about 90-95% implementation of skins for osu! and osu!catch, while osu!mania and osu!taiko need further work.

Will the new interface be skinnable?

We will be adding more skin support after we’re happy with the existing gameplay and support for legacy skins, including the ability to customise more of the menu screens.

Will there be touch controls to allow playing fast streams etc. on mobile?

Yes, these are in development. For now, you can connect a hardware keyboard.

Will there be touch controls for taiko/catch/mania?

Yes, these are also in development!

Will circles be larger on mobile?

Gameplay will likely get larger, yes.

We still have a lot of work to provide a good experience on mobile. We are primarily developing for PC first, and will address mobile UX issues over time.

I don’t like the new gameplay UI layout

Don’t worry, we are adding support for customisation of the UI layout, to the point you’ll be able to make it look like it used to (or potentially completely different!).

For example, a setting for the density graph is currently in development, which will allow it to either be hidden or less tall. Likewise, moving elements around the screen will eventually be present.

Leaderboards and Scoring

Will leaderboards / multiplayer be shared between PC and mobile platforms?

We haven’t made final decisions about the future of the global leaderboards. Personally I strongly believe in cross-platform play and hope to make this feasible.

Will lazer only use “score v2”?

We have implemented a new scoring algorithm which takes elements from v2 but is not the same. It is made with backwards compatibility in mind.

There is already an option in lazer to switch the display mode of scoring between quadratic (classic) and 1mil cap (standardised), and eventually this will be supported game-wide to cater to user preference.

Will the global leaderboards potentially get reset?

We will never reset the leaderboards. No final decisions are made yet, but the closest to a reset would be to freeze them in some accessible way. We value the rich and long history of osu! as much as you do.

Will scoring changes (like slider judgements) be applied to existing beatmaps?

We haven’t made a decision about this one yet. It may be applied only in places it doesn’t affect the playability of the map (using predetermined rules) or may only be applied to new beatmaps.

Where do I get lazer?

You can find the latest release available for Windows, macOS, iOS and android here.

Thanks for reading until the end. I hope this covered a lot of your concerns. If you happen to have questions which haven’t been answered here, feel free to reach out on twitter or in the comments below!


Enjoying osu! with others

published 27 Dec 2018

As 2018 draws to a close, I want to touch on something which I’ve felt strongly about for a while now. While I’ve described my thoughts and intentions in private to people I’ve met up with, this will be the first time discussing this publicly.

The reasoning behind not covering this topic until now is that I don’t like to dream about a future and all the cool things it could be, making big promises without being able to deliver soon enough for people in the present to enjoy.

osu! is a heavily community focused online game. Some players lean towards the competitive side while others play more casually, but I think everyone finds motivation from being around others striving hard to perform their best.

On any given day there are up to 800 concurrent multiplayer games, creating localised competition and enjoyment for their participants. Even so, I think we can all agree that the ecosystem created around Performance Points (or PP) and the global rankings are the focus for most.

Through all the complaints and criticisms of the system and its method of calculation, people of all skill levels are still striving for the highest pp they can achieve (quite often focusing on beatmaps which specifically reward more pp for less effort, known as “farm” maps).

This is one area which is quite unique in osu! – we don’t have the concept of seasons where leaderboards are reset. Scores remain for years and history is retained forever. It really lets users feel like they have achieved something solid, in my opinion – set a record and it’s there in stone (unless you beat it yourself!). But it also does lead to unavoidable stagnation.

I’m not about to go into the topic of PP calculation. The performance point (v2) system was made with the aim of objectively calculating difficulty and it did a pretty good job of that. Over the last month we have seen a huge increase in activity over at the dev discord aiming to improve the calculations while maintaining this overall goal. Things are in good hands and you will likely see changes applied over the next month.

The leaderboards and pp will always have their place in this community, but I don’t think it is healthy to focus exclusively on them. Let me introduce my plans for expanding osu! horizons via multiplayer.

From today, a new multiplayer system is live in lazer. I have been working on the logistics of this for several months on paper, and the team has spent the month of December focusing on this internally in relative secrecy.

As lazer is a very open development project, we don’t want to develop the whole system behind a curtain. Therefore we have developed a MVP (minimum viable product) that you can enjoy from today, but will continue to build on it more openly going forward. Huge props to smoogipoo, nanaya, notbakaneko, nekodex who all came together to make this happen in a relatively short period!


Introducing “timeshift” multiplayer

“timeshift” brings multiplayer rooms that can run over a longer period of time and involve large numbers of players. The room owner will be able to set the specifications for the room, along with a duration. During this specified time range, players will be able to join the room and compete in a localised leaderboard instance.

Think of it as asynchronous multiplayer.

Each room gains an overall leaderboard and a chat room which persist even after the room has expired. History of rooms will be kept indefinitely, and chat will be available for discussion after the results are finalised. I’m sure many of you are already imagining the possibilities of how this can be used. Let me touch on the concepts (and my personal goals) behind this system and how things will be developed going forward.

total attempts

Creating more ways for people to enjoy osu!

This will allow for leaderboards to be made by users. We plan on adding the tools to limit the scope of who can participate (private, friends only, limiting by rank, by country etc.)

Making tournaments easier to run

This should allow for a new generation of tournaments (well, competitions) which don’t need to be run in real-time. It can also provide a way for seeding players into groups before the main portion of a tournament starts, for instance. We will continue to flesh the system out with tournaments being one of the focuses, so if it’s missing features you may need, expect them to come in the future.

Giving players new areas to focus on competitively

As a player participates in multiplayer rooms, they will be visible in their “Participated” listing and (soon) show on their profile. It should hopefully add an extra dimension in where players place focus on when playing online.

room ranking

Increasing the creativity of gameplay

Rooms can be created with a variety of conditions. Currently quite limited, you can choose:

  • The beatmap to play
  • Which mods to play with
  • The duration the room is active for

Going forward, here are a few of the coming options (the system has been designed with these in mind from ground-up):

  • A playlist of beatmaps to play
  • A progression mode (marathon, or beatmap-by-beatmap)
  • A selection of required mods, plus a selection of allowed (options) mods. Like FreeMod but more choice. Mods apply per beatmap in the playlist.
  • A maximum number of attempts per player
  • A maximum number of players participating
  • A choice of leaderboard metric (currently locked to total score, but will include accuracy, pp and max combo)
  • A start time, for scheduled matches in the future.

Current room creation limitations:

  • 1 concurrent room for non-supporters, 5 rooms for supporters.
  • Duration is limited to 24 hours while we stabilise things. Eventually this will be allowed to be much longer.

We have more ideas, but these are the base ones we will be working on in the coming weeks and months. Also, important to note that lazer already has four or five unique mods not found in stable, and this number is only set to increase! User contributed mods are being openly accepted and each new mod increases the number of creative possibilities when it comes to designing a fun multiplayer room!

And that’s as far as I’m willing to go when it comes to explaining the progression of this system. From here on it will depend on how the community uses the new rooms and the feedback we get as to how we mould this into a permanent and (hopefully) well enjoyed aspect of the osu! ecosystem.

Blinds mod, a new lazer-only hardcore mod selection

Let me preempt some questions

When is it available?

Now. Go get lazer if you haven’t already and install it alongside your stable install! It’s more stable than you may think, and available for windows and macOS (linux too if you compile it yourself).

Global leaderboard submission is of course not in lazer yet, so enjoy a different and possibly more chill experience than you’re used to.

What about pp ranking in lazer?

For the time being this is not a focus. We have everything in place to make that eventually happen, but for now I want to make a new unique experience in lazer that differs from what one can play on stable.

So I guess that means I can’t join “timeshift” rooms from stable?

No, this is and will always be lazer exclusive.

What about real-time multiplayer? This feels like the wrong direction!

I can understand this concern, but I have already planned a logical progression from this system into a fully real-time system. All the features that will be made for timeshift will be 100% available in real-time rooms once we get to that point. Thing of this implementation as a prerequisite to adding real-time play modes in the extensible way we want to do it.

Of course, realtime/non-realtime will be a user choice when creating a room in the future!

What about coop play?

Once we flesh out the timeshift system and get to the point we can add cooperative modes, it will happen. Mark my words.

Can we view this from the web site / our profiles?

We are working on adding participation history and overall visibility on the web. This will come a bit later!

Will room history be maintained from the outset?

We’ll do our best to maintain all history, but as this is a new system do understand if we need to delete data for one reason or another. Possibly multiple times.

I don’t play multiplayer / I don’t like this concept.

That’s perfectly fine! In fact, if you enjoy solo play for leaderboards you may actually like this concept more than the multiplayer available on stable.

If you don’t, nothing is forcing you to play it. Global leaderboards are not going anywhere.

It’s buggy

We have plenty of work ahead of us. Go easy on it in the mean time.

This is our final surprise for 2018

I hope we can all agree it’s been an amazing year and we’ve ended it with some amazing notes!

Between Featured Artist releases, this announcement, completely operational iOS lazer builds (alongside windows/macOS/linux), web design renewals and some long-awaited features being added to stable, we’ve made some phenomenal progress.

A huge pat on the back to the whole osu! team (devs, community managers, BNs, QATs, moderators and anyone I missed).

I was planning on doing a blog catching up on Q4 2018, but due to time constraints it will likely happen sometime in January. Hopefully this makes up for it :).

Wishing you all a great new year period. Much of the team will be away on a well-deserved break for next week or so, but we’ll be in touch soon!


August 2018 Development Update

published 14 Aug 2018

Welcome to the August development update. A bit late but you are probably used to that by now. Been focused on keeping things running as usual.

July 2018 Review

osu The osu!lazer project. The future of the osu! client. Download preview releases from here and follow along!

July was a bit of a slow month, and we didn’t really cover the scope we planned on.

You can view a full digest of the changes from July using our new changelog system.

Basic structural implementation of osu!mania editor COMPLETE

Work on the basic structure for the osu!mania editor has been completed. It is now at a point in development roughly equal to the osu! mode editor. While it’s not yet in a usable state, the goal was to get a framework in place for structuring the different editor overlays for each ruleset, which was definitely achieved by this work.

I’m not sure if we’ll continue adding the editors for the remaining two rulesets or expand functionality for osu!/mania to begin with, but you can expect further improvements coming to the editor over the coming months.

Add automatic error reporting on hard crash PENDING REVIEW

Sentry support has been implemented, along with allowing singular errors to result in a non-crash result (similar to how osu-stable avoids many unnecessary crashes). Not yet merged, as we found quite a few cases where exceptions were not being handled correctly along the way.

Should be in soon!

Full-blown custom sample support COMPLETE

All hitsounds should now correctly be played back during gameplay. This includes per-object samples (commonly used in mania) as well as unlimited custom sample sets associated with timing sections.

Selectable full-screen resolutions PENDING REVIEW

While framework support has been added for this, the UI is not yet merged. Should happen this month.

More skinnable elements COMPLETE

Skin support was added for slider ticks, reverse arrows, follow circles, slider balls and follow points (in addition to existing support for hit circles, approach circles and judgements). If you already have a skin loaded, you should see the new elements displaying. If you haven’t yet imported a skin, it’s as simple as dragging an .osk file into the lazer window.

Gameplay visual fixes COMPLETE

Slider repeats were being shown too late, causing visual uneasiness when trying to read sliders with one of more repeat points. This has been resolved and should now match stable.

The slider follow circle was visually (and logically) far too large, making sliders much easier than they should be. This has now been resolved and should match stable.

Sliders which use a linear curve style felt a bit ugly when snaking in due to the reserve arrow changing its rotation in a single frame. Interpolation has been added to make these corners smoother.

osu-framework The game framework behind osu!lazer. Written from scratch with osu! in mind, but very extensible.

Move both osu and osu-framework to target netcore2.1 COMPLETE

All framework and osu! deployments are now 100% netcore2.1! This is a huge step forward and allows us to focus on making the net core experience as good as possible. It also means that release builds are running on the same runtime as we are using to debug, which makes it easier to locate issues before they are deployed to a production environment.

To make this happen, I ended up forking and publishing out own squirrel.windows packages to keep automatic update support working on windows builds.

The performance issue I mentioned in the last blog post (macOS debugging) has not yet been resolved, but I have filed an issue to track this. Hopefully it will get some attention soon.

Fix remaining autosize / update order issues PENDING REVIEW

A pull request is open which fixes the majority of these issues. It still needs to be reviewed and tested for any performance regressions, but should be arriving soon.

Ongoing input subsystem improvements

Huge improvements have been made to the input handling. There is still some further changes we are looking to make which will likely be completely over the next month.

osu-web The new web front-end. Already live but hasn’t yet replaced the old site, pending feature parity.

You can view a full digest of the changes from July using our new changelog system.

Almost half the milestone tasks were unachieved. Pretty slow progress. Not really much to say here…

Move osu-web score lookups to elasticsearch COMPLETE

This has now been deployed. You may notice that profile pages are now quite a lot faster to load (especially for players with a lot of scores). Eventually we will update other pieces of the infrastructure to benefit from this performance improvement, such as in-game song select lookups.

New private messaging system WIP

Work on this was slow this month, as attention was diverted to code review instead.

Add ability to report users from profile PENDING REVIEW

We’re finally on track for adding back one of the few missing features from the new profile pages. You will soon be able to report users again.

Test and deploy new registration process

Did not happen.

August 2018 Goals




Leaving our August goals blank for the time being. Already halfway through the month, we’re playing catch-up on missed goals and outstanding bug fixes. You can still click in to the headings above to see what is remaining to catch up on missed milestone tasks, though!

Thanks for following along.


July 2018 Development Update

published 03 Jul 2018

Hello everyone! Back again one month later with a review of how the last month went, along with a report on our direction for this coming month.

June 2018 Review

osu The osu!lazer project. The future of the osu! client. Download preview releases from here and follow along!

As you will see below, we failed a few of the milestone goals, but I think we can all agree that amazing progress was made. You can view full changelogs by checking out each build from the month of June (I hope we can present this on a single page next month!):

Changes for the month: 2018.629.0, 2018.628.0, 2018.627.0, 2018.626.0, 2018.622.0, 2018.621.0, 2018.618.0, 2018.616.0, 2018.614.0, 2018.612.0, 2018.611.0, 2018.608.0, 2018.607.0, 2018.606.0.

That’s a lot of builds!

Basic structural implementation of osu!mania editor WIP

Some structural changes were made to get things ready for implementation of the mania editor, but actual editor work was slower than expected. Further work on this will continue into July.

Add automatic error reporting on hard crash SKIPPED

This was not completed. That said, we were able to fix most of the startup crashes that were reported. I hope to complete this in July as it will help us act quickly on reports, with less back-and-forth to figure out what went wrong.

Begin work on multiplayer lobby system SKIPPED

Still don’t feel like we are in a state where I want to start adding multiplayer. I’ll know when it’s ready. There’s just too many other issues I want to fix first.

osu-framework The game framework behind osu!lazer. Written from scratch with osu! in mind, but very extensible.

While we haven’t been adding much new functionality to the framework, bug-fixes are ongoing. We still have a lot of outstanding issues surrounding platform compatibility and single-frame drawable glitches due to incorrect invalidation scenarios.

Move both osu and osu-framework to target netcore2.1 WIP 80%

We were successful in migrating all the development configurations to be netcore2.1 only. This means that all development going forward will be done under the netcore runtime. Windows releases are still targeting net471 for the time being to ensure we don’t hastily break auto-updates. We also need to still ensure that all tests still perform correctly under netcore when run on our CI infrastructure. These remaining tasks should be completed over the month of July.

That said, we did run into a performance degradation during debug sessions on macOS that we are currently tracking in hope of a fix.

Add automatic nuget package builds of osu and osu-framework LIVE

We completely made the switch to nuget dependencies for the framework. This turned out to be a really positive move forward, further isolating development of the two projects and allowing for better testing at each end. We are now pushing out regular builds to nuget.org

For the time being, we’re holding off the osu-side nuget packages until there’s a demand for them. If you’re working on a custom ruleset and are interested in seeing this happen then please let us know and we’ll expedite it!

osu-web The new web front-end. Already live but hasn’t yet replaced the old site, pending feature parity.

Changes for this month:

2018.702.0, 2018.701.0, 2018.629.0, 2018.628.0, 2018.627.0, 2018.626.0, 2018.625.0, 2018.622.0, 2018.621.0, 2018.620.0.

Complete the new changelog page LIVE

The new changelog page is now live! It has already been a great help in visualisation the status of our various projects. I hope you guys are enjoying viewing and commenting on individual releases as we are making the regular releases! Click any of the change links in this post to check it out for yourself, if you haven’t already.

Move osu-web score lookups to elasticsearch WIP 80%

The back-end component for this was written and is deployed. The front-end changes are also ready to be deployed; we’re just making a few final touches and quality checks before deploying this. It should be the start of a lot more flexibility in score lookups, though. Exiting stuff!

New private messaging system WIP ~50%

The new PM system is coming along nicely. Due to the fact this is part of an experimental push for changing the architecture of our more interactive web-based components, it will be taking a bit longer than a month to get in place. This was expected; hopefully we’ll see some results live this month, though!

July 2018 Goals


Check out the milestone for a full list of tasks. We’re keeping actual listed tasks below to a minimum to ensure we can hit goals this month. We also still have some lingering issues from June that need to be completed.

Full-blown custom sample support

While lazer has had the basic level of custom hitsounds for quite a while, it has not supported infinite custom banks or per-hitobject hitsounds (commonly used in mania). This will be added this month.

Selectable full-screen resolutions

All the building blocks for this are coming together, so this should be feasible to have implemented during this milestone.


Fix remaining autosize / update order issues

There are a lot of remaining cases where incorrect update orders between various low-level framework components means that layouts or styles are incorrect for single frames. Here’s a couple of examples (notice the single-frame flicker of an incorrect state):

This is turning up in many places and requires some dedicated attention to fix. Likely not going to be an easy task, but something that needs to be addressed sooner rather than later.

Ongoing input subsystem improvements

We have begun moving away from passing around full input states between input sources (handlers) and the drawable hierarchy, but there are some regressions and WIP elements which still require attention. The completed result should give us a much more stable and manageable input framework. Extensive unit tests are being added as we go to avoid regressing int he future.


Add ability to report users from profile

This has been missing for the longest time and remains one of the few features we need to truly turn off the old site.

Test and deploy new registration process

While the API already contains the logic to handle user registrations, they are still currently occurring through the old site. This is a really immersion-breaking experience which new users should not be exposed to.

I hope to move registration to inside the client (with an implementation in osu-stable) to address this.

Thanks for following along. Will report back in a month!