I'd like to reply to some comments in this thread, particularly to the
ones about SETs and NOS DACs. These are not in the same category as
electret cream, they are real life engineering tradeoffs. I'll give some
examples of both.

I'll start with the SET. I have  built MANY amplifiers over the years,
both solid state and tube, covering  many different architectures. They
all have different tradeoffs, doing one aspect better than others.
Different people will prefer different implementations, they really like
the  optimization of one and are willing to live with it's deficiencies.


The SET is an interesting case, from a sound perspective it does some
things extrordinarily well that some people absolutely  love, and does
other things poorly, some people are willing to live with the 
deficiencies to to get what it does well. Other people are not. Part of
the issue here is that what it does poorly is very easy to measure, and
what it does well is very difficult to measure (not impossible, just
difficult). It took a long time for the engineers to figure out how to
measure what it was doing well. Now that what an SET does well is mostly
known some designers have been able to come up with solid state circuits
that sound similar, but they still have a some of the same deficiencies.


In my case my current main amp is an SEP, Single Ended Pentode, it has a
similar good characteristics to an SET, but it also has vastly lower
distortion, less than  0.1% over the whole  power range (up to 25 watts)
and  frequency range. This has been possible by eventually understanding
what it is  about SETs  that make them sound  good, and then figuring
out how to keep that while improving other aspects. This turns out not
to be easy, improving one aspecct very often makes the other worse. It's
good old fashioned engineering  tradeoffs. Unfortunately it's not cheap,
there are about $3000 worth of parts in this thing and  it weighs  150
pounds, but boy does  it sound good when driving the speakers it was 
designed to drive. With $3000 worth of parts it would be very expensive
if it were commercially produced, and that  would NOT be exhorbitant
profits, it just plain costs a lot of money to get that level of
performance.  At least for now. Over time designers might figure out 
how to do it for a lot less money, but that hasn't happened yet.

On the DAC side, many years ago I (and others) noticed  an interesting
fact, that when you bypassed the internal digital filter in most DAC
chips things sounded better in some ways, and also worse  in others. 
Without the filter the sound was more musical,  more alive, more
realistic, BUT it also sounded "dirtier". Some people are willing to
live with the "dirtier" in order to get the  "more alive" sound. The
dirtier of course comes from the aliases, but nobody yet knows what  the
digital filters are doing that squashes the "alivenes". Again it's a
case of nobody  knows how to measure "the goodness" but it is very easy
to measure the  "badness".

I have spent the last 7 years trying to find out what it  is that  the
digital filters are doing that causes the problem. My currrent
understanding is that it is NOT the fact that it is a digitalfilter in 
general that is the problem, but  the  implementation that is used in
ALL DAC  chips that have a builtin filter. In  order to get extremely
good numbers in the spec sheets for certain parameters the designers
have  taken  to using complex filters in the chips. Getting these spec
sheet numbers with the traditional mathemetical function you read about
in  DSP textbooks takes a lot of processing  power which would
significantly increase the price of the chips. So they have come up with
DSP "tricks" to get those numbers, but now the filter function is much
more complex, and this "complexity" seems somehow responsible for the
sonic degradation. At this point I have no idea WHY this is, just that
it is. 

I have now built several DACs using external digital filters and DAC
chips that either don't have digital filters or it can be turned off.
When I use the basic simple function and give it enough processing power
(either in software on a computer or in an FPGA) the results are
amazing, you get the "aliveness" of the NOS DAC, but all the dirt is
gone, the result is stunning. 

Again the NOS DAC is a tradeoff, people willing to live with the
negative aspect to get the positive aspect. As we start to understand
what is going on it starts to get possible to do away with the tradeoff
and get both aspects done well. 

Neither the SET nor the NOS DAC are an attempt to bilk the public for
monitary gain, they are legitimate engineering tradeoffs that some
people are willing to take. Hopefully as we understand what is actually
going on we will be  able to do away with the tradeoffs, but that is not
fully there yet. 

John S.


------------------------------------------------------------------------
JohnSwenson's Profile: http://forums.slimdevices.com/member.php?userid=5974
View this thread: http://forums.slimdevices.com/showthread.php?t=99360

_______________________________________________
audiophiles mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/audiophiles

Reply via email to