osu-lazer

Fix several issues with incorrect sample playback (#35685 )

* Add failing test coverage for layered hit samples not playing in mania when beatmap is converted

Adding the `osu.Game.Rulesets.Osu` reference to the mania test project
is required so that `HitObjectSampleTest` base logic doesn't die on

https://github.com/ppy/osu/blob/f0aeeeea966f06add12cf2bca3dd48dac8573e82/osu.Game/Tests/Beatmaps/HitObjectSampleTest.cs#L88-L91

* Fix layered hit sounds not playing on converted beatmaps in mania

Compare
https://github.com/peppy/osu-stable-reference/blob/f9e58b4864a10f801393199e7652b2192c7342c3/osu!/GameplayElements/HitObjects/HitObject.cs#L476-L477.

In case of converted beatmaps, the last condition there
(`BeatmapManager.Current.PlayMode != PlayModes.OsuMania`) fails,
and thus layered hitsounds are allowed to play.

* Add failing test coverage for mania beatmap conversion assigning wrong samples to spinners

* Fix mania beatmap conversion assigning wrong samples to spinners

A spinner is never `IHasRepeats`. It was a dead condition, leading to
the hitobject generating fallback `NodeSamples`, which in particular
feature a silent tail which stable doesn't do.

Noticeably, stable also appears to force the head of the generated hold
note to have no addition sounds:

https://github.com/peppy/osu-stable-reference/blob/f9e58b4864a10f801393199e7652b2192c7342c3/osu!/GameplayElements/HitObjects/Mania/SpinnerMania.cs#L86-L89

* Add failing test coverage for file hit sample not falling back to plain samples if file missing

* Allow `FileHitSampleInfo` to fall back to standard samples if the file is not found (or not allowed to be looked up)

I'm honestly not 100% as to how closely this matches stable because I
reached the point wherein I'd rather not look at stable code anymore, so
as long as this passes tests I'm fine to wait for someone else to report
new breakage.

* Use alternative workaround for lack of osu! ruleset assembly in mania test project

* Fix encode stability test failures

Bartłomiej Dach · 2025-11-15 16:19:08 +09:00

bd4ed49c06

Adjust gameplay sample models to support custom sample sets

This is a set of model changes which is supposed to facilitate support
for custom sample sets to the beatmap editor that is on par with stable.

It is the minimal set of changes. Because of this, it can probably be
considered "ugly" or however else you want to put it - but before you
say that, I want to try and pre-empt that criticism by explaining where
the problems lie.

Problem #1: duality in sample models
---

There is currently a weird duality of what a `HitObject`'s samples will
be.

- If an object has just been placed in the editor, and not saved /
  decoded yet, it will use `HitSampleInfo`.

- If an object has already been encoded to the beatmap at least once, it
  will use `ConvertHitObjectParser.LegacyHitSampleInfo`.

As long as that state of affairs remains, `HitSampleInfo` must be able
to represent anything that `LegacyHitSampleInfo` can, if feature parity
is to be achieved.

Problem 2: The 0 & 1 sample banks
---

Custom sample banks of 2 and above are a pretty clean affair. They map to
a suffix on the sample filename, and said samples are allowed to be
looked up from the beatmap skin. `Suffix` already exists in
`HitSampleInfo`.

However, the 1 custom sample bank is evil. It uses *non-suffixed*
samples, *allows lookups from the beatmap skins*, contrary to no bank /
bank 0, which *also* uses non-suffixed samples, but *doesn't* allow them
to be looked up from the beatmap skin.

This is why `HitSampleInfo.UseBeatmapSamples` has been called to
existence - without it there is no way to represent the ability of using
or not using the beatmap skin assets.

As has been stated previously in discussions about this feature, it's
both a *mapping* and a *skinning* concern.

There are many things you could do about either of these problems, but I
am pretty sure tackling either one is going to take *many* more lines of
code than this commit does. Which is why this is the starting point of
negotiation.

Bartłomiej Dach · 2025-10-22 14:58:08 +02:00

c361e6d3c2

Disallow zero-length sliders from specifying a non-zero number of repeats (#35220 )

Bartłomiej Dach · 2025-10-08 16:01:04 +09:00

1a522a19f1

Fix lack of slider encode-decode stability due to truncating control point coordinates

Mostly closes https://github.com/ppy/osu/issues/33505.

Compare

	https://github.com/ppy/osu/blob/97e6187f0d7c3dbee896596a623e34627135bf0e/osu.Game/Rulesets/Objects/Legacy/ConvertHitObjectParser.cs#L56-L59

I say "mostly" here because I'm rather skeptical that this is 100% rock
solid still, for one reason - namely that the game stores path control
point coordinates relative to the head, then turns them into absolute
coordinates when encoding, and then on decoding turns them back into
coordinates relative to the head, which in floating-point world is a Bad
Idea because of round-off error. But I'm not fixing that without
introducing a completely new beatmap format or rewriting half the
editor to address that, so I'll just pretend that I don't know any
of this until someone notices.

Bartłomiej Dach · 2025-06-10 14:09:13 +02:00

eab02a2aa5

Add setters to hitobject coordinate interfaces

Bartłomiej Dach · 2024-12-27 11:01:47 +01:00

0d16ed028b

Merge pull request #30984 from smoogipoo/fix-mania-spinner-conversion

Fix "spinner" conversion for mania-specific beatmaps

Bartłomiej Dach · 2024-12-19 15:30:10 +01:00

1057aa867f

Populate default LegacyType value on convert hitobjects

Normally not an issue, but some tests create their own hitobjects
deriving from `ConvertHitObject`.

Dan Balasescu · 2024-12-06 17:47:14 +09:00

8dda5aada8

Apply minor performance rules

Huo Yaoyuan · 2024-11-28 20:41:44 +08:00

f5a7716509

Remove unused params

Dan Balasescu · 2024-11-11 16:10:15 +09:00

8c68db0a36

Add IHasLegacyHitObjectType to ConvertHitObject

Dan Balasescu · 2024-11-11 16:10:15 +09:00

7206e97b7b

Merge implementations of ConvertHitObjectParser

Having these be separate implementations sounded awesome at the time,
but it only ever led to confusion. There's no practical difference if,
for example, catch sees hitobjects with `IHasPosition` instead of
`IHasXPosition`.

Dan Balasescu · 2024-11-11 15:09:13 +09:00

e1d93a7d9c

Fix hitobjects without custom sample banks parsing as not auto sample bank

OliBomby · 2024-10-01 12:11:44 +02:00

2381c2c72c

Implement auto additions editor-only

OliBomby · 2024-10-01 16:40:48 +09:00

0a78eb9628

add back protection against perfect curve segments with > 3 points

OliBomby · 2024-08-21 12:03:15 +02:00

eefd7cf083

Fix perfect curve anchors losing type between reloads

OliBomby · 2024-08-16 14:24:55 +02:00

a2e26ba9ff

Fix hit object coordinates being truncated to int values

Closes https://github.com/ppy/osu/issues/29340.

Dean Herbert · 2024-08-09 16:34:39 +09:00

d072c6a743

Update HasFlag usages

Dean Herbert · 2024-07-03 00:19:04 +09:00

5350945340

Replace usages of is null with == null

Dean Herbert · 2024-04-03 01:50:39 +08:00

94cbe1838f

Use Slice method instead of index range operators for readability

Salman Ahmed · 2024-03-19 02:43:36 +03:00

a8ce6a0bba

Make the code more clear

Huo Yaoyuan · 2024-02-29 20:02:04 +08:00

a11e63b184

Fix method indent size

Co-authored-by: Salman Ahmed <frenzibyte@gmail.com>

Huo Yaoyuan · 2024-02-29 10:47:16 +08:00

f28f19ed7e

Fix formatting

Huo Yaoyuan · 2024-02-29 00:24:24 +08:00

e86ebd6cdb

The overhead of LINQ is not ignorable

Huo Yaoyuan · 2024-02-29 00:07:00 +08:00

470d2be2e1

Use ArrayPool instead of stackalloc

Huo Yaoyuan · 2024-02-28 22:51:36 +08:00

bcb91f348d

Update parsing.

Huo Yaoyuan · 2024-02-28 22:42:08 +08:00

fe34577ee2

Parse points and segments for path string

Huo Yaoyuan · 2024-02-28 22:01:39 +08:00

f49aa4d815

Fix incorrect score conversion on selected beatmaps due to incorrect difficultyPeppyStars rounding

Fixes issue that occurs on *about* 246 beatmaps and was first described
by me on discord:

    https://discord.com/channels/188630481301012481/188630652340404224/1154367700378865715

and then rediscovered again during work on
https://github.com/ppy/osu/pull/26405:

    https://gist.github.com/bdach/414d5289f65b0399fa8f9732245a4f7c#venenog-on-ultmate-end-by-blacky-overdose-631

It so happens that in stable, due to .NET Framework internals, float
math would be performed using x87 registers and opcodes.
.NET (Core) however uses SSE instructions on 32- and 64-bit words.
x87 registers are _80 bits_ wide. Which is notably wider than _both_
float and double. Therefore, on a significant number of beatmaps,
the rounding would not produce correct values due to insufficient
precision.

See following gist for corroboration of the above:

    https://gist.github.com/bdach/dcde58d5a3607b0408faa3aa2b67bf10

Thus, to crudely - but, seemingly accurately, after checking across
all ranked maps - emulate this, use `decimal`, which is slow, but has
bigger precision than `double`. The single known exception beatmap
in whose case this results in an incorrect result is

    https://osu.ppy.sh/beatmapsets/1156087#osu/2625853

which is considered an "acceptable casualty" of sorts.

Doing this requires some fooling of the compiler / runtime (see second
inline comment in new method). To corroborate that this is required,
you can try the following code snippet:

    Console.WriteLine(string.Join(' ', BitConverter.GetBytes(1.3f).Select(x => x.ToString("X2"))));
    Console.WriteLine(string.Join(' ', BitConverter.GetBytes(1.3).Select(x => x.ToString("X2"))));
    Console.WriteLine();

    decimal d1 = (decimal)1.3f;
    decimal d2 = (decimal)1.3;
    decimal d3 = (decimal)(double)1.3f;

    Console.WriteLine(string.Join(' ', decimal.GetBits(d1).SelectMany(BitConverter.GetBytes).Select(x => x.ToString("X2"))));
    Console.WriteLine(string.Join(' ', decimal.GetBits(d2).SelectMany(BitConverter.GetBytes).Select(x => x.ToString("X2"))));
    Console.WriteLine(string.Join(' ', decimal.GetBits(d3).SelectMany(BitConverter.GetBytes).Select(x => x.ToString("X2"))));

which will print

    66 66 A6 3F
    CD CC CC CC CC CC F4 3F

    0D 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00
    0D 00 00 00 00 00 00 00 00 00 00 00 00 00 01 00
    8C 5D 89 FB 3B 76 00 00 00 00 00 00 00 00 0E 00

Note that despite `d1` being converted from a less-precise floating-
-point value than `d2`, it still is represented 100% accurately as
a decimal number.

After applying this change, recomputation of legacy scoring attributes
for *all* rulesets will be required.

Bartłomiej Dach · 2024-01-10 19:30:18 +01:00

7c9adc7ad3

Fix handling of combo offset without new combo, and incorrect lazer tests

Dan Balasescu · 2023-11-24 09:46:06 +09:00

10e16e4b04

Merge branch 'master' into fix-combo-handling

Dan Balasescu · 2023-11-23 16:22:01 +09:00

7bc2d6e6cd

Fix combo/combo colouring issues around spinners

Dan Balasescu · 2023-11-23 13:54:05 +09:00

7998204cfe

Fix first object after break not starting a new combo

Dan Balasescu · 2023-11-22 17:22:48 +09:00

cb4568c4a1

Fix incorrect parsing of BSpline curve types

Dean Herbert · 2023-11-21 15:38:12 +09:00

3d094f84ad

Use static BEZIER instead of allocating new every time

Bartłomiej Dach · 2023-11-20 13:35:07 +09:00

80a3225bb2

Fix code style/quality issues

cs · 2023-11-13 08:25:27 +01:00

fa976a5aa0

CI Fixup

cs · 2023-11-11 15:02:06 +01:00

54b8244a18

Generalize Bezier curves to BSplines of Nth degree

Thomas Müller-Höhne · 2023-11-11 13:32:53 +01:00

926636cc03

Fix fudge not being applied

Dean Herbert · 2023-10-20 20:18:14 +09:00

906b700aca

Move helper method to LegacyRulesetExtensions and stop applying rounding allowance to catch

As discussed, it isn't used in stable like this. Was a mistake.

Dean Herbert · 2023-10-20 18:57:14 +09:00

3fb74cb5f9

Fix various other inspections

Dean Herbert · 2023-10-17 17:48:51 +09:00

e081fa48a2

Move LegacyLastTickOffset specification to generation code and stop passing everywhere

Dean Herbert · 2023-09-29 14:32:29 +09:00

81485c548c

Merge pull request #24800 from sw1tchbl4d3r/invalid_bank

Default to normal bank if invalid sample bank is specified

Dean Herbert · 2023-09-19 22:28:34 +09:00

aea7f81f1c

Replace sample bank fix with more correct fix

stable does not treat unknown enum members as `None` / `Auto`, it treats
them as `Normal`:

    switch (sampleSet)
    {
        case SampleSet.Normal:
        default:
            sample = 0;
            break;
        case SampleSet.None:
        case SampleSet.Soft:
            sample = 1;
            break;
        case SampleSet.Drum:
            sample = 2;
            break;
    }

    (from https://github.com/peppy/osu-stable-reference/blob/1531237b63392e82c003c712faa028406073aa8f/osu!/Audio/AudioEngine.cs#L1158-L1171).

Bartłomiej Dach · 2023-09-19 13:53:49 +02:00

c4a0ca326e

Make not extension method

Dean Herbert · 2023-09-15 18:13:04 +09:00

56cc2b62f0

Move to extension method and throw on non-legacy ruleset

Dean Herbert · 2023-09-15 17:38:34 +09:00

0031da76ff

Default to none bank if invalid samplebank is specified