Mod Archive Forums Mod Archive Forums
Advanced search  

News:

Please note: Your main modarchive.org account will not work here, you must create a forum account to post on the forums.

Pages: [1]   Go Down

Author Topic: A possibly really dumb question about drum loops.  (Read 27138 times)

0 Members and 2 Guests are viewing this topic.

an0va

  • New User
  • Offline Offline
  • Posts: 11
    • View Profile
    • an0va on Facebook
A possibly really dumb question about drum loops.
« on: August 18, 2011, 02:19:17 »

So I really want to start working more with breaks and 9xx madness and what have you. So I'm viewing a bunch of .xms and I usually see that people have a whole drum loop loaded as a sample into one of the instruments. Okay, that's fine, but I usually notice that the tempo of the drum loop is synced perfectly to the tempo of MilkyTracker. How does one go about doing this?

I read that there's no sync or time stretching feature in Milky, so what methods do you use to allign your drum loops?


(also, I'm well aware of the beatslicing each hit to a different instrument method. But that's not what I'm looking for with this here)
« Last Edit: August 18, 2011, 03:03:31 by an0va »
Logged
http:/facebook.com/an0va

raina

  • Milkytracker Moderator
  • Great Poster
  • ***
  • Offline Offline
  • Posts: 719
    • View Profile
    • raina.kapsi.fi
Re: A possibly really dumb question about drum loops.
« Reply #1 on: August 18, 2011, 08:31:52 »

You adjust the sample pitch (note the finetune slider in the instrument editor) and the song spd/bpm until it's good. Sometimes it doesn't work because the "natural" sample and song speeds are too far off. Not wanting to sacrifice your song, that's when you try other samples.

For 9xx the traditional approach is trial and error. Try different values and observe where the sample starts off playing in the sample editor to give you a visual hint on which way to go. Many trackers scale the 9xx scale to cover the length of the sample but in .XM it is fixed to 256 byte increments. See http://milkytracker.org/docs/MilkyTracker.html#fx9xx. Using the resampler to get exactly 0x10000 sample length is tedious because the dialog preview doesn't reflect the changes accurately enough when you directly input the frequency for example. But again, with trial and error, it can be done.
« Last Edit: August 18, 2011, 08:34:25 by raina »
Logged

an0va

  • New User
  • Offline Offline
  • Posts: 11
    • View Profile
    • an0va on Facebook
Re: A possibly really dumb question about drum loops.
« Reply #2 on: August 19, 2011, 01:33:48 »

Thank you for the quick and informative reply!

What is the relationship between sample speed and pitches?
For example, if I were to tune a drum loop at C-4 to play for an even four beats, would this now mean that C-5 would play two beats (doubling the speed evenly) and C-4 would play eight (halving the speed)?
Logged
http:/facebook.com/an0va

8ch

  • Newbie
  • Offline Offline
  • Posts: 52
  • A#4
    • View Profile
Re: A possibly really dumb question about drum loops.
« Reply #3 on: August 20, 2011, 16:29:13 »

For example, if I were to tune a drum loop at C-4 to play for an even four beats, would this now mean that C-5 would play two beats (doubling the speed evenly) and C-4 would play eight (halving the speed)?

that's right. except for C-6 playing eight, not C-4...
it is generally a good idea to make the loops as small as possible. 16 steps (or beats, as you say) is a good average. better 8 steps. using the resampler trick as raina described is very useful. i usually resample to values just below 10000 (don't want to go too crazy about this). if, after finetuning, resampling and finetuning again your loop just don't want to fit in, sometimes funktempo could be a solution.
Logged

an0va

  • New User
  • Offline Offline
  • Posts: 11
    • View Profile
    • an0va on Facebook
Re: A possibly really dumb question about drum loops.
« Reply #4 on: August 21, 2011, 02:52:08 »

that's right. except for C-6 playing eight, not C-4...
it is generally a good idea to make the loops as small as possible. 16 steps (or beats, as you say) is a good average. better 8 steps. using the resampler trick as raina described is very useful. i usually resample to values just below 10000 (don't want to go too crazy about this). if, after finetuning, resampling and finetuning again your loop just don't want to fit in, sometimes funktempo could be a solution.

Wait, don't we both mean C-3?


Also, resampling and funk tempo-are these natuve features within milky?
Logged
http:/facebook.com/an0va

8ch

  • Newbie
  • Offline Offline
  • Posts: 52
  • A#4
    • View Profile
Re: A possibly really dumb question about drum loops.
« Reply #5 on: August 22, 2011, 00:29:59 »

oh sorry, of course, you're right again ;)

please take a look at the attached example. you can see a typical 16 steps loop is loaded into intrument 1. the loop is in 44.1 khz, mono, 16 bit. milkytracker automatically finetuned this sample to "-28" and adjusted the relative note to F-6 (+29). this makes C-4 the "base-note" or default note. that means you play the original loop in it's original speed/pitch (so to say) by entering a C-4 note. C-3 would half the pitch (making the loop play twice as long) while C-5 doubles the pitch (making the loop play twice as fast as a C-4 note)...

you can see that the sample-size for instument 1 is 1B6C0 in hex (roughly 220 kb). you can see this in the bottom right corner in the sample editor where it says "length". a tiny "h" is printed there. you can hit the "h" to cycle through hexadecimal, decimal and time display. but let's stick to hex in this case (make sure the "h" is visible).

leaving the size as it is won't allow you to use full-scale 9xx effects here (as descibed above). in order to use full-scale 9xx effects you need to "shrink" the sample-size to 64 kb (or 10000 in hex). the resampler (which is a native feature in milky ;) will do the job. simply right-click into the waveform, select "advanced" and "resampler..." and you're in the resampler dialog. now adjust the relative-note value (or any other) until the "new size" value is below 10000. something like FFDF or so. FFFF is hard to get but would be the optimal size. just try and see what you get and remember there's an undo button in the sample editor ;) hit OK and you're done.

ok, now back to the example loop. the loop's bpm is somewhere around 94. you can adjust the song's bpm value to 94 and notice that it is almost perfectly looping without further finetuning and all that. also try 188 bpm and 47 bpm (i did that for you in the example file).

while instument 1 is the original, unaltered sample, instument number 2 is the resampled one (FFDF). a 96E command on instument 1 would play the loop starting with the "snare". you can see that in the example pattern. the same 96E command played on instument 2 would start from somewhere near the second "drum" in the loop. so you cannot use the same 9XX commands for both instuments, because instrument 1's size is above the range of the 9XX command. a 9FF command (the maximum) on instument 1 will leave you somewhere 3/4 far in the loop while a 9FF on instument 2 is useless because the sample size is just below FFFF hex (10000 in decimal).

it is very important to do all the resampling and finetuning BEFORE you start entering 9XX commands.

now, with the resampled instument 2 you have full control over each of the 256 possible offset values (9XX). ok, i know 2 values are missing in this case ;) in reality you don't need to be that precise. normally you would just need to know which offset is the "kick", "snare" and "hihat" (depending on what kind of sample you want to use, but i'm talking about drum loops here).

this is a very simple example. the next step would be to adjust the pitch of the loop to your song. let's say your song is in 125 bpm and the loop is in 94 bpm. in this example your "base-note" would then be F-4 in order to make a smooth loop (more or less)...

you see it is acctually quite easy (but difficult to explain). when entering 9XX commands it is a good idea to have the sample editor open so that you can see which value triggers which offset in the sample. place your cursor on the note and use the ALT-SPACE shortcut (press and hold). once you release the keycombo you'll be thrown back to where you came from. once you got the "right" offset values it is just a matter of copy and paste...

"funktempo" is sometimes called "shuffle" in other (commercial) programs. it is not a feature but a trick that trackers sometimes use to create a dynamic shuffle rhythm. it is easily done by placing a couple of FXX commands on every row in the pattern like this: F07, F03, F07, F03, F07 and so on. this would produce some kind of reggae-style emphasis on every even row. if, for example, your loop plays just a little longer than 94 bpm, but too short for 95 bpm you could try 94.5 bpm using funktempo.

« Last Edit: August 22, 2011, 00:36:52 by 8ch »
Logged

an0va

  • New User
  • Offline Offline
  • Posts: 11
    • View Profile
    • an0va on Facebook
Re: A possibly really dumb question about drum loops.
« Reply #6 on: August 22, 2011, 03:35:50 »

8ch, that post was brilliant. Sticky worthy IMO as well. Everything was explained so perfectly and the example was awesome and very easy to follow, thank you!!!  :)

One last question: Say you have a bunch of drum loops now that you have cropped and what not to make loop correctly-before we begin into resampling and what have you, what is a good suggested method to find out that loops original BPM? Trial and error in the tracker till you find a BPM where it loops accurately? I could just dump it in Ableton to see what it warps the tempo as, but that seems like it could be more steps than necessary.
Logged
http:/facebook.com/an0va

raina

  • Milkytracker Moderator
  • Great Poster
  • ***
  • Offline Offline
  • Posts: 719
    • View Profile
    • raina.kapsi.fi
Re: A possibly really dumb question about drum loops.
« Reply #7 on: August 22, 2011, 09:01:37 »

You could use an external app to find the sample BPM but since the BPM value in the tracker isn't related to real world BPM, the results won't be perfect. You can get close by using the Spd value of 3, 6, 12... but it'll get out of sync before you can say "goddamned obscure legacy software". So staying true to my schtick, Imma say trial and error.

an0va

  • New User
  • Offline Offline
  • Posts: 11
    • View Profile
    • an0va on Facebook
Re: A possibly really dumb question about drum loops.
« Reply #8 on: August 22, 2011, 09:07:47 »

Word, thank you so much guys!
Logged
http:/facebook.com/an0va

8ch

  • Newbie
  • Offline Offline
  • Posts: 52
  • A#4
    • View Profile
Re: A possibly really dumb question about drum loops.
« Reply #9 on: August 22, 2011, 19:53:45 »

glad to see it was helpful. though i made a mistake: the resampled instument's size is not 64 kb but 128 kb. 64 kb is for 8 bit samples...
Logged
Pages: [1]   Go Up