On 2008-11-10 Lars Finsen wrote:
> the long intervocalic stops, those that paradoxically are written 
> _kk_, _pp_ and _tt_, are fully voiced, not perceptibly less voiced 
> than the Norwegian voiced stops. 

It's the conditioning to register (as opposed to
actually hearing) all lenes as 'voiced' which
makes it sound so to us Norwegians and Swedes.[^1]
They really are voiceless lenes, modulo that it
seems **any** voiceless obstruent can apparently
become voiced between two vowels in modern Danish,
so that you can even get a minimal surface contrast
between [v] /f/ and [v\] /v/.  You can also get
both [4] and [T] beside [d_0] and [t_s] as allophones
of /t/.  Evidently Danish lenition is far from
a closed chapter: it will probably go on till
all Danish words are CV syllables and then start
fusing into longer words, so that apocope,
syncope and then a new helping of lenition
can start all over again...

I just got my hands on a little book I warmly

Type:  	  Book; Danish
Publisher:	KÝbenhavn : Akademisk Forlag, ©2007.
Editions:	3 Editions
ISBN:	8750039180 9788750039181
OCLC:	145568215
Related Subjects: 	Danish language -- Phonetics.

Close scrutiny of the relation between
phonetics, phonology and orthography makes
our cousins across the water **much**
easier to understand!

[^1]: I've come to the conclusion that we
probably have an instinctive notion that
voiced stops are the most typical (as in most
distinctive) lenes which creates this
illusion.  The gods know it's hard to unlearn,
but it gets us dead wrong when listening to
Danish -- and most Romance varieties: in the
dialect of Rome intervocalic /p t tS k/ are or
can be [b_0 d_0 dZ_0 g_0] while /b d dZ g/
are fully voiced.  I heard them all as
'voiced' and assumed a merger, which my
Italian acquaintances vehemently denied.
It just so happened that I was conditioned
by my L1 to hear those two phone types as
'the same'.