You’d describe the encoding, not the source. The fun part is that it also applies to audio. “At 256 kbps, MP3 is transparent.”
It only applies to lossy codecs. Lossless codecs, by definition, have no error. “Error” itself being a borrowed term. Good encodings don’t have fewer errors… they have less error. For example, measured as mean squared error, where an individual sample being very wrong counts more than many samples being slightly wrong.
You’d describe the encoding, not the source. The fun part is that it also applies to audio. “At 256 kbps, MP3 is transparent.”
It only applies to lossy codecs. Lossless codecs, by definition, have no error. “Error” itself being a borrowed term. Good encodings don’t have fewer errors… they have less error. For example, measured as mean squared error, where an individual sample being very wrong counts more than many samples being slightly wrong.