>> I have tested just those samples (among many others) >> by sampling my XL-7 via the S/PDIF out with all >> engine components disabled or at nominal levels... >> >> the basic analog waveshapes are far below 0dbFS, so >> that is why I assumed there is a large degree of >> gain scaling occurring in the engine itself...even >> when boosting a layer via +10 gain, it is still below >> full scale.on 1/13/04 4:32 PM, Aaron Eppolito wrote: > Right, remember the samples are 16 bit while the output > is somewhere between 20 and 24 bits (can't remember). > Playing a single 16 bit waveform should theoretically > be much below 0dbFS, yet still retain its full 16 bit > resolution. nope... 0dbFS is the baseline, and it goes down from there... theoretically 96dB for the 16-bit, and 144dB for the 24-bit, but in either case, the first bit of the sample corresponds to the first 6dB of dynamic range... so regardless of the fact the samples are 16-bit, if they are at full scale and the digital output is not, the E-mu engine is doing gain scaling that cannot be overcome/compensated for by adjusting engine parameters. cheers, aeon
Message
Re: [xl7] Sample bit depth?
2004-01-13 by aeon