This is an old anecdote (of mine), I generally avoided tantalums during recent decades. Back in the old days, tantalums were popular for their density (high capacitance for small size), but notorious as fire-starters when they short circuited across PS rails. Tantalums were notorious for poor dielectric absorption (DA), another specification blown out of proportion by the audiophile crowd. BUT back in the 70s I offered to do a kit article for Popular Electronics, in return for a cover story (you can't buy that kind of advertising). The design they wanted was a CX record decoder. A now obscure encode/decode vinyl noise reduction system. This technology was developed by CBS Records and I was provided a free license with full support package.
This was not very difficult, a simple 1:2 downward expander to complement the encoder 2:1 compression used to master the vinyl. As part of the licensee documentation I received a copy of the circuit design of the mastering 2:1 encoder designed by Urie (well respected for signal processing). To my unpleasant surprise I found a tantalum capacitor in their side chain primary time constant. In my experience DA is pretty innocuous in dc blocking audio paths, when used in a side chain, however the DA can actually make a difference. I arguably over engineered my decoder. I used a tantalum capacitor in my decoder side chain so any DA errors introduced to the encode control voltage would be symmetrically cancelled out in the decoder.
In the process of dialing in the time constants I found a mistake in the pro-forma decoder schematics that CBS provided (about a 10% error to the time constant). I made my decoder accurate to the published CX standard time constant, and notified CBS of the error in their published recommended circuits.
I delivered my finished article to popular electronics in time for the Christmas issue. Shortly thereafter while I was attending the AES show in NYC I wandered into the Urie booth and introduced myself to them. They responded "so you're the guy who found the mistake"
. Then they shared that because multiple consumer companies had already manufactured a few 10k playback decoders using the wrong time constants, CBS decided to change the encode time constant to agree with the mistake...
. making me the odd man out, because I did it right. As soon as I caught my breath I called the editor of Popular Electronics and tweaked my side chain (one resistor value) to agree with the new wrong time constant. Of course CBS never notified me of the change, I only found out by my chance meeting with the Urie design engineers.
JR