Talk:Differential pulse-code modulation

Djpark0115 (talk) 07:00, 21 May 2009 (UTC) Q^-1 is necessary in the encoder/ecoder diagram of option2 ??

Yes, good catch; it's needed in the type 2 encoder as well. Maybe you can alert the editor who made the figure... Dicklyon (talk) 17:19, 21 May 2009 (UTC)

Q^-1 is horribly wrong! Quantization is not invertible. Option 1 is only possible if no quantization is done and the compression algorithm is lossless. The whole article has a lot of errors. — Preceding unsigned comment added by 188.155.108.64 (talk • contribs)

You were right to remove the bad figure from option 1. I edited the text to make it correct, I think. Dicklyon (talk) 18:07, 8 January 2011 (UTC)

Please make it more 'audible'

Please explain the roles of Z^-1, the circuitry. —Preceding unsigned comment added by Piotroxp (talk • contribs) 06:51, 19 June 2009 (UTC)

I noted in the article that it represents a sample delay; it's obscure conventional DSP notation based on the Z transform. Dicklyon (talk) 18:16, 8 January 2011 (UTC)

Local model

what does "local model" means in the option 2 section? --190.60.93.218 (talk) 18:53, 29 May 2014 (UTC)

Entropy is not reduced

Hello,

the statement "because the entropy of the difference signal is much smaller than that of the original discrete signal" is wrong. The idea of a DPCM is to reduce the variance by building a difference signal. The entropy is not reduced (it is not changed at all by calculating the difference signal). Entropy is only reduced in the quantized signal (obviously). — Preceding unsigned comment added by 130.75.31.216 (talk) 09:21, 20 August 2018 (UTC)