The reliability, clarity, and fixed low latency were nothing to sniff at, and ISDN didn't lose on quality by any means. In a different universe, it might have gone on to be a dominant technology.
The data rate was 64kbps. The audio frequency bandwidth of G.722, the most widely compatible codec over ISDN, is about 7kHz. Typical analog phone line bandwidth was about 3kHz, so a single ISDN BRI D-channel call was already twice the bandwidth of analog. This was sufficient for ISDN to be suitable for links between sports arenas and FM broadcast stations, for example. Other codecs with better compression (like MPEG) over bonded D-channels for 128kbps, could provide more like 20kHz audio bandwidth (in stereo!), which is close enough to "broadcast quality" 22kHz to be useful for coast-to-coast broadcasts and remote studio recording.
In our timeline, however, Carterfone and the breakup of AT&T opened the doors for all kinds of development in voice-band modem technology. 56K modems were good enough, and they worked with existing last-mile equipment on existing POTS lines, leaving ISDN to find a niche with small business and broadcast. When "always-on" DSL entered the market to compete for Internet subscribers, ISDN was finished. People cared more about data rates than latency, more about the Internet than point-to-point links, and the market ISDN was aiming for had moved on by the time it arrived. Rapid deployment of fiber and T1/E1 quickly ate up whatever was left of ISDN's apple... not because ISDN wasn't very good at what it did, just because people didn't much care about what it was good at.
Which, I suppose, brings us back to the point of the article: advances come with drawbacks.
The data rate was 64kbps. The audio frequency bandwidth of G.722, the most widely compatible codec over ISDN, is about 7kHz. Typical analog phone line bandwidth was about 3kHz, so a single ISDN BRI D-channel call was already twice the bandwidth of analog. This was sufficient for ISDN to be suitable for links between sports arenas and FM broadcast stations, for example. Other codecs with better compression (like MPEG) over bonded D-channels for 128kbps, could provide more like 20kHz audio bandwidth (in stereo!), which is close enough to "broadcast quality" 22kHz to be useful for coast-to-coast broadcasts and remote studio recording.
In our timeline, however, Carterfone and the breakup of AT&T opened the doors for all kinds of development in voice-band modem technology. 56K modems were good enough, and they worked with existing last-mile equipment on existing POTS lines, leaving ISDN to find a niche with small business and broadcast. When "always-on" DSL entered the market to compete for Internet subscribers, ISDN was finished. People cared more about data rates than latency, more about the Internet than point-to-point links, and the market ISDN was aiming for had moved on by the time it arrived. Rapid deployment of fiber and T1/E1 quickly ate up whatever was left of ISDN's apple... not because ISDN wasn't very good at what it did, just because people didn't much care about what it was good at.
Which, I suppose, brings us back to the point of the article: advances come with drawbacks.