darTZeel Audio NHB-108 B version - Manual (Page 22)

darTZeel NHB-108 model one
User Manual, version B 1.0 Page 22 of 35
strumentation measurement, assuming this result
necessarily leads to outstanding sound purity. We
take this with some caution, since it recalls for us
the 70-80' Japanese philosophy, by which "per-
formance" implies "quality". We have to be fair by
acknowledging that such designs can be liked and
even loved by audiophiles. The world is big enough
for multi musical tastes.
The idea is very nice indeed, and we do admire
such a philosophical approach, even if it is not
ours: the problem is that those designs are highly
sophisticated – too much for our ears – involving
lots of added transistors, op amps, and higher than
reasonable NFB, leading finally to altered sound
structure.
These electronics can be considered as very pure
and detailed for the first listening hours or days,
but once noticed, their sonic signature becomes
harder and harder to bear.
So if THD does not seem to be a determinant factor
in the accuracy of the reproduced sound, we have to
look elsewhere.
T5.1.2. Inter Modulation Distortion
Now, let us listen to two flute players. The flute is
well known for its less rich than usual sound - har-
monically speaking, of course! A kind of exception
confirming the rule.
If each player produces a different, sustained note,
what will we hear? One intermediary note, two
distinct notes, or more than this?
In theory we should hear 2 distinctive notes. In
theory only? Well yes, because in practice our hear-
ing is not perfect. Truly wonderful indeed, but less
than perfect.
Without entering into details, we can nevertheless
say that in fact we will not only hear 2 distinctive
sounds, but also combinations of those primary
notes.
We beg musicians and music lovers to forgive us,
but for a while we need to replace notes by fre-
quencies. Not very musical, but much easier for the
purposes of our explanation.
If the first flute plays a 1,000 Hz tone, and the sec-
ond one a 2,500 Hz tone, we will not only hear
those two discrete tones, but also the following
combinations:
2,500-1,000=1,500 Hz, also called the beating fre-
quency, and also the mirroring part, say
2,500+1,000=3,500 Hz. Fortunately, those combi-
nations are of much less amplitude than the discrete
notes themselves. But that’s not all! We will also
hear the harmonics of these combinations! So, fre-
quencies of 1,500, 3,000, 4,500 Hz, but also 3,500,
7,000 and 10,500 Hz will be perceived. Once again,
their level will be very low compared to the 2 initial
notes played.
It seems hardly believable, but it is true… In prac-
tice, though, this is far less embarrassing than you
might think at first glance. As said above, the rela-
tive level of those "ghosts" is much lower than the
basic notes. The result will be heard as if the flutes
had some tremolo, or vibrating behaviour, a very
easily perceived phenomenon. And the combina-
tion’s harmonics will add some "warming" factor,
or on the contrary some "coldness" or "dryness" to
the perceived sound, depending on how they will
combine together.
So you can see that lMD is even less easy to under-
stand and quantify than THD. And how can we
appreciate IMD, can we hear it, at what level does it
become a nuisance?
Well, the first thing we can admit is that it is not
musical instruments that generate inter modulation
effects, but our own hearing – brain included – that
is the cause. Some acousticians say than our inter-
nal ear can "produce" inter modulation artefacts at
levels between 25 to 40%! Of course the brain then
does some necessary "correction".
However, this kind of "correction" cannot be com-
pared with an electronic chain, say the microphone
and analysis system. So, we feel very uncomfort-
able about advancing any IMD value from external
origin for which our ear is sensitive. In practice it
seems that IMD less than 2% (3 to 5% for THD) is
considered as not disturbing, even not audible in a
musical context, judging by the Single Ended Tri-
ode aficionados or by all the psycho acoustical
experiments conducted over the past few decades.
By way of conclusion, it appears that IMD, while
perhaps more disturbing than THD, remains practi-
cally harmless, provided the levels are not outra-
geously high.
Okay. Now bring on the next suspect…
T5.1.3. Temporal Distortion
In the ‘seventies, the famous Matti Otala high-
lighted a new kind of distortion, not so easy to
measure, but clearly audible, called Transient In-
termodulation Distortion, TID (also called TIM, for
Transient Inter Modulation).
TID occurs when the negative feedback (NFB) loop
is in a state of overflow, something that arises more
often than you might think since the NFB correc-
tion always applies after the phenomenon to be
corrected appears. During these very short instants,
the amplifier can produce more than 100% THD
and/or IMD.
At darTZeel, we prefer to call it Temporal Distor-
tion, or TD, because this is obviously what it actu-