ALPHA will definitely color the sound. But it is based off the measurements of the change in slope of the curve and apparently accounts for several samples via a buffer. How it takes that data and adds it to the original data is audio-hand waving.
Then, the second processing it does is applying a dynamic digital filter than it applies, not unlike the PMD100 of yesteryear. The hype is that the coloration is pleasant based upon development using Nippon Columbia‘s recordings and knowing how to dynamically switch digital filters on the fly apparently is also a bit of art.
JVC K2HD is similar.
https://www.victorstudio.jp/flair/e/k2hd/k2hd.html
Here you see that they combine analog out from the old Sony PCM recorder in addition to original data source to guestimate how to manipulate the signal. But it is clear that the algorithm works with a specific PCM encoder.
No one knows if the Alpha branding (AL32) in receivers or SACD players is similar to the ALPHA used originally.
Anyway, for the Denon ALPHA gear I sent
@amirm, I am predicting a shrugging shoulder panther. By the numbers, it won’t beat the Levinson 360S. But I do think it will reproduce the 16/44 test tones from Alpha marketing brochures better than an Octo Dac8. The problem is that we don’t listen to test tones.
That is images 3 and 5 should be how Alpha processing reproduces those signals, with the consumer grade analog components impacting the measurements. Figures 4 and 6 and probably achievable with modern DACs with the right digital filter selected. ALPHA dynamically picks the filter for the music.
Denon’s final ALPHA demo was to take a Nippon Columbia musical recording and master it at -60dB since music easily dips that low in real content. The claim was that Alpha processing could take that real world content and make it sound better than traditional oversampling or upsampling. It will be interesting to see how ALPHA processing performs under the test bench and when the metric is 16/44 only.