CD Upsampling on DVD Player question

[iLuvLandingStrips]

Can someone explain to me what exactly upsampling does?
What is the difference between playing it "upsampled" and playing it at standard rate on a normal CD player? Better sound?

My DVD player has these 2 options..

88.2KHZ (x2)
176.4KHZ (x4)

Thanks for any help.

It's a Yamaha 5750S player.

[iLuvLandingStrips]

No one knows anything about this?
C'mon...where's all the PROS around here?

[trossin]

Ok. I'll give it a try. The answer is there should be no difference. The only difference is in the way you market your product. Oversampling is a technique that is used to make up for a non-ideal filter. By filter I mean the circuit that is used to convert digital to analog.

This old math god Shanon figured out long before you were born that it is possible to take a continous waveform (analog), convert it to discretly sampled data points (digital) and then convert it back to an exact replica of the original continous waveform (analog) if you have and ideal filter and if you sample the waveform at twice the highest frequency component of the signal.

The fun comes into play when reality raises its head and you find out that it is not possible to construct an ideal filter. Folks have found out that you can get by with a cheaper, less ideal, filter by raising the sampling rate. In the old days when microprocessors were slow this was a bogus option. No machines are fast so you can sample and playback very fast and use a dirt cheap filter and it still sounds good.

More fun comes into play because the CD spec came around in the early 80s so they are sampled at just a hair faster than the minumum Shanon said to use (44KHz is just 2.2x of 20KHz). 20KHz is the max freq most humans under the age of 20 can hear. Old farts like me (40 something) can't hear much over 16KHz.

So, how do you recreate a waveform using a cheap filter when the source data is sampled at a slow rate? The answer is to create extra samples and push them out to the cheap filter at a higher rate. But, how is this done? There are many ways to go. One way is to linear interpolate between the data points and create new samples along the line. Other techniques are to look at adjacent sample and map a curve to the points and then sample the curve. Anything you do may create distortion because you a making assumptions about data points that do not exist.

Remember that of this is done to make up for the fact that the designer was too cheap to put in a high quality filter. But, clever marketing people will try to lead you to believe that a higher playback rate is good when in reality it just show how far their filter is away from ideal.

Help me. I'm typing and I can't stop.

[edDV]

Sounds like hype to me. What good does upsampling a 44KHz CD do for just playback? Didn't they document the hype in the product literature?

I can see doing it if you needed to mix it under a 24 bit 96KHz track.

Try an audiophile forum.

[iLuvLandingStrips]

Thanks guys! Appreciated.

[edDV]

Ok. I'll give it a try. The answer is there should be no difference. The only difference is in the way you market your product. Oversampling is a technique that is used to make up for a non-ideal filter. By filter I mean the circuit that is used to convert digital to analog....

Great post trossin.

Bringing this back to video, the same techniques are used for cheap capture cards to aviod use of expensive high quality low pass filters ahead of the A/D. Video is oversampled, poorly filtered and then later digitally filtered and down sampled on the digital side. The result is noise.

That is why consumer capture cards sell for $50 when just one of the three component low pass filters on a "broadcast quality" capture card costs more than $50.

http://www.allenavionics.com/VidFilt/pdf/AVS.pdf