The KillerDAC Audio forum
KillerDac => General Information => Topic started by: Drew on November 03, 2010, 09:36:14 AM
-
Hi All,
Since listening to kajaks system and talking to him about oversampling......have to wonder how and why the wider hifi world has become some entranced with something that effectively degrades sound.....
There would be very few production CD players that don't perform some kind of oversampling and most do upscaling as well!!??
How do you best describe the difference that oversampling makes.....smearing?
Drew.
-
Hi Drew and welcome to the forums.
This is a tough one to answer. I guess in theory and on paper/graphs, oversampling and upsampling provides the end user with more impressive result. Maybe something for companies to use as their main marketting point?
But for the end users/tuners who base their decisions on how it sounds (mainly by ear), then it sounds most impressive. These sort of people (like myself) aren't particularly worried about specs and values (to a point of course!) printed on paper, but more about it how it sounds.
Have a read through the following link:
http://forums.klipsch.com/forums/p/56463/1164182.aspx (http://forums.klipsch.com/forums/p/56463/1164182.aspx)
It has the 2 different groups of people I mention above, battling it out :)
To me, the oversampling on my marantz cd94 seemed to viel, smear the sound slightly. Removing the oversampling chip gave the sound more clarity, sharper/more precise decays of notes. The sound seemed to be more open. No sonic disadvantages that I could tell.
Please don't get me started on upsampling!! First 48 then 96, 192, 384, 768kbps upsampling.. when will they stop??? :)
This is just in my experience/opinion though.. so it's not something to take seriously :)
Enjoy your stay,
Tuyen
-
At the end of the day it's money... All about cost....
Oversampling apparently reduces noise, not hard to do, could mean everything else on the player can be made cheaper as there is less noise to deal with?..
Upsampling gives more (fake?) data points. In a sense it's a trade off between processing and storage. less points requires more processing as the processor needs to know what happens in between 2 data points (is it a curve, or a straight line?)... The more data points you have, the closer the line in between approaches a straight line that is close to the original wave form.
Over all that, they probably apply filters on top to get the final results...
So really quite a lot of processing, so many places where things can go wrong...
I would like to think the theory is sound, but the application/implementation is wrong.. Eventually we will get there, just not sure if it's within our lifetime.. :p
-
NOS hi-res also bad?...
what would be the options here?
-
Non oversampling it is for me..
I listened to the same dac....after changed to nos i found the differences downright huge. :o
Do not put a scope on it.....but your ears.. ;)
-
I have no idea what you mean...
what dacs are you talking about?
-
I mean that non oversampling sounded much more like music to my ears.
In this particular case it was an old dac 3 from Audionote.....they used 2 pcm63k iirc
My guess on this is that all the processing in order to get a nice scope wave result does more harm than using a "non-digital-filter" (NDF) and getting it nos....with a "horrible" wave on the scope.
Hope you get it...I am a foreigner ;)
Just a personal opinion. ;D