Delta Mix-e-Load....Here's a thought......

[Ziphoid]

I sure wanna have a go at this as well. It may help me get things rolling on the remixing side again, since I won't need to do a full remix straight off...

Any deadline?

[skitz]

Sounds ideal.

I would suggest a small change to the structure tho. Take the name of pack from the filename and move it to a folder instead.

bin\ - location of demo
bin\packs\name of pack\
bin\packs\name of pack\lead.zip
bin\packs\name of pack\bass.zip

WIll make it easier to determine what packs are installed plus the user needs only to pack the folder containing their work rather than to select file names.

and yea... if you supply the first versions - gives people something to work from (structure etc)

The LEAD, BASS and DRUM are all independent of each other so using this structure would not really work.

The lead-[name of pack].zip, bass-[name of pack].zip and drum-[name of pack].zip keeps everything seperated.

for example you wouldn't have bin\packs\funky\bass.zip as this doesn't exist in the standard list. My thought was that you would scan the "packs" directoy for lead-*.zip and list those in the LEAD section, bass-*.zip and list those in the BASS section and finally drum-*.zip for the DRUMS.

Does that make sense and did you under how the key works and how randomly pick one of the 7 sequences at ever 8 beat mark like the original demo does. I was thinking of just using a sequence number mapped to the actual key sequence. I am mapping them like this (1=G2, 2=G#2,3=A#2, 4=C3, 5=D3,6=D#3 and finally 7=F3) - so every 8 beats @ 114BPM you would need to generate a random number between 1-7 and pickup the selected LEAD and BASS loop based on this formular to play next loop. The drum-*.zip file is slightly different in that it only contains one loopfile for the drums and I will call that file drumloop.ogg or drumloop.wav).

Does that make sense? It should make more sense once I create the default packs with some random instrument choices (and they really are random for the moment just for testing!).

[Infamous]

Let me get me head around for a second, this working as a sort of mix and match type thing so someone at a later date could say do a whole set of new leads using the basic key given by skitz there so it doesnt sound out of tune and the bpm (114?) and then import it and make up their own stuff to the delta loader?

Cos if thats the case i'll just throw in a crap load of drums seeing as im quite good at that and leave the synthy stuff to everyone else (bar the other work im doing for it ofc).

[StuC_Ovine]

Ok Skitz - we'll roll with your structure and see how it rolls

Yea the key change I understand (just glad im not doing the music side of things) Yea a random number would be easy enough and like you say mapping 1-7 to the keys with maybe a bias towards not always changing the key so it doesnt sound too random.

I'll wait for the first pack get it working then release a quick beta version so people can join in creating.....

[skitz]

Let me get me head around for a second, this working as a sort of mix and match type thing so someone at a later date could say do a whole set of new leads using the basic key given by skitz there so it doesnt sound out of tune and the bpm (114?) and then import it and make up their own stuff to the delta loader?

Actually the basic key sequences (G2,G#2,A#2,C3,D3,D#3 and F3) are not mine but Rob's and as I said I really don't know the significant of these key notes apart from maybe those 7 simply work well as a random progression?

I think you misunderstand the role of these key notes though as both the lead and bass part keys appear to be are linked to those key notes so for example the Funky-G2 pattern is linked to Level43-G2 pattern and the intro just randomly picks the key note whenever you change the different Lead / Bass in the intro.

I would love for somebody to try and replicate the original 4 drum riffs (BONHAM, DEPECHE, MARCH and RUMBA) using some real drums as my attempts are pretty bad so far sounding like a Casio Drum machine You can find these 4 WAV files in the attachment I posted earlier in the thread.

So maybe you could start off with that?

[skitz]

So, I have done a quick edit for the original sequence packs which contain the following zip files:

Bass-BlahBlah.zip
Bass-Level43.zip
Bass-Nolans.zip
Bass-Spacial.zip
Bass-Terendy.zip
Bass-VLTone.zip
Drum-Bonham.zip
Drum-Depeche.zip
Drum-March.zip
Drum-Rumba.zip
Lead-Arpeggio.zip
Lead-Bounce.zip
Lead-Funky.zip
Lead-Mainshum.zip

The Bass-*.zip files contain the bass sequences from 1 to 7 (G2,G#2,A#2,C3,D3,D#3,F3 respectively)
The Lead-*.zip files contain the lead sequences from 1 to 7 (G2,G#2,A#2,C3,D3,D#3,F3 respectively)
The Drum-*.zip files contain just a single drumloop file called "Drumloop.wav".

I think we will definitely want to settle on the OGG file format as even these default packs weigh in at a whopping 66MB as WAV files. In theory it should not be an issue to use OGG but I thought I would post them as WAVs to allow you to test the engine first with clean PCM data and then maybe convert them to OGG yourself and try using the BASS Audio Library to stream them?
edit: I just realised I exported all the wavs at 24-bit as this is how I had left my export settings to in FLStudio so the wavs would be a lot smaller at 16-bit

I've also created a couple of example FLStudio projects and MP3 files again showing how these WAVs are structured so if you are still confused grab a copy of FLStudio and load up the projects (expanding the WAV files in the same directory first of course).

Ok, so here are the (very) temporary Basic Loop Packs:

http://www.djskitz.com/delta/Packs.rar

and here are a couple of example FLStudio projects showing how this all works

http://www.djskitz.com/delta/Examples.zip

[StuC_Ovine]

Ok - all the ZIPs are loading fine and playing - a couple of them sound like they are not looping 100% - not sure if its the player or the sample tbh

I'll package a version up tonight for testing.

Note:
Inside the zip, I first check for WAV, then OGG so upto the user which format they want to use. As its zipped the uncompressed WAV would be smaller in size, probably about the same as the OGG anyhoo without loss.

Also Ive put in a 75% chance of a key change between samples.

It now scans the PACKS folder and looks for BASS-*.zip, LEAD-*.zip and DRUM-*.zip, adds the

Be warned tho, the samples have to be cached as playing from the zip file would cause a slight delay as its unziped and decoded. As they are loaded into memory it will take a fair chunk of RAM if there are masses of PACKS

Id suggest only copying in the packs you actually require. The amount of RAM would depend on the quality of the samples of course.

[StuC_Ovine]

ive OGG'd the sample ones so far - 44100Hz 128kbps (0.4g) stereo

Down to 3MB instead of the > 60MB.

[skitz]

Ok - all the ZIPs are loading fine and playing - a couple of them sound like they are not looping 100% - not sure if its the player or the sample tbh

Actually you bring up a good point and although the samples should be effectively an 8 beat loop running @ 114BPM. For the test pack I thought that I had "wrapped" the loop meaning that the FX (Delay, Reverb, etc) get merged at the beginning of the sample which is why they all abruptly end but you can hear tailing FX at the start of the sample.

Ideally what we need is for the player to have an alternating player for the lead and bass so that the samples can have the tailing FX. For example:

Code: Select all

Time @114BPM:  [1-2-3-4-5-6-7-8[1-2-3-4-5-6-7-8][1-2-3-4-5-6-7-8][1-2-3-4-5-6-7-8][1-2.....
LEAD Player 1: [Lead1.WAV----------]            [Lead3.WAV-----------]
LEAD Player 2:                 [Lead2.WAV----------]             [Lead1.WAV-----------]
BASS Player 1: [Bass1.WAV----------]            [Bass3.WAV-----------]
BASS Player 2:                 [Bass2.WAV----------]             [Bass1.WAV-----------]

So as you can see your routine will need to keep time (114BPM) and if possible maybe fade out the out going sample by the second or third beat of the sample coming in. Does that make sense?
This way the Lead and Bass samples can have their tailing FX left in tact so the transition between sequences is not so harsh.

I don't think we need this on the drumloop because as long as we export them in a "wrapped" fashion (i.e. tailing fx get put at the front) then the transitions will not be so harsh.

Note: Inside the zip, I first check for WAV, then OGG so upto the user which format they want to use. As its zipped the uncompressed WAV would be smaller in size, probably about the same as the OGG anyhoo without loss.

Excellent and as I mentioned the only reason those WAV files were so huge is that I encoded them @24-Bit - my bad

Also Ive put in a 75% chance of a key change between samples.

Excellent!

It now scans the PACKS folder and looks for BASS-*.zip, LEAD-*.zip and DRUM-*.zip, adds them

Excellent!!

Be warned though, the samples have to be cached as playing from the zip file would cause a slight delay as its unziped and decoded. As they are loaded into memory it will take a fair chunk of RAM if there are masses of PACKS Id suggest only copying in the packs you actually require. The amount of RAM would depend on the quality of the samples of course.

How quickly can you unzip a pack and load it's samples into the cache? In theory the only time you need to perform this action is when the user changes the lead, bass or drum selection. If you listen to the original demo these actually only change after a given sequence is completed (i.e. when you switch from FUNKY to ARPEGGIO it doesn't actually change to ARPEGGIO until the current FUNKY sequence is complete). Again this all depends on the demo keeping time (114BPM) and how long it takes to unpack and load the samples but could you not simply raise a "flag" that the user wants to change the LEAD, BASS or DRUM and have a "FIFO" based caching system?

For example when first loaded the demo will only cache in MAINSHUM, BLAHBLAH and WHINE which are the defaults in the original. Now the user will immediately get really fucking annoyed with the WHINE and instantly switch to say RUMBA. At that point you immediately unzip and cache in the RUMBA Drumloop and then on the next sequence cycle you switch to this.

Depending on how long it takes to unzip and load the samples into cache would still mean there is a small chance that a user switches too close to the next sequence cycle so I guess you would need two flags - one to say the user has switched something and another to say that the unzip and load into cache of that change is completed and is ready for playing. Worst case it would mean the newly selected sequence would have to wait another 8beat cycle.

This way you would could maintain say 4 sample caches per part (lead, bass and drum) and just clear them out based on a First In First Out approach - is this possible?


I bet you thought this was going to be easy!

[trace]

I was just thinking of how we should make the leads , bass etc, if there is reverb and stuff on them. Using monphonic and no effects, that won't be as nice as it should be
Else we could add an fx part where we can add reverb and stuff to each channel, but I guess that will be toooo much? let's make it "simple" for the programmer

Your ideá is great skitz overlapping the wavs like that

EDIT: I made a quick test just for fun, the drums is own made

http://trace.ziphoid.com/WIP/test_mixeload.mp3

[skitz]

YAY Trace - you got it son!

Maybe we should all start out with trying to remix the original lead, bass and drums

So maybe we simply have the demo open with original SIDs for MAINSHUM, BLAHBLAH and WHINE (Annoying!).

Then we all do our own takes on the other leads, bass and drumriffs for example:

lead-funky.zip
lead-funky(skitz).zip
lead-funky(trace).zip
lead-funky(tomsk).zip
bass-level43.zip
bass-level43(skitz).zip
bass-level43(trace).zip
bass-level43(tomsk).zip
drum-rumba.zip
drum-rumba(skitz).zip
drum-rumba(trace).zip
drum-rumba(tomsk).zip

All of this obviously depends on you Stu and how feasible all of this is

[StuC_Ovine]

More than feesible.

Ive just implemented the cross fade of the 2 leads and 2 bass which is no problem. The problem I have got with the current samples are that im fading out the first lead then fading in the new lead, as the new lead is playing before the drum beat has wrapped, the 2 become out of scync. I presume I have the incorrect dae out/fade in setup ?

The zip/load is already done, the format you have specified there is ok - except I dont have the room for all those descriptions think I can get about 10 chars in its current form.

[StuC_Ovine]

I'll have to post a version.... I think ive got the timing correct but I dont have an ear for music like you guys

I'll upload (remember this is a WIP so theres no error checking etc)

[StuC_Ovine]

http://baa.ovine.net/d2.zip

Unzip all files and include the folder paths.


Use cursor keys left right to change which track you are altering, up down to select from the track list.


I think the fade is still knocking the timing out, but like I say I think im tone deaf :S

[trace]

For me the bass and the lead is starting sometimes starting alittle earlie and sometimes alittle late and sometimes on the spot, don't know why...
Anyway a great job so far *Tumbs up*