Re: Luma discussions

From: Justin Cordesman <shadow_at_darksideresearch.com>
Date: Mon, 20 Feb 2017 07:49:21 -0800
Message-Id: <zarafa.58ab1001.09fe.542fee5d2ae6805c@atlas.darksideresearch.com>
Only if it is repeatable, which I'd suspect it is not.  Things like color gamut being reduced on an old plasma, and quantization errors from the scaler are not indicative of actual performance by the C64.  It is also entirely possible for features in the scaler that are intended to do things like reduce compression artifacts or normalize analog noise will misbehave and generate new artifacts from legitimate signal or in attempts to reduce artifacts present in the original signal. Early ones are simple edge detection algorithms that do smoothing.

Also, artifacts that are invisible under normal use are irrelevant, and not all signal defects are subjectively a problem.  For example, adding a small amount of analog noise to a noiseless video stream makes people score it as more detailed than the noiseless version when they are untrained viewers.

Justin

On Feb 20, 2017, at 17:34, smf <smf@null.net <mailto:smf@null.net> > wrote:

On 20/02/2017 14:33, Justin Cordesman wrote:
> I agree that the issue with that old article is clearly that it is describing behavior of the scaler and quantization of the signals coming out of the 64 and cannot be relied on to describe the behavior of different revisions.

If putting it through a scaler makes display noise more visible that is 
usually hidden with a combination of your crt, your eyes and your brain 
then it is perfectly valid to use that to determine which revision is 
better than any other.



       Message was sent through the cbm-hackers mailing list




       Message was sent through the cbm-hackers mailing list
Received on 2017-02-20 16:03:42

Archive generated by hypermail 2.2.0.