Upsampleing?


I have a little problem understanding the concept of upsampling. Maybe someone can explain it to me. If a redbook CD is 44.1hz how can upsampling produce a 192hz out of a 44hz disc? Doesn't the disc only have 44hz on it, so therefor is the upsampling just bridging the "gaps" and essentially smoothing out the sampling rate segments?
I wonder if the improved sound is really a fabrication more than an unrevealing of music that is supposedly there in the disc and being missed by most players.
128x128theo
Of course. Upsampling cannot get more information out of the original since there is no more information. It can help with the necessary filtering and other processing.
This might help you:-

http://forum.audiogon.com/cgi-bin/fr.pl?ddgtl&1059490592&read&keyw&zzupsampling

Also, search these archives using keyword "upsampling" & there should be plenty of talk on this subject here on A'gon.

FWIW
You are right on.

You've probably seen those graphs of a curving soundwave. The original musician's sound will be a true smooth curve.

You can think of a CD recording as a curving stairway used to model a true curve. Hz means times per second and in the case of a CD, it is sound samples per second. A 44.1 Khz CD will have 44,100 steps (or sound samples) per second. If the sample rate were 192khz it would be 192,000 samples for that second of music. Your stairway would more closely match the original curve for that second.

An upsampler only has the 44,100 samples to work with because that's what comes on the CD. Based on the look of the curve, it guesses what the smaller steps would be and adds these little stairs in. When the music is played your ear hears the smoother sound of 192,000 stairs verses the 44,100.

To me upsampling improves the sound significantly. I recently purchased an upsampling cd player that lets me change the upsampling rate using a front panel switch. Myself and 5-6 others who have heard me change the upsampling have consistently preferred the higher setting.
John thank you for adding clearity to my confusion. I felt my perception was correct and you "stairstep" analogy simplifies the description of what I was clinging to.
So in a perfect world where the steps are taken in even stride by redusing the step size the processor fills in those gaps and therefor a cleaner or should I say more continued sound along that projected path?
Actually, the stair-step analogy is not accurate as the reconstruction filter smooths the outputs to the same result. Now, the nature of that filter can be affected by the output sample rate but that's another story.
More to discover