Dan gave you a good answer, but I'll elaborate on what he said.. FM HD uses a total bit rate of 144kbp/s. Within that 144, 96kbps are reserved for the HD1 audio. Within that 96kbps are some packets reserved for FEC (Forward Error Correction-bits) Assuming a station has an HD2 channel, there are 48kbps left for an individual ancillary HD2 channel with very little FEC. If the station runs and HD3 channel, then the HD2 and HD3 splits 48kbps, 24 each, with pretty much no FEC.
The delay Bruce complains about frequently, is caused by caching or buffering of the stream at the receiver until enough bits in the correct order arrive to continue the stream playing. There is very little room for extra (redundant) bits for the HD2 and HD3-24kbps to replace potential missing ones lost to things like: multipath, interference, whatever.
Hmm, that's interesting. So the HD1 is fixed at 96 or higher kbps? If the leftovers can be divvied up between any additional subchannels, I'm wondering if the FEC can be tweaked to allow one channel to have more error correction than another, then… that would explain why the HD2 is more robust than the HD3 on my local stations. On the opposite end of my state, a friend has noted that one of the iHeart stations in Huntsville is running and HD2 country format and an HD3 religious format, and the HD3 is the one that works well while the HD2 is much more dropout-prone.
2. Radio station owners aren't taking HD-2 seriously.
Getting listenership on the subchannels has always been a chicken-and-egg situation. The stations won't program compelling content unless people are listening, but people won't listen unless there's compelling content. And that's assuming anyone has an HD radio at all. I'm
really starting to see it in a lot of cars now, so it's definitely making headway in the automotive space. But at home, quality HD radios are few and far between.
One thing I've never been able to figure out is why iBiquity decided table radios (horrible unitaskers that few buy anymore) were the preferred home solution instead of getting HD fitted to home theater receivers. It's the perfect place for it. Home theaters are more popular than ever, it's a stationary device so dropouts will not be a big issue and the majority of them already come with AM/FM as standard. But few if any offer HD built-in. At best it's a $100 add-on that no one would dare touch.
I do believe radio stations owners ARE taking HD-2 and HD-3 feeds seriously now, because they've found a way to monetize them, via translators and leases. In this sense they're nothing more than glorified SCAs, but apparently those have been profitable for broadcasters, too. HD radios are still easier to acquire than SCA tuners, that's for sure.
You talk about the ubiquity of internet, but keep in mind not all internet connections are the same. When I got my first "smart" phone in 2010, the my carrier's network was EV-DO 3G and it rarely hit 1 Mbps, and even then only for fits and spurts. (It was Cellular South, later C-Spire, a regional southern CDMA carrier.) AT&T was tops here, with 2 Mbps through their HSPA 3G. There was no backbone infrastructure to support anything faster.
Nowadays I do have 4G LTE through Verizon and it's lightning fast — I've hit 80 Mbps a few times in download speed tests — but when moving even it tends to drop streams and rebuffer a lot. There's no way streaming can handle the local listenership of a typical modest radio station without crashing to a halt. Of course even if it did work flawlessly, I'm still staring down a data cap, one I'd rather reserve for watching Netflix away from wifi
!
5. Consumers still won't listen because of 4.
Now I know this is somewhat logical, but it has not been true in my observation. 250 watt translators are still a pittance compared to a typical urban class B/C's coverage abilities. They will have gaps in challenging terrain that the HD can fill in. In Birmingham, I know of a handful who have embraced the HD-2 feeds of a couple "metro signal" broadcasts due to terrain shadowing of the analog translators. In that market, some translators are 250 watts but others are only 99 watts and the power levels are trending down as space becomes scarce. A relative recently bought a new-used Mazda and I was happy to show her how to listen to her beloved rock station via the HD2 feed when she couldn't pick up the translator in the fringe areas. Now she has dropout-free service via one or the other for her entire commute to work.
6. Broadband internet is a much cheaper and better way of distributing signals to translators, so HD-2 isn't necessary.
That'll mean something if the FCC ever allows stations to feed a translator directly that way. Right now the rules say they can only rebroadcast something else terrestrial: an analog FM/AM or digital station.
Personally, I'm more convinced than ever that IBOC HD was never actually meant to succeed the way it was hyped in the early days. I think secretly the thinking at iBiquity was "if it works the way we'll claim it does, great. But if it doesn't, there's always using it to skirt ownership caps, which is a guaranteed success."
The issue with car radios with C-Quam was that by the time any of them got on the road, most music was off FM. And, for the first number of years, C-Quam had awful platform motion; I witnessed one person have to stop a car while we were doing signal measurements, get out, and vomit. It was rather awful in that respect and many of us believe it made fringe AM listening in the nulls of a directional very flangy and unpleasant.
It's sad that the technology never got off the ground. From what I've read, platform motion is pretty much a non-issue with the most modern chipsets available. Too bad no one uses them anymore. It seems like that happens a lot in the world of technology. A good idea is released a bit before it's fully baked, then flounders to obscurity while steadily improving. In addition to AM stereo, I personally embraced another example of this: minidisc. I got in probably in the second or third generation of ATRAC, when the sound quality was starting to really pick up. The technology continued to advance, and had Sony not been so stubborn with their baby, it might have seen wider success. By the end of the line (in the US at least), minidiscs could hold hours and hours of high fidelity audio thanks to many ATRAC improvements.
Just as I have a little-used AMAX Walkman sitting in a drawer, I have an entire closet full of MDs in cases, that I'll probably never listen to again.