I have been looking into using DPS to help recover some of the damage in CD recordings over the past 20 years with ever increasing amounts of over-compression and limiting. Thankfully, the streaming services have begun to put a halt to the “loudness war”.
I’ll never forget buying the CD soundtrack to “SlumDog Millionaire” only to find a measly 3 dB of dynamic range in the best songs. I wrote a complaint to the mastering engineer, who defended it as stating that “people liked it”. Yeah, the underlying songs were pretty good. Too bad about the mastering though…
I studied the Bongiovi patents to see what is going on. Then I came up with a roll-your-own approximation to DPS so that I could look for what parts of it produce the apparent re-inflation of damaged recordings. (damaged – meaning hyper-compressed). You can read about my experiment here.
There is undoubtedly some re-inflation produced from DPS. But I fear that most of it is due to treble enhancement in the graphic EQ. And that isn’t really true restoration of the missing peaks. You won’t get booming bass drums back that way.
I’m coming at this slightly differently than most people. I have a background in image restoration from the early days of the Hubble Telescope, when it was discovered that the primary mirror was incorrectly figured. Damaging amounts of spherical aberration in every image.
Yet, by way of something called phase-retrieval – an acknowledgement that every real object (sound, image, electrical signal, etc) has certain dispersion relations that must be obeyed. This relation allows us to resynthesize the missing phase information and produce corrected images.
I was hoping for something along those lines for dynamic range recovery too. And the Bongiovi DPS looks like a promising start.
Thoughts?
.
