forum

Beatmapping - timing: by the number of beats

posted
Total Posts
28
This is a feature request. Feature requests can be voted up by supporters.
Current Priority: +0
Topic Starter
elie520
Edit: Start of the real conversation : https://osu.ppy.sh/forum/t/654396&start=18

Hello,

I'm totally new to beatmapping, so I hope the feature I ask doesn't already exist or I'll look like a fool (but i'll learn something :D)
Picture a map with a lot of tempo changes. For me, the most efficient way to time it would be likewise :
1) try to very precisely mark the time of some important notes (at the beginnning of a time changge, or every 16 beat, etc...),
2) count the number of beat between each marked note
3) deduce the BPM in each section.

That's what I'm doing, and it's necessary in my opinion as soon as the song is played by humans not remaining exactly precise when playing. In 10 seconds, they can vary the bpm from 155 to 155.8. It's quite noticeable when you try to time perfectly a song.
But doing so implies that for every marked note, you have to deduce the bpm (by applying 60*(number of beats)/(time between the end and start note of section)), and it's quite painful, while it could be automatic, and only requires a new field "number of beats before next section" (maybe with a check box to indicate to osu! that it's the parameter that must remain fix?)
Furthermore, after a try, if you feel that you misplaced a note by maybe 10 or 5ms, then you have to compute again the bpm of the section before and after this note. An automatic BPM based on the number of notes would be instantly corrected, and it'd be WAY faster/easier to time a song.

So... what do you think ?

Thanks for your interest, I hope I'm understandable :s
Kondou-Shinichi
what feature did you actually want

for me timing is basically:
1. find the correct bpm
2. correct offset
3. you’re done

most music isnt that fucking variable. unless it’s a super horribly done and offbeat, it isn’t the case (piano pieces excluded)

not sure how does your method work, it just looks weird to count beats

variable timing is when basically you have a bpm change every beat lo
Topic Starter
elie520
Hello,

Thank you for your answer :)
While what you're saying might be true on some songs, it's almost impossible for a musician to keep a beat to the 1/100 second for 5 minutes. Especially in classical music or any kind of music with "respiration".
The thing is that almost all maps I looked at with the editor were mapped as you say, and the result is pretty terrible. If you play the song in the timing section, with a computer beep every beat, it never matches almost. That's why in a LOT of maps, it's almost impossible to have an UR less than 100/120 in HD (that is when you base your clicks on what you hear), and completely impossible to go lower than 80 for example.

I really mean no disrespect, but looking at your profile, I don't think you're according that much importance to the precision of the ticks. This is fine, everyone focusses on different aspects of the game.

Thanks :)

Edit: Btw, I'm trying to map a music played by an orchestra, that's why tempo varies a lot, sometimes voluntarily, sometimes not.
Saltssaumure
Not sure if this feature would be useful, as the mapper would still need to manually count the number of notes, overall not saving much time. Also, for orchestral music, to have it truly perfectly timed you'd have to change bpm every note due to human error in playing. And even if the map was perfectly timed, the player's UR would still be high as the player can't predict all the little bpm changes.


p.s. if this is your first beatmap, I strongly do not recommend variable bpm music like classical stuff, because it's a pain the ass as you rightly said. Here's a map of the 3rd movement of the Moonlight Sonata, and as you can see the timeline is jam packed with red lines.
https://osu.ppy.sh/b/1101611
Topic Starter
elie520
Thanks for your answer.
Indeed, you'd still have to count the number of notes, but that's a constant once you have it (and takes one listening with paper to take notes as the song progresses). However, if you change offset to a beat point by 10ms, the bpm before and after both change, and have to recompute both of them. In a 5 minute song, I have about 50 bpm changes, and what takes me the most time is to recompute bpm every 30 seconds to ensure that I set the starting note and the end note right.
You're pretty wrong about the BPM changes, because players play maps a lot when they want to train on it, or just love it. When you heard an interpretation a thousand times, you know the fluctuations, for sure.

Yes, and if the feature I'm asking was implemented, the composer of that said beatmap would have been pleased, I'm sure.
Endaris
By my personal experience, musicians, especially classical ones, are always playing some extremely light rubato in every single phrase.
That means even if you mark these important notes you have to time the ones inbetween separately as well.
It also has to be noted that many notes doesn't equal high bpm.
The bpm value directly affects how distance snap plays out and how fast the sliderball rolls and as such it does not determine the amount of notes but the actual tempo of the song. I know, kind of self-explanatory but for that exact reason I don't think your method is a good approach as it doesn't really care about that.
I definitely agree that the editor could need some kind of automatical bpm adjustment for variable bpm songs but that should only apply after the mapper made the call on how fast the song actually feels in that moment.
Kondou-Shinichi

elie520 wrote:

Hello,

Thank you for your answer :)
While what you're saying might be true on some songs, it's almost impossible for a musician to keep a beat to the 1/100 second for 5 minutes. Especially in classical music or any kind of music with "respiration".
The thing is that almost all maps I looked at with the editor were mapped as you say, and the result is pretty terrible. If you play the song in the timing section, with a computer beep every beat, it never matches almost. That's why in a LOT of maps, it's almost impossible to have an UR less than 100/120 in HD (that is when you base your clicks on what you hear), and completely impossible to go lower than 80 for example.

I really mean no disrespect, but looking at your profile, I don't think you're according that much importance to the precision of the ticks. This is fine, everyone focusses on different aspects of the game.

Thanks :)

Edit: Btw, I'm trying to map a music played by an orchestra, that's why tempo varies a lot, sometimes voluntarily, sometimes not.
not boasting but im p sure i’m better at timing than many ppl
You should said earlier youre timing a orchestra! saves a lot of misunderstanding
also, sometimes the timing may not be 100% accurate, but that’s ok. You won’t notice when playing (unless the timing is very bad)
there’s also those offbeat timing lo

elie520 wrote:

Thanks for your answer.
Indeed, you'd still have to count the number of notes, but that's a constant once you have it (and takes one listening with paper to take notes as the song progresses). However, if you change offset to a beat point by 10ms, the bpm before and after both change, and have to recompute both of them. In a 5 minute song, I have about 50 bpm changes, and what takes me the most time is to recompute bpm every 30 seconds to ensure that I set the starting note and the end note right.
You're pretty wrong about the BPM changes, because players play maps a lot when they want to train on it, or just love it. When you heard an interpretation a thousand times, you know the fluctuations, for sure.

Yes, and if the feature I'm asking was implemented, the composer of that said beatmap would have been pleased, I'm sure.
5 minute song, only having 50bpm changes? for a classial music perhaps that’s not enough lol
btw I’d fucking love to see a feature that do the timing for you, but variable... thats hard
Topic Starter
elie520
Thanks to both of you.

First of all, a solo performer will have a tendency to add a lot of rubato, true, but that doesn't mean that he/she will change the beat every second note. This becomes especially true with on orchestra, where rubato are less possible as a group. But in concerto for example, there may be moments when the soloist is alone, and enjoys his/her "freedom" with more rubato; while he/she won't be lingering freely on note when the orchestra is following.
I invite you to hear this song that I'm talkibng about https://www.youtube.com/watch?v=ZSOMXrdTGEE. I finished the timing with 74 sections, and I'm very happy with the result.
In the first 23 seconds for example, it's barely noticeable but the baet changes several times. You just can't keep one bpm, or it'l be ugly. I changed it 5 times in this period, with BPMs going from 154.589 to 155.844. There is no rubato, but still, a tempo irregularity.

To focus on the feature I ask for, it's very simple to implement, and to use, and won't disturb any one. It follows this way :
Instead of one "BPM" field, there would be one BPM field plus a "number of notes before next marked point". Moreover, there would be 2 check boxes, one for each field. And then, the mapper just checks the box he/she wants to be manually set. The other one is automatically computed. Simple as that :)

Cheers!
Endaris
The song you linked has nothing to do with orchestra performance or anything. It lacks the fluctuation of an actual artist performance.
It is music composed and played by a computer at the end of the day.
Compare it to something like https://osu.ppy.sh/s/92795 which fluctuates too much to be represented by a single 60bpm section although it is 60bpm throughout.
An actual instrumental song played turns out as an abomination like https://osu.ppy.sh/b/671988

For an actual orchestra it is not too much different compared to a solo performer as the character of the timing is still determined by one person alone: the conductor.

Also, the point I made still stands: many notes does not equal high bpm.
Topic Starter
elie520
Played by a computer ? Hmm no.
Your "point" I don't see why it's there, noone ever mentioned high BPM.
And you are mistaking obvious fluctuation (due to rubatos for example) with unvoluntary flunctuations due to the humans performing, who can't keep a bpm to the 0.1 because none on earth can (and it wouldn't even make sense to dishumanize the music this way).

So... I don't know what to tell you, I don't understand what you want me to say, I still stand by what I said, and that such a little feature would be a great tool.
Endaris
If you don't think it is played by a computer, then why does the artist say something like this?


If there are 5 notes in 1 second, it doesn't mean the actual speed of the song at that spot is 300bpm.
Depending on their meter, the actual rhythm and everything its actual bpm could be everything. It is not feasible to middle over short timeframes.
And if you have long timeframes, then why aren't you simple using the way of tapping to the beat as offered by the editor through the T key within the timing window?
Topic Starter
elie520
Lol, that has nothing to do MIDI or not. There are lots of musics played by people, turned into midi, because people like the retro look of it, and it's adaptable to lots of supports. It IN NO WAY shows it's computer played. And if you just listen, you know it's human played.

Y agree with you, high quantity of notes doesn't mean high BPM but again, who mentionned that ? I don't supporrt that claim. Please let us not deviate from the original subject, which is the useful feature.
I don't use the beat technique because it's way not procise enough on 5 second timeframes, and it's way more precise to do by hearing. I adjusted my timing tonight. I had to move some notes y 0.010 some times, thus changing the bpm by 0.200 or less, and it really matters, the difference is huge to the keen listener. It may not be noticeable to some people by hearing the song, but it'll definitely bring more quality to the gameplay.

Edit: By the way, here is what computer music sounds like when it comes to classical music : https://www.youtube.com/watch?v=cX2AbbEwTbs
Endaris
You act like it's not possible to manipulate sounds on their own instead of just letting a dumb program run it through a sheet.

Anyway - I don't have problem with having the option to automatically adjust BPM for such changes if desired by the mapper. But this doesn't change the fact that "timing by the number of beats" is nonsense in the way you ask for it.
That is what you get when you put 2 different feature requests in one thread.
Topic Starter
elie520
1) Thank you for your interest in my post,
2) there is only one request here,
3) You either didn't understand me (I have trouble making clear what I mean, so that could be the problem :( ), or you're wrong.
Endaris
_Meep_
I feel like this discussion is getting more complicated than timing the beatmap itself
Topic Starter
elie520
By the way Endaris you don't seem to be aware of that, but "beats per minute" basically means that you count the number of beat in one minute so... nonsense ?
It's basically what the autoBPM of osu does when you tap the beat, except that maybe it removes a few start and end beats if it makes the mean deviation too big (I'm not sure, idk how it's written but that's how I'd do it).
In any case, this is not appropriate in short time sections (like with maybe 16/24 beats or shorter) as it can't give you enough precision. However, if you spend a few minutes listening very carefully to the first and last beat of a section, you can precisely time it, and then be really procise on the BPM by counting the number of beats.

I agree _Meep_, and I deplore it. The reaquest is simple, but as apparently I'm not able to make myself clear enough in one post, I have to discuss with people "attacking" me to try and make myself understood. I still think that it'd be a great tool in timing for a lot of songs/musics.

Cheers :)

Edit: Endaris, if you could please not pollute the thread with your image, it's inappropriate.
Kondou-Shinichi
hey who WOULD count BEATS for a whole MINUTE for the bpm? LOOOOOL
can you clear up once again what do you want

btw electronic music is different from like irl played music
Topic Starter
elie520
Sorry for the obvious misunderstanding. I was just talking about "BPM" to point out that the primary method is counting the beats (which osu does for you in automatic mode) so that's not nonsense. Because a long as you're timing a short section (for example 5 ou 10 seconds), you can't rely on the precision of the auto BPM, which is of course pretty poor in this situation. Thus, you would indeed have to count yourself the beats.

Clarifying what I ask for :

In the timing panel, for each section, there are "offset" and "BPM" boxes. The idea is to replace the lonely "BPM" box with 2 boxes + 2 checkboxes. There would still be one "BPM" box, with its check box next to it. And the new box would be the number of beats before mext secion (if there is one, otherwise left blank), with its checkbox. The effect of the checkbox is that if the offset is changed, the checked box won't change value, but the other one will adjust automatically.
Example no. 1 : BPM box checked : If you change the offset, the bpm doesn't change, but then mechanically, the number of beats before the next section will change (increase if offset decreases and other way around in the other case).
Example no. 2 : Number of beats checked : If the offset is changed then the BPM will automatically ajust to keep the same number of beats before next section (which makes sense because usually if you barely change the offset, it's not to add more beats, but to adjust the beep sound to the beats). Then for example if the offset is increased, the BPM will decrease, and the other way around in the opposite case.

Is it clearer now ? Maybe I should have presented this like this earlier, instead of trying to motivate it.

Motivation : The point of this is that on a map hard to time, with a lot of short time sections when you can't rely on autoBPM, you will have to do trial and error by adjusting the offset often. With this new way, you won't have to recompute by hand the BPM every time.

Thanks for your interest.
Skylish
I take a general but not in depth look of the whole thread, your concept and idea of making a new timing panel is really innovative, but it is unnecessary to do so:

Ordinary music, which accounts for the majority of the music genres, have constant and stable BPM, so as offset. One single red timing point (uninherited timing point) is already fine in timing the song.

While, there are indeed some scenarios that the music contains multi-bpm with different offsets. They can be well fixed by a really careful ear and mind, with the consideration of musical theory which involves in bars counting (not beats conunting). Bars counting is already achieved in timing panel, in F2. X:Y is located at the top right corner. X represents the no. of bars, starting from 0; Y represents the beats of a bar.

In time signature 4/4, 4 beats= 1 bar. In time signature 3/4, 3 beats= 1 bar. In the expression of music, usually 3/4 songs have different bars cutting edges which make the common '4 bars conunting method' not working. For example, https://osu.ppy.sh/b/1242057&m=1

Shifting of offsets (+- 15ms, a general acceptable range for offset) does not interfere the bars distribution and BPM at all. It is not really a big worry while timing a multi- BPM music.

Timing does not really require any musical theories, common senses should work well in common cases. The main concept of an accurate timing is that: the timing is beneficial for a mapper to put their notes as accurate as possible meanwhile the gameplay and mapping effect (you will know it later if you are involved in slider placement across multi-BPM red timing points) would not be on expense.
show more
Please sign in to reply.

New reply