In article , Tim Anderson
wrote:
I'm interested in SACD technology. But I'm puzzled by this:
http://interprod5.imgusa.com/son-637/technology.asp
"Unfortunately, the PCM process exposes the music to a "decimation"
filter during recording and an "interpolation" filter during playback.
These two filters can smear the sound, corrupt the sense of space around
the instruments and sacrifice the subtleties of live music."
- What do the quotation marks signify?
Not seen the page, but I'd assume the quote marks either genuinely mean
they are a quote from somewhere else, or the author is using them to imply
something - e.g. that he is writing in a descriptive manner that may not be
reliable. The comment seems to include some opinions after the "can..."
which may be of debatable relevance or reliability in any given case.
I could just as easily write something like "The SACD approach tends
to inherently produce large amounts of ultrasonic/hf hash/noise
that can cause problems for amplifiers and loudpeakers and hence
smear the sound." Putting this in quotes means I could then say
I'd expressed an opinion qualified with "can" that can be argued
to be supported by some evidence.
- "Decimation" means "killing every tenth man", according to my
dictionary. What's that all about?
In this context I suspect it means that the analog original has been
sampled at a high sampling rate, and this data steam is then reprocessed
and filtered to provide a lower sample rate output. By implication the
output may have a lower bandwidth. The historic meaning of decimation has
been ovetaken by engineers who now use the word for their own purposes. :-)
- I think I understand "Interpolation" which is about filling in a
missing value with a guessed value. Isn't that more to do with error
handling?
Not necessarily. It can mean inserting samples at intermediate instants
between the input data samples, but these do not have to be 'guesses', nor
are they just to correct errors. The idea may be to (using methods based
upon information theory) to produce the values *you would genuinely have
obtained* if you'd actually sampled the original input waveform more
frequently, but limited it beforehand to the bandwidth implied by the
sampling rate to which you are applying the 'interpolation'. The purpose is
to aid the filtering/processing process. It can make the task of the DAC
and following analog filters less demanding, so give (nominally) a 'better'
result.
Above said, It *can* mean attempts to replace values which have been lost
due to errors, so this will depend upon the context.
Or this:
"The Direct Stream DigitalT pulse train "looks" remarkably like the
analog waveform it represents. More pulses point up as the wave goes
positive and down as the wave goes negative."
- How do I "look" at a pulse train?
In a diagram or on the screen of something like a 'scope. :-)
- What are the benefits of this "look" to the audible results?
The closer the output of the DAC approaches the shape of the original
analog waveform, the less work the following filters have to do.
Or is this mumbo-jumbo to impress those who do not understand it? Having
said which, I don't understand it, but I'm not impressed either.
It just sounds to me to be unclearly expressed. However I've not had a
chance to examine the page, so can't say for sure.
Slainte,
Jim
--
Electronics
http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc
http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio
http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc.
http://www.st-and.demon.co.uk/JBSoc/JBSoc.html