emusonacid! legacy

Hi there,

first I have to say I'm nowhere near my sampler but after experiencing the lack of different envelopes per zone on my MPC 4000 (which maybe will have to go for this reason at least) I made some thoughts about layering drums in the emu.

First of all one very big advantage of the emu over the mpc is that it is not limited to just 4 layers per key but instead I guess 128 voices (and its 128mb RAM in my case) and its possibility to set up different cords/filters/chorus/envelopes/LFOs/etc per voice/layer. Incredible as we're still speaking about a single preset.

There's just one thing the MPC has a nice trick for layering up drums: You can set a sample offset individually per zone (which equals to voices in emu-language). So you can set the first layer on a certain key to trigger immediately at the push of that key and the second layer to start a few (or a lot of) samples later. This is especially nice if samples aren't cut up perfectly or if you like/want that sloppy sound (which I do).

So actually it would be nice to combine this sample offset feature into the emu, but I have no parameter in mind which could do that. There is the possibility to alter the samplestartpoint (via midi for example) but you'd cut off the very beginning of the sound instead of moving the whole sample before the other one like it is on the MPC 4000.
So basically what the MPC does is adding silence in front of a layer if you shift it behind or the other layer forwards. How to do that in the Emu?
Well I think, as it is not possible to add silence before the sample with its parameters set in its voice (like above mentioned envelopes and filters and so on)
there might be a compromise.
We need a third sample which only consists of a few samples/ms of silence which will be pasted in front of our two individual drumsamples.
You then could use the "SStart" Cord Destination wired with a very small amount (which should give a higher resolution and we just need to scroll around the beginning of the transient) to a midicontroller. After finding a good offset for the samples via that midicontroller assigned to SStart, we'd alter the envelopes a bit and resample the whole new sample and cut off excess silence in the beginning of our new created sound.

Sounds fairly complex in the beginng but I think loading up predefined presets for that purpose might give that worklow quite a boost in time. Also with resampling internally and the quite good truncating algorithm, we'd not introduce much noise or a delayed sample.

Of course there is the possibility to use your sequencer and offset the two samples via two seperate midi tracks. This way you wouldn't have to to alter the envelopes after finding a good offset between the two midinotes. The disadvantage might be any sort of bad miditiming so you'd have to resample a few passes and picking the best one. Also you'd need to utilize another tool the sequencer but I think most of you guys are working with a DAW so this won't be a big drawback.

Maybe one of the variations of the possibility to do the sample offset thing with a midi sequencer might be the emus INTERNAL sequencer but I never even touched it so I'm not quite sure whether it is possible to shift midi notes in a very small amount. Maybe anyone of you guys might now this.

I'd love to hear your thoughts on this. Perhaps someone might try one of the three suggested workflows as I'll be able to touch my emu again in three weeks the earliest.

Cheers

Within Preset Edit you can set a "Delay" time from 0 to 10.000 milliseconds (= 0.001 - 10.0 seconds) in steps of 1 millisecond for all selected voices.
On the page "Voice Modifiers" you find the parameter "Delay".
AFAIK is the "Delay" parameter not available as destination in the cord settings.

Without transposition the (virtually) inserted silence before sample playback starts is about 1/1.000 of the sounds sampling rate.
So 28 samples at about 28 Khz, 44.1 samples at 44.1 Khz and 48 samples at 48 Khz.
I hope this is fine enough.

I don't really get what you're talking about in your second paragraph, do you mean the samplerate I'd have to apply to the silence or do you mean the samples being altered by that delay parameter? (Actually I really didn't know the emu got that...)

How accurate can you set that delay? 1ms increments? Is it possible to audition the layers while setting up that delay? pleeease try that for me :)
Try matching 2 different drumsounds, one cut up properly and the other one with to much in front of the transient.. if that delay parameter has enough resolution you should be able to align the drums almost perfectly

If you want a drum sound to come in later than the midi note on message you fire it with - use the Voice Modifiers / Delay page as mosrob pointed out. This has a very wide range in milliseconds. You can alter and listen to how this shifts the groove in real time.

If you want to have a sample play further into the waveform then you can use a cord as follows
source-------------------destination-------------------amount
CONTROLLER A----------SSTART-----------------------+100
You could use any controller such as velocity, mod wheel, or controllers A to L
If the hit does not start at the beginning of the sample this silence can be altered with a knob/slider until you are happy with the groove. Setting a value that goes beyond the hits start point can yield some good results too especially on snares.

You could build a template that uses 2 controllers per voice
eg In Group 1 S/Start assigned to controller A, and also set controller B to envelope decay.
If you copy this Group into the same Preset and change the Group number to G2 it would be possible to alter the sample, key group, C/C letters to set up another drum sound.
With 12 controllers it would be possible to do this 6 times.
Just remember to give each new sample its own unique Group number.
A Preset like this would allow 6 different drum hits to have start points, amp decay on separate knobs and perhaps all six could have the filter cutoff altered by the mod wheel simultaneously. Plus sample playback delay could be altered per hit as mosrob has explained.

This could be used to chop up a break by assigning the same break to different groups and using the controllers to find each hit point and then shape the tail of it.

Save your preset and use this as a template for new drum sounds/breaks

Another cool trick is to have the hit sound a few milliseconds before the timing of the beat. This sounds good on claps that hit slightly ahead of a snare. It would be possible to add some delay to the snare so it comes in late but this might wreck the feel you may want as it is perceived as slower, creating a more laid back groove.

Try this instead......

Have the snare hit on beats 2 and 4, then program the clap on beats 1 and 3
Use the delay feature of the sampler to move the clap a quarter note later than the programmed midi note on time.
By adjusting the timing of the delay you can move the clap to fall just before the snare hit.
This also sounds cool on a reverse clap.
When this technique is done well it really pushes the groove and turns an average snare into a WTF snare.

Of coarse this method can also be used on synth grooves to, and works great when combining a forward synth hit with a reverse synth hit as you are able to get some tight sequences running between the two.

Give it a try and you will see what I mean.

Hi Jahfunk, thanks for joining in.
I do understand the SStart cord, which is absolutely useful for triming sounds realtime (and then resampling it for example)

What you explained is quite different from what I actually wanted to achieve but the thoughts on the 1/3 clap delaying it so it just falls in front of a 2/4 snare is very interesting :)

Btw, I think you got the groupfeature a bit wrong. The grouping of voices is to alter many voices at once. You could alter voices themselves with only highlighting the V# string in the preset edit screen. So there actually is no need of setting different groups.

Damn I wish I had my emu with me :D

ja-ki wrote:Btw, I think you got the group feature a bit wrong. The grouping of voices is to alter many voices at once. You could alter voices themselves with only highlighting the V# string in the preset edit screen. So there actually is no need of setting different groups.

Yeh its part of my workflow, I always do this so I know what is happening in a preset when I come back to a some time later.
Seeing G1, G2 etc I have some sort of visual reference to the samples that have separate cord settings.
I know that I only have to do the group thing for multi-samples, but I find it useful as a reminder to how the preset was originally set up.
It is also easier to combine new samples into the same cord settings and have them shift up the list so they sit with existing sample/samples that use the same Group number.
It seems logical to me, but I am left handed and my wife thinks I do everything "kinda weird"

actually this is a pretty cool idea organising your presets this way :)

btw, I had another idea in accomplishing my first task:


adding 64 samples of silence in front of the transient, using the sstart cord with a very small amount of 1% to scroll around in the first few samples around the transient and using the delay option in the voice modifiers I should get a pretty nice resolution. 1 ms = 48samples delayresolution (I work always with 48khz, as I used to do a lot of video stuff, so I've gotten used to it) and the sstart cord should complement each other very well!

Okay, news: Finally fired my machine up today and had a little go. I must say it works perfectly, better than I expected.
The delay parameter is the key, also the idea of adding a few samples of silence before the individual drumhits for scrolling thru with the sstart cord works out great but isn't that necessary. Also I found out you don't have to use the sstart cord as you can set a sample start offset below the delay parameter. There has been so much been thought of when this software was written that it really is a shame E-MU no longer builds samplers. Imagine further development into this machines.

lol seriously. coincidentally i was just reading this sound on sound article before i saw your post. here's what it said in 2002 xP

"Dave Rossum and Scott Wedge have moved on from designing synths and samplers, and they have been replaced by a new generation of music technologists, who are poised to launch new ideas that combine the best of software and hardware. The first of these will be Emu's Emulator 5 software sampler later in 2002, offering 24-bit sample streaming and integration with the hardware Emulator Ultrasampler. This will be followed by further exciting combinations of software and hardware, and we hope to see new digital filters and better control surfaces — maybe even an SP2400 and a virtual Modular or Audity in 2003!"

that would have been cool. fortunately they made a pretty awesome machine as is

Hahahahaha... You know I just deleted the sentence about the E5 in my above post because I thought this would be bollocks. Didn't know there were such plans.
it's still quite powerful and I'm realizing more and more how powerful the cords are. It's a modular synth, seriously - PER VOICE!