emusonacid! legacy

Sorry if this is a double post, but I thought I may have originally posted in the wrong section...

I want to add primitive waves (sine, square, saw etc) to my sampler and then use its synthesis abilities to create new sounds.

I have a few questions on this?

1) I will create the waves in Wave Lab. Should I normalise them?

2) How many cycles should I use.....Would one looped cycle give a different sound to a number of cycles per sample?

3) Should I multi-sample the simple waves....Like say create a wave every octave or every other note?

My aim is to get as close to the raw set of sound other synthesizers use as building blocks.

Thanks for any help...Cheers.

Can't believe no one here knows about this kinda thang!

Of course you can do everything you listed. The results, however, are a bit more subjective. My recommendation would be to experiment with all of the options and determine with your own ear what sounds good. Additionally, you may want to try sampling from the test tones generated by your Ultra.

Actually, from memory, I believe there may actually be an issue with using really small sample sizes, though you can verify this yourself by searching the forum.

Congrats on the new Ultra.

ikechukwu wrote:Sorry if this is a double post, but I thought I may have originally posted in the wrong section...

I want to add primitive waves (sine, square, saw etc) to my sampler and then use its synthesis abilities to create new sounds.

I have a few questions on this?

1) I will create the waves in Wave Lab. Should I normalise them? Normalize them where? In wave lab? In the sampler. I would maybe normalize them in the sampler but really they are going to be so freaking loud that you will not be able to play your tones with full out put. I typically have to turn my patches down to 30 or 40 in the muli view.

2) How many cycles should I use.....Would one looped cycle give a different sound to a number of cycles per sample? You can use one if you want. you can use a million if you want. The sound will not change if you are sampling just your straight typical waves. They care a constant tone and thus looping one cycle or mutliple cycles really shouldn't make a difference. I loop one cycle.

3) Should I multi-sample the simple waves....Like say create a wave every octave or every other note? I am thinking yes on this one. See there are errors that can occure when the sampler interpolates the sample. Usually what it does is make the note you play sound higher then what it should. The problem with this idea though is your going to spend hours sampling and hours looping. Even with the loop calculator telling you how many samples to loop, you still have to make adjustments. At 440 hz you need to loop 100 samples to get a single cycle. but in reality you need to loop 100.22. I am going to resample all my test tones again and maybe sample them at a higher rate to see if I can get better loops. Because I am lazy, and I use my emu for breaks and bass synthesis, I am going to sample all my testones at low frequences, maybe around 55hz. I orginally sampled them at 440 hz and I kind of felt that my sounds didn't seem... sharp enough. I looked at my finished choon's bline in the spectrum analyzer and found that my overtone rolled off at 8k. The reason why is because I sampled them all at 440hz. As I pitch down, that window of harmonic content remains the same. It rolled off at 20k or whatever. so when I play a note at 50hz that roll off that would have been at 20k if the note was played at 440hz moves down along with the pitch. So sampling at 440hz only and interpolating that sample will only yeild about 10k of the actual sound. If I sample at 55hz, I will get about 19k of the actual sound. Which means my eqing of the patch in post perduction will be more successful because I will actually have harmonic content to work with. So I would say sample every other note. That is actually the best way to do it. Most high profile piano sample librays do this. it's the most accurate way. but I would say, if you are making a lead, sample at 440hz, if you are making a bass patch, I would sample as low as possible, say 55 hz, because that happens to coincede with the multiples that lead up to 440hz. One other thing I would have to say... You should sample the test tones in mono (sample just one side of the wave form from wave lab) then convert to mono. This will give you the appearance of a stereo sample... with out the polyphony being eaten up and I recently found that stereo to mono conversion on the emu is not the best choice in the matter. I wouldn't recommend doing stereo to mono conversions ever. See one side may run slightly behind the other and when you convert the sample you may end up losing over tones and such. You end up having phase cancellation issues. also, when you lay multiple wave forms that are stereo and you use the lowpass fitler. The filters become unstable do to the sampler dealing with the huge amount of polyphony. The fitlers losing their minds sounds alot like digital distortion. Loud nasty snaps and pops. So converting to mono is from a mono sample is best. Reduces polyphony and keeps the bird happy.

My aim is to get as close to the raw set of sound other synthesizers use as building blocks.

Thanks for any help...Cheers.

Of course we know about this.. I was just off line for a while. I have been doing this for years. Matter the fact, it's pretty much the only way I make sounds. I sample test tones and loop them and detune and filter and use the matrix mod and adsrs just like a real synth. No... refer to your quote and I will add answers like this

Thank you

ikechukwu wrote:
2) How many cycles should I use.....Would one looped cycle give a different sound to a number of cycles per sample?

If you want to work with the Modulation Chord -Option Sample-Cycle , to get PulseWide Modulation FX, ( like PWM Analog ), i think there is a limit in the EMU EOS of 13ms ( or 30ms ?- can?t exactly remember yet ) .

So i think that ?s a good generaly Time.

But in Mono !...it ?s the best what i found out...just try it out.

I have gotten the Pulsewidth modulation from using single cycle loops.

snh

I am wanting to use my ultra like this


i have been using wavefroms and multi sample of synth patches ive made

can any walk me through setting my ulta up like a synth

like the post say i want to be able to have a constant sine wave or whatever

mapped proper acros the keyborad

also

are there any good emu sample cds set up in this way ?

To make the Emu like a synth... you set it up just as if you where to use it as a sampler. All you do is get a waveform, loop a cycle or so of it, and then set the origin to A4 or A3 (if it is A4 or A3... depends on what freq you sample at) and then set the range to the max. You have a sine wave. Copy the voice to create a multi oscillator "synth."

You should have it good after doing that.

Any questions... just ask.

snh

When you are talking about "do it in mono" are you talking about mono sample? or a spesific mono voice allocation setting on the emu ultra?

Mono samples. This maintains a higher level of fidelity and also increases polyphoney. If you layer too many stereo samples together and start to work the filters in the emu, the sampler sort of starts to tweak out.

snh

thx snh



so it just tiny cycle like those found in cakewalk dimension on a loop



i always use mono for bass

I prepared a synth preset with cords and control change assigned to some parameters but when i save and reopen a patch, my previous real time settings are not kept and it depends on how are set knobs on my midicontroller, someone have and idea to fix it ?

thx

Have you tried to save from the multi view... you know the screen that the sampler defaults to when booted up initially?

Just a thought.

Hope this helps.

snh

Do you have the initial controller values you want saved on the Preset Globals page in the preset? Pity it will only store values for A-D controllers, not the full A-L