Emax

jammie

March 15, 2014 1:51 PM

never record hot they distort

in the emax i always set my samples to about -6db then use the normalise function in the emax that way you get a non clipped signal and the emax then maximises it

clipping in the digital relm is terrible and makes samples sound terrible

also its not good to have all your samples at 0db as when more than one is played at same time then the headroom is overloaded adding to clipping on the out

the emax has 8 poly so you need to devide 0db by 8 so that all 8 samples can be played at same time with out going over 0db

its the same with all samplers and why so many people get it wrong with mixing and levels

always mix low level so you can here the detail then when you put it loud you will still hear the detail

if you mix at high level when you drop the level a lot of the time you cant here the bass as you lowered it at high level becuase you thought it was to loud

when i mix and master i always try to get my levels to -12db for each mixer channel then when summed i try and get the master to about -6db

this then allows me in the mastering stage with a buss comp and m/s and eq to bring levels closer to the 0db range but i always try and get mine to -1to -2 db so that it can never clip any system

now you can layer 2 samples as a and b but they both have to use the same filter and vca settings if not they will reduce the 8 poly to 4 note poly

but you can detune them and pan them and add chorus with out losing poly

now your noise problem racks suffer noise more than keyboards due to the cable from the digiboard to the analog board

as some of the cables actually ligh over the power regs

and get induced noise

you can split the cables so that a ferrite bead vcan be fitted that stops the interference and should make all your outputs at the same noise level

also with sampling look at the frequency content and then double it for nyquist and sample at that rate first it saves memory and second if sampling something with max content of 5k then a 12khz sample will cover its spectrum no point in sampling at 44.1 as its only still going to catch the conent of 5k

the reason for 96khz and 192khz is actually for over sampling in dsp systems as by oversampling at these rates errors in the algo are deminished due to the oversampling

and why its done as it pushes the aliasing frequencies further away from the 0-20khz human ear range but this is for vsti and emulated synths and samplers in software

in the hardware there was brickwall filters put in place for this and why synths and samplers had what was called a playback engine

like the korgs right up until the trinity

they used a engine that had a play back rate of 64khz allowing sample playback rate of 32khz at a octave higher

32khz is a good sampling frequency as there are not many sounds that go above the 16khz mark

my asr10 samples i do in 22.5khz and they sound amazing as the voice chip makes every thing sound good it gives me double the sample time of 44.1 and even asr10 had a 30khz sample setting

so sample at -6db coming from your mixer max thats what i would do and the worst thing about the emax is that you cant monitor the actual sampled sound in the emax so you cant tell if its distorting until after you sampled it

i always find that on the emax and the dss1 it always samples to the same keys on the dss1 its c3 always then you can to assign it to the key you want

Carl Lofgren

March 15, 2014 7:55 AM

Recording too hot? Oh... I planned doing that. And with different rates as well. It's just that I would like to set up the optimal settings first. But no matter - I tried again a few hours using expensive shielded cables and practically measuring all outputs. The noise of the sampler when nothing is playing is highest on the right output, noticeably lower on the left and quite low on the individual outputs. When when sampling the noise/artifacts is pretty much the same give or take one or two dbs.

/C

Carl Lofgren

March 14, 2014 9:20 PM

I am not that used to its operation. And after three days trying to
understand how it works, I'm a little wiser but far from all-seeing.
It's really a quirky little machine with lots of oddness going on.

>From my part, this is obviously partly a case of RTFM - but I honestly
cannot stand the manual. Craig Anderton might be a legend, but I
honestly cannot get my head around it.

1) Individual outputs vs mix outputs. The individual outputs are way
less noisy. My Babyface registers about 10 decibels less noise - which
might/might not be the entire ruth since the levels are different
between the individual and mix outputs. I've been listening to my ears
almost start to bleed, but I cannot hear any difference in sound quality
(noise level yes) between let's say output 8 and the mix. How about you
guys? What's your experience?

2) Replacing samples. Can it be done? Let's say I've got a 18s sample
(max memory) and I would like to do another one. Master + 4 erase memory
would be one way, but that nukes some of my settings, such as selection
of output connector. Is there any way to sample another 18s and still
keeping the settings? When erasing the machine, maybe I shouldn't say
yes when the Emax asks to make a new default setting? (just occurred to me).

3) Getting the samples on the right keys. I know I am going to sound
like I am stupid, but the root position of the samples I make are not
constant. Sometimes they land on G1 - sometimes on B1(!). Can someone
please please just write down step by step how you do it to sample
something on C1.

4) Setting the levels. Oh man what I've been banging my head on this
one. I send a sinewave (normalized to -0.05 db) to calibrate the input
level. I set the sampling level in the Emax to 0 and I always have to
back off on my sound card before sampling. I never go all the way to the
right, but I settle one or two pixels from the left/max. This causes the
Emax to distort - very slightly. I could clearly hear it and even see a
small part of the falling top slope of the sine wave being straight. The
solution was to back off a tiny bit more - and the distortion
disappeared. How do you guys set the levels? I assume setting the Emax
input level to 0 is the optimal (it even says so in the manual). But I
haven't had the time to experiment with it.

5) Ok. Now it's really RTFM time. Primary / Secondary samples - how are
you supposed to work with that? To me it sounds like you can sample one
sample that says AAAA and another sample that says OOOOO and have them
playing together like a normal two oscillator synth. Can you detune
them? Set the levels? Different envelopes?

/C