Emax

Hi Nik and Esynthesis! 

Actually, I've tried all the possible combinations of 12bit dithering available in Ozone and I found that the noise it mixes into the sound is way louder than the Waves L1 in either moderate or ultra setting. The L1 does the same 12bit thing by the way and I found that even the moderate setting is sufficient for the samples I've tried them on. Unfortunately, when I try to apply it on 24/28 kHz samples the DAW says L1 won't support these sample rates so I still have to process them in 44.1 and convert afterwards, which brings back some of the annoying artifacts. 

 BUT!....I was so focused on this whole dithering/sample conversion issue that I forgot to try sampling into the Emax via the adc AFTER I applied the 12bit dithering to the 44.1 kHz samples -without converting them afterwards. I tried it now and it seems to be working pretty well, I can only recommend it! In fact my new recipe for sampling percussive sounds in 28kHz into the emax is this: keep the samples in the DAW in their original sampling rate, apply some heavy limiting (6-8 dB) to bring up the tail and overall noise floor, then use the 12bit moderate dithering of L1 (I like type1 for deep and type2 for high) and then sample into the Emax ADC, then give back some transients shaping the sample with the VCA envelopes. I found this method optimal for 28 kHz, but if I can afford 44 kHz, simply applying L1 12bit moderate, bouncing in 44.1 and letting EMXP do the conversion is the sh*t!

I have to admit, the ADC does give the samples some vibe though! I'll keep experimenting!

Thanks for your attention! 



--- In emax@yahoogroups.com, "esynthesist" <esynthesist@...> wrote:
>
> Yes I'm on this group ;)
> 
> Actually the codec is converting 16 bit directly into the 8 bit data as required by the Emax. The 8 bit audio data in the Emax is a non-linear conversion of the original higher resolution audio data.
> It uses a smart conversion algorithm (belonging to the u-law family as used in analog telephony. It tries to make the 8 bit audio data sound like "12-bit quality" instead of just linear "8-bit quality" as can be found in cheap samplers like the Ensoniq Mirage.
> 
> This compression algorithm is also applied within the Emax itself and by the Emax-II when it compresses/saves the bank into 8-bit. The same principle (but different algorithms) are also applied in the Emulator-II and the Emulator-I.
> The algorithm has been chosen by E-Mu in such a way that it gives the best results for the most common musical sounds.
> But the algorithm is weak when applied to some specific sounds, especially percussion sounds.
> You get similar problems when you sample such sounds on the Emax itself, although I have to admit that from time to time they sound somewhat better than the software conversion versions generated by Emax-II or EMXP.
> 
> E-Mu must have been very aware of the weakness of their compression technique for percussion sounds, because the SP-12/1200 drum computers are NOT using 8-bit compression on their sampled sounds. The drum computers store their samples in 12 bit resolution, which means sampling time was sacrificed in favour of sound quality.
> Which is a good choice, because percussion sounds are incredible on the SP-12 ! 
> 
> ///E-Synthesist 
> 
> --- In emax@yahoogroups.com, Nik Kinloch <nikkinloch@> wrote:
> >
> > The EMXP manual says:
> > 
> > "Resolution conversion: the bit depth of the sound data may be 
> > converted, e.g. from 16 bit to 12 bit. However, with the current sampler 
> > formats supported, this conversion only takes place when converting to 
> > Emulator-II and Emax-I. In these conversions a special codec is being 
> > used since the resolution conversion is not linear."
> > 
> > So it sounds like some conversion is happening, but from our experience 
> > it is not optimal.  Perhaps someone can ask Kris Van de Cappelle (is he 
> > on this group?) about this? Maybe it just needs a tweak. I hear the 
> > aliasing as well esp on low frequency sounds with a tail (for example, 
> > long bass drums).
> > 
> > izotope Ozone apparently can dither to 12 bits then save to 16 bit file 
> > (last 4 bits are zeros), this would seem ideal. Anyone tried it? 
> > http://www.izotope.com/support/help/ozone/pages/modules_dithering.htm
> > 
> > Nik
> > 
> > 
> > 
> > 
> > 
> > On 16/02/2012 15:34, zrennaya wrote:
> > >
> > > Hi!
> > >
> > > Thanks for your input! I gave up on the 12 bit render thing as you mentioned there's no tool that converts to 12 bits.
> > >
> > > On the other hand I experimented with some dithering types and I actually managed to reach the quality of factory samples in terms of noise and aliasing (using the L1). So far I couldn't get that quality from the emax ADC inputs!
> >
>