Originally Posted by
Chucky-X 
When I developed the Tera, I was thinking along the same lines:
Low impedance, direct coupled output with no capacitors between amp and headphones, so that the output can run straight down to DC (zero Hz). And a multi-bit DAC is capable of DC
So, I made Tera-Player prototypes which were running amps in parallel (to lower impedance just like on the Attraction DAC) and used an active virtual ground (8 OPamps in total) to eliminate the output capacitors. Also, there were no resistors in series to the output.
While the sound of this configuration was not
bad, it was not really good either. I had the sense of the music coming from inside my head, with lack of dispersion and soundstage. It also sounded kind of brutal, just like: 'see how great everything is under control'. The flow of the music was not good, it was sounding too controlled. It was as if the feedback of the amp was forcefully controlling the musical performance and it reminded me of my experiments when I did the SPLIF amplifier topology many years ago.
I was getting better sound quality from other Tera prototypes which had a much simpler output stage and were using output capacitors and I tuned the output resistance for best sound, as said earlier.
In order to
see a flat curve under load, you'd need of course a low output impedance. But that curve is a voltage only, and is not necessarily linear to the acoustic output of the monitor in terms of sound pressure.
Also, I thought and still think that many of those IEMs on the market are really short-lived, and most of them are made to do at least
something with an inferior input. In other words, make them work with mp3 players and smartphones, give more bass, be louder than the others. These things may not necessarily be designed to sound best with a stiff output.
Charles
