forum

[Editor] Improve Song playback speed with 25%, 50%, 75%.

posted
Total Posts
65
This is a feature request. Feature requests can be voted up by supporters.
Current Priority: +78
show more
abraker
No, but changing the pitch on any speed would be helpful. Also the editor needs to have smooth playback on slower speeds. Mapping to crackled audio is hard at times when there is a lot going on.
Fullerene-

abraker wrote:

No, but changing the pitch on any speed would be helpful. Also the editor needs to have smooth playback on slower speeds. Mapping to crackled audio is hard at times when there is a lot going on.
It is not possible to both maintain the original pitch and have smooth audio at the same time, you either pick one or the other. Pitch shift is miles better as it makes it a lot easier to determine pitch and incorrect snaps, I seriously don't know what valid argument could be brought up to keep it the way it is right now.
Kibbleru
sorry, i would prefer just to slow the song lol, its gunna be super annoying
Topic Starter
WingSilent
I tried with another program and i heard clearly everything from the music.
Just try before rushing on "NO".
Score_Under
For anyone who doesn't want to have to do this themselves, here's a song (t+pazolite - GLITCH MANIAXXXX)
Original speed (20 second sample): http://a.pomf.se/jwrdup.webm
Same clip slowed to 25%: http://a.pomf.se/illmmh.webm
Stretched without pitch shift: http://a.pomf.se/ccdmmz.webm (not sure if this is the same algorithm osu uses but it sounds very similar)

Encoded in webm in the hopes that your browser can play them.

I think the beats sound very clearly here. There are some beats you simply can't hear in the version without the pitch shift.
It does seem useful, but then again I don't map (yet?) so I wouldn't know.
abraker
Jeez the one without the pitch shift is crap. What did you use?

This is how it is supposed to sound like and this is the action osu is missing (25% speed of original):
http://a.pomf.se/arjbxp.webm
Bara-
This should never happen
Ignore the reason why, I was incorrect
Also, it would just sound terrible if you use it to check a map/mod
I strongly disagree with this, and I'll try to do as much to ever prevent this from happening
Redon
Shoegazer

baraatje123 wrote:

Back to my point for Taiko mappers
This should never happen
It makes distinguishing little different pitches near impossible
Read Halogen-'s post. Have you tried slowing down the pitch of a song? There's no difference in pitch. There are mappers with absolute (or relative) pitch who can prove this.

Note: I do not know how to call the pitch of a music, so I just use random number
Let's say, the pitch of a low tuned note is 80, a high tuned note is 100
Has a difference of 20, so good to distinguish
If you then play at 25%, the pitch will also be 25%
80 becomes 20, 100 becomes 25
Now there is only 5 difference, which is really low, and it's really hard to find out whether it's high or low
Since it goes too fast at 100%, even at 50%, most people use 25% to do this
Read Halogen-'s post. Slowed down or not the root note will be the same, and the differences in pitch between notes will be the same. The difference in pitch doesn't change because of the song is slowed down, it just lowers the octave of the music. Listening to it on a lower speed will make the music a lower octave and unarguably easier to catch pitch differences slowed down. Slowing down the track while keeping the pitch creates distortion and other unnecessary artifacts that create unnecessary clutter and difficulty.

Also, it would just sound terrible if you use it to check a map/mod
I strongly disagree with this, and I'll try to do as much to ever prevent this from happening
Completely subjective and irrelevant to the topic. Please stop.
Bara-
Oh
Oops
I always thought pitches work like that
Sorry

But still, it'll sound terrible, and it'll be annoying as heck for modders (and probably almost all mappers)
lemonguy

baraatje123 wrote:

But still, it'll sound terrible, and it'll be annoying as heck for modders (and probably almost all mappers)
Then why not just have it as a toggled option? That way, everyone can be happy.
Bara-
Meh
As long as there'll be a toggle, and it's not forced it then seems fine
But only as long as there'll be a toggle
ziin
If you ask for a toggle peppy will come and hit you with a bat.

abraker wrote:

Jeez the one without the pitch shift is crap. What did you use?

This is how it is supposed to sound like and this is the action osu is missing (25% speed of original):
http://a.pomf.se/arjbxp.webm
How did you do it?
I tried 4 methods in audacity:
Change Tempo (sounds a little worse than osu)
Change pitch +300%, set rate to 11025 (sounds like change tempo)
set rate to 11025, change pitch +300% (sounds OK, maybe this is how you did it?)
set rate to 11025, impossible to hear properly.

-50% in pitch is acceptable IMO, but -75% is REALLY hard to hear properly. The most I could support is a linear -75% speed/-50% pitch, -50% speed/-33% pitch, -25% speed/-16% pitch.
Kurokami
I'm not completely against this. There are song where just slowing down doesn't help or even make notes harder to hear. But it would be nice to have as a tick besides the speedbar since there are songs where this will only make them less hearable. So because there are possibilities for both type of song I think this shouldn't be forced, just an optional choice.

Have my starless support.
abraker

ziin wrote:

I tried 4 methods in audacity
I used wavepad
lenpai
Oh this. This will work well in certain hectic songs with a lot of distortions and syth crap. This will be good.

If toggle is not a viable option then have two of those 25%-100% stuff. Or at least place the pitch control somewhere in the editor.
Halogen-
going to bump because this is a feature that needs to happen, it will benefit all mappers

also going to shoot down the below post, because if you're going to have a negative reason for something, you need to make sure you're right:

baraatje123 wrote:

Back to my point for Taiko mappers
This should never happen
It makes distinguishing little different pitches near impossible
Note: I do not know how to call the pitch of a music, so I just use random number
Let's say, the pitch of a low tuned note is 80, a high tuned note is 100
Has a difference of 20, so good to distinguish
If you then play at 25%, the pitch will also be 25%
80 becomes 20, 100 becomes 25
Now there is only 5 difference, which is really low, and it's really hard to find out whether it's high or low
Since it goes too fast at 100%, even at 50%, most people use 25% to do this

Also, it would just sound terrible if you use it to check a map/mod
I strongly disagree with this, and I'll try to do as much to ever prevent this from happening
- Taiko is nowhere near as pitch-intensive as osu!mania; if mapping for pitch is improved in osu!mania, it will most certainly be improved in Taiko

- You absolutely should not be advocating against this idea when those in support of it actually have musical ears/backgrounds and understand differences in tonality and you admit that you don't at least know how to "call the pitch of a number".

Let's say, the pitch of a low tuned note is 80, a high tuned note is 100
Has a difference of 20, so good to distinguish
If you then play at 25%, the pitch will also be 25%
80 becomes 20, 100 becomes 25
Now there is only 5 difference, which is really low, and it's really hard to find out whether it's high or low
Since it goes too fast at 100%, even at 50%, most people use 25% to do this
This does not work at all in practice. You cannot arbitrarily say that a note is tuned at X frequency because musical notes have fixed and finite frequencies which increase by a multiple of 2. Additionally, numerical difference in frequencies are not directly correlative in difficulty to separate reasonable tones -- it will be equally easy to discern a 7th octave A at 3520 hz and 7th octave A# at 3729 hz (difference: 209 hz) as it will be to discern a 3rd octave A at 220 hz and 3rd octave A# at 233.1 Hz (difference: 13.1 hz, or < 1/10 the previous difference). Because your argument relies on the "separation" of "note tuning" and doesn't account for no difficulty change in discerning notes within audibility, it's completely invalid.

To apply your numbers and illustrate how this is erroneous, we'll take two notes: A4 (440 hz) as your "100" and E4 (~330 hz) as 75 and not 80 because no such frequency perfectly aligns on a musical note without the application of fine tuning. If we proportion the speed and tempo reduction by multiplying the percentage of speed reduced by the frequency of a given note, we can prove that the root note of a given melody does not change. For the purpose of this example, we'll do 50% speed. When adjusting linearly, we adjust the frequencies of the given notes by the same speed factor:

50% of 440 = 220 Hz
50% of 330 = 165 Hz

The frequencies of 220 Hz and 165 Hz are 50% of the originals (440 and 330), and these frequencies also represent musical notes themselves: 220 Hz = 3rd octave A, 165 Hz = 3rd octave E.

25% of 440 = 110 Hz
25% of 330 = 82.5 Hz

The frequencies of 110 Hz and 82.5 Hz are 25% of the originals (440 and 330), and 50% of the frequencies before them (220 and 165).

Coincidentally enough, if you were to double the speed of a song (and therefore, double the frequencies of any given note in this suggestion), quick calculation and research will have you know that the root tone of the note is preserved, simply raised an octave. This means that if the speed is increased/decreased by any square of 2, the root note will be preserved, but the note's octave will change. This indicates that the representation of a given melody, within the constraints of osu!'s editor, will be preserved at 25%, 50%, and 100%. Cutting the speed/frequency to 75% will scale the notes to a 7th of the root note (music theory, go look it up).

Application:

For the exact purpose of this example, I transcribed a small piece of Soleily's Violet Soul piano breakdown (no grace notes or anything special to ensure that all notes are CLEAR), located at this puush link here: http://puu.sh/hvEV5/93c957c9cf.mp3 -> the melody is as follows:



Note that the first note of this sequence is F, at the 7th octave --> a frequency of 2794 Hz; we don't need to focus on the rest of the notes because we are going to be scaling ALL notes down by the same proportion. We can apply the same two examples as earlier: cutting the speed/frequency by 25% and 50% for the entire melody.

At a start of 7th octave F, referred to from now as F7 (2794 Hz), the first note's frequencies will be as follows:

- 50% of 2794 Hz = 1397 Hz; the frequency of 1397 Hz is known to be equal to the same musical note identified as F6, or exactly one octave below. If you cut the the tempo and pitch by 50%, the notes used will be the same as this version, but played 50% slower.

- 25% of 2794 Hz = 698.5 Hz; the frequency of 698.5 Hz is known to be equal to the same musical note identified as F5, or exactly two octaves below the original note, and one octave below the notes resulting in a 50% speed cut. The notes used will be the same as this version/, but played 25% slower than the original, and 50% slower than the first cut.

I preserved the original speeds to allow for users testing this theory to compare the tonality of the mathematically adjusted version they will inevitably create, just to indicate that there is a difference in pitch. Me preserving original tempo allows me to prove that I already know what the pitches will be after the math is applied.

There is absolutely no reason to keep osu!'s current method of speed reduction, because pitch can be reasonably preserved with this concept iin such a way that pattern construction in pitch-sensitive mapping will not change because the root notes of the patterns themselves do not change, even though the frequencies do. The octave of given note absolutely does not matter given that there are no game modes were we can facilitate a need to map more than 12 notes relevantly at the same time (there are 12 notes within a given octave of notes).
Bara-

Halogen- wrote:

going to bump because this is a feature that needs to happen, it will benefit all mappers It does not. How will it help people when they are strongly annoyed by it? Yeah, it does not

also going to shoot down the below post, because if you're going to have a negative reason for something, you need to make sure you're right:

baraatje123 wrote:

Back to my point for Taiko mappers
This should never happen
<<Some tekst which was incorrect>>

Also, it would just sound terrible if you use it to check a map/mod
I strongly disagree with this, and I'll try to do as much to ever prevent this from happening
- Taiko is nowhere near as pitch-intensive as osu!mania; if mapping for pitch is improved in osu!mania, it will most certainly be improved in Taiko Seriously? I honestly never knew that. Good to know tho

- You absolutely should not be advocating against this idea when those in support of it actually have musical ears/backgrounds and understand differences in tonality and you admit that you don't at least know how to "call the pitch of a number".
I already said I made a mistake, because I thought (due to personal experience with Audacity, which then apparently didn't raise percentally, like I thought it did, as when I hightened pitch, it SEEMED that the higher pitched notes were much higher then the lower pitched note, then they were before, but I guess that only seemed so. Due to that, I thought this was the case. Also, I DO know "how to call the pitch of a number", which is hz, but I do NOT know what values are appropriate for music, which is why I said that, and used completely random values

<<Some tekst which was incorrect>>
This does not work at all in practice. You cannot arbitrarily say that a note is tuned at X frequency because musical notes have fixed and finite frequencies which increase by a multiple of 2. Additionally, numerical difference in frequencies are not directly correlative in difficulty to separate reasonable tones -- it will be equally easy to discern a 7th octave A at 3520 hz and 7th octave A# at 3729 hz (difference: 209 hz) as it will be to discern a 3rd octave A at 220 hz and 3rd octave A# at 233.1 Hz (difference: 13.1 hz, or < 1/10 the previous difference). Because your argument relies on the "separation" of "note tuning" and doesn't account for no difficulty change in discerning notes within audibility, it's completely invalid. I know already because Shoegazer explained That I should read you previous post, which I before my post didn't honestly read completely, only skimmed Also, these numbers nearly give me a headache, which such big numbers >_< holiday is bad for the brains xD
Thanks for replying, but there was absolutely no need to go into my incorrect point of view, which I already knew was incorrect due to others.
It also sounded a little harsh ;-;
But still, subjectively, I still have an opinion, that it would just ruin the song and sound bad. That's an opinion you can't shoot, as it's not a factual opinion but a... uhm... a regular opinion.
I just dislike having lower pitches. Take a look at the DSi's and the 3DS's Music player.

Going up increases the pitch, going down decreases the pitch
Going right makes it go faster, going left makes it go slower
Bottom left, is probably this feature at 33%, and believe me. It sounds Terrible
Higher pitched slow, is also terrible. Normal sound slow, sounds good. With fast, High and normal sound good, while low still feels terrible, but quite feasible here though

Also, I love it, how most of the people who respond here negatively are active mappers (with the exception of mania mappers and the OP), which already shows most mappers strongly disagree, and that I'm not the only one who does.
Due to reason stated above, I think Taiko and Mania are fine by this, but Std/CtB are not, because they follow rhythm, which at a lower pitch can be harder to hear
Kurokami
Though I already said I'm not against this, actually I think what you can't hear normally that shouldn't be mapped either. Players just want to enjoy good maps. If you try to force player to follow something which they can't hear, that makes the mapping bad. Still, in some really rare case, this feature can be useful.
lolcubes
The function of the 25/50/75% playback is for the mapper to notice certain music elements better and to get a better timing/synchronization. Pitch change can only make that more difficult. If you need a pitch change to map something, you are probably not doing it right.
osu! isn't a music player, it's a game.
-Maus-

baraatje123 wrote:

post
Well, we're not speaking about full octaves here.
PyaKura

lolcubes wrote:

The function of the 25/50/75% playback is for the mapper to notice certain music elements better and to get a better timing/synchronization. Pitch change can only make that more difficult. If you need a pitch change to map something, you are probably not doing it right.
You... What ? No offence but it's as though you didn't read anything of what's been posted before.



While I do agree that altering pitch makes songs sound horrible it does help to notice some beats and sounds which you'd miss at normal pitch. This is not an issue as the result will be heard and played at 100% speed so I couldn't care less. I'm talking from an o!m mapper point of view.
Bara-
Perhaps only add this for the O!M editor
The only mappers here who agree are O!M mappers (and OP)
lolcubes

PyaKura wrote:

While I do agree that altering pitch makes songs sound horrible it does help to notice some beats and sounds which you'd miss at normal pitch.
This is exactly what's wrong. If you need pitch adjustment to hear a sound you either have a very terrible sound equipment or that sound probably shouldn't be mapped cause you don't hear it normally.

PyaKura wrote:

This is not an issue as the result will be heard and played at 100% speed so I couldn't care less.
How is that not an issue? It's making things difficult to pretty much everyone except for you in this case.Since you don't even care about others, why should everyone care about this then?
Full Tablet

lolcubes wrote:

PyaKura wrote:

While I do agree that altering pitch makes songs sound horrible it does help to notice some beats and sounds which you'd miss at normal pitch.
This is exactly what's wrong. If you need pitch adjustment to hear a sound you either have a very terrible sound equipment or that sound probably shouldn't be mapped cause you don't hear it normally.
I don't think the purpose of using a pitch-changing playback speed algorithm is helping notice sounds a mapper otherwise wouldn't hear.

Current algorithm (pitch-preserving): The music sounds more similar to the original song, but pitch preserving algorithms have a risk of introducing sound artifacts (the current algorithm is not bad in that regard as far as I have heard, but there might be some edge cases; the worst case I recall is sounds that are highly transient, such as percussion sounds)

Pitch-altering algorithm: The music sounds considerably different, but there is no risk of introducing sound artifacts.

Edit: an example that shows how can osu! algorithm fail.

Original soundfile (mostly silent with a loud transient sound in the middle): http://a.pomf.se/tqwjvl.wav
osu! algorithm at 25% playback speed: http://a.pomf.se/nxbrnx.wav (recorded directly from game audio, silent parts partly cropped)
Pitch-altering algorithm (25% playback speed): http://a.pomf.se/cchuxa.wav
PyaKura
@lolcubes : mania allows for heavy pattern layering. If I want to map something hard and end up using those almost unhearable sounds for the sake of it I will do so because these sounds exist and can be justified as not being overmapped as these sounds, not matter how hard they are to notice simply exist.

This will most likely be helpful for mania mappers only due to the gameplay and nothing else. I do understand the other modes' mappers who see this as not being useful at all as you're going for a clear melody or beat 99% of the time. This is just how it is. If I can find some "hidden" mapping material in a song I'm certain to try to find a use for it.
Redon
ziin
So it's annoying, but you can achieve exactly what you want by doing this:

Start osu! with a removable audio device such as a bluetooth speaker or usb sound card. Change the playback rate to the desired amount and play the song. While playing, remove the audio device. Osu should automatically pickup your next default audio device (usually onboard audio) and it will be playing without the audio pitch change. As long as you don't change this playback speed, it will stay at that new super low pitch.
BilliumMoto
To everyone saying that changing the pitch and the speed simultaneously would sound horrible because it doesn't work with a DS, please educate yourself as to how sound works.

If you want a short handwavy explanation, read Halogen's post. This is a bit more technical if you don't believe him.
In short, all audio is just a waveform, it's a position on a single axis as a function of time. A graph where time is the X axis. But sound is not a function you can represent with an equation, so the computer instead simply takes points along every 'n' interval on the time axis, commonly 1/44100ths of a second. These points are then stored in a file known as an mp3 or a wav.
In order to reproduce sound, the computer takes the data points from the file, and plots them back on a new graph. It then scrolls through the points at a speed of 44100 points per second, which then fully reproduces the original sound.

The current granular algorithms to slow down the music is a problem because it results in major distortion of tone and loss of transients, which is a direct result of trying to process the sound in a way that keeps the pitch the same, which is a lot like trying to force a circular rod into a triangular hole. The largest problem with granular stretching is that the transients are lost since matching these transients is exactly what we are trying to do as charters.

What are transients you ask?
TRANSIENTS ARE THE REASON WHY RHYTHMS EXIST. If you distort or lose the transients, you distort or lose the rhythm. Simple as that.

The granular algorithm abraker used in his demonstration is a much better algorithm that what the current editor has. Unfortunately, it's very computationally taxing to do in real time and it still greatly distorts the tone and transients.

The OP's suggestion here is instead of using a badly written algorithm designed to transform and distort the sound data, we simply play the original data back at a lower rate than 44100.
So for 50% speed, we play the data back at 22050 points per second.
This will result in TWO THINGS:
1) The music will be half as fast.
2) The pitch will automatically be taken an octave lower. This is due to the fact that frequencies and tones are nothing more than repeated patterns of points in the data, and slowing the data down will also slow down the patterns, causing the halving of the frequency and thus lowering the pitch of everything in the music by an octave.
But most importantly, the transients will be preserved. This is why we want to slow down the music to begin with: to hear the transients better.

This is why lolcube's argument is painful for me to read. It demonstrates an inherent ignorance of how sound works.
Thanks for reading, hopefully at least people have some idea as to what OP is trying to suggest now.

tl;dr lolcube is wrong, please understand what OP is trying to say and realize it's a great idea which should be implemented.
Bara-
1.) There is already confirmation here about how pitch works. Poor you, for typing such long story
2.) This'll sound so firetrucking annoying I'll do anything to prevent this unless it'll be a toggle
Why did you have to bring this back up, such a terrible idea
BilliumMoto
Because granular stretching sounds worse
and leads to charting rhythms incorrectly
Bara-
You do know this is purely subjective?
I really hate low pitch switches
I love high pitches, as in Nightcore
But I HATE low pitches, especially at slow speed
BilliumMoto
Transients and rhythms don't care whether or now you like the way they sound. They're still there, and you can hear them better this way.
But you're right, mappers should have the choice to use the granular algorithm if they so choose. Just be aware you're shooting yourself in the foot by doing so.
Bara-
Why?
Many mappers have achieved great results without the 'new' speed
Especially for standard/ctb
It might be so for Taiko/Mania
BilliumMoto
My experience with mapping is very mania orientated. And let me tell you, a lot of us mania mappers use many tools outside of the osu editor to help us achieve the great results. Advanced granular stretching (abraker's post) and slowing the music down like I described using other programs, as well as a waveform and spectral visual aid are among a few of these tools.

You're right, it's much less important for ctb and standard to follow the music's rhythm very accurately, but that's no reason to make Taiko and Mania mappers suffer.
roufou
make a toggle I guess, I imagine this will help a lot but some people seem to be religious to keep it the way it is.

in the current editor it can get really hard to figure the proper pitch for a song and noises don't sound accurate to original speed, and this sounds like a solution.

seriously sometimes I need to map at 50% speed opposed to 25% cause I can't hear the song properly at 25%
abraker

BilliumMoto wrote:

You're right, it's much less important for ctb and standard to follow the music's rhythm very accurately, but that's no reason to make Taiko and Mania mappers suffer.
I agree. While not a taiko player, in mania you have to respond to the tiniest elements within music from around 4.7 stars and up. Degrading that element, whatever it may be, that is 100ms or so in length to utter crap is just not right.
Starry-
I'm still supporting this as it will be a huge help to a majority of mania mappers. I shouldn't be having to switch to audacity to slow a part of an mp3 down using pitch shift just to hear it's snap. And no, in mania you are required to have the right snap on anything when mapping - you aren't overmapping at all by doing so.

I honestly don't see what the fuss is with everyone saying they hate the song when slowed down in pitch shift. If necessary a toggle can be added but 'listening to nice music when trying to find correct snaps' should be the least of your concerns when making a beatmap. It's quite a shame people are using this reason for the basis of their argument against this feature request.
Halogen-

Starry- wrote:

I'm still supporting this as it will be a huge help to a majority of mania mappers. I shouldn't be having to switch to audacity to slow a part of an mp3 down using pitch shift just to hear it's snap. And no, in mania you are required to have the right snap on anything when mapping - you aren't overmapping at all by doing so.

I honestly don't see what the fuss is with everyone saying they hate the song when slowed down in pitch shift. If necessary a toggle can be added but 'listening to nice music when trying to find correct snaps' should be the least of your concerns when making a beatmap. It's quite a shame people are using this reason for the basis of their argument against this feature request.
Thank you, I mean -- it's nice to see someone who is an active QAT for the part of the game that requires the most detail in beatmapping in support of this idea. It's equally as important to realize that there are multiple proficient users, including Starry-, who again, is a QAT who are having to utilize external tools.

Those who are in support of this are utilizing objectivity to explain their situation; those who are not in support of this are being completely subjective. For those of you who are talking about "listening to the song", you do realize that you are only slowing the song down for yourself to ensure correctness in your beatmap, right? This has nothing to do with the enjoyment of the song -- if you want to enjoy the song, you actually sit down and listen to it outside of osu!, or you play the song in osu! at the correct speed that it is intended at. The editor is only for the purpose of ensuring correctness, and nothing else.
Bara-
I strongly disagree
When I mod/map, and it plays at 25%, I don't want my ears to die because of horrible pitch
I know, technically, that's how it should be, but that doesn't mean people will like it
Yes, it can be useful for Taiko/Mania, but not for the other modes, and it'll only be annoying as heck to listen to, make ensuring correctness even harder
show more
Please sign in to reply.

New reply