HD FMs that don't multicast: What's the point?

[kc0ltv]

Here in Duluth, we have three stations transmitting HD Radio:

91.3 KUWS, Wisconsin Public Radio / College Station
92.9 WSCD, Minnesota Public Radio Classical
100.5 WSCN, Minnesota Public Radio News

KUWS-HD went on the air over two years ago, WSCD-HD and WSCN-HD seem to have gone on some time in April or May. Neither WSCD nor WSCN transmit ANY subchannels: just a simulcast of their analog, with no real improvement in audio quality. Likewise, KUWS-HD DOES have an HD-2 channel, but it seems to be of the WPR Classical Network, which is carried on WHSA, Brule, WI, which has almost all of Superior within its 60 dBu contour is audible almost anywhere in the Duluth area on any radio (except perhap a cheap portable) worthy of the name.

What's the point of merely simulcasting perfectly-fine analog FM in digital? Why can't WSCD or WSCN import MPR's third network, The Current, to Duluth, or multicast the BBC or MPR Classical 24 as they do in Minneapolis?

 

[dumber than a box of hair]

Here in Duluth, we have three stations transmitting HD Radio:

91.3 KUWS, Wisconsin Public Radio / College Station
92.9 WSCD, Minnesota Public Radio Classical
100.5 WSCN, Minnesota Public Radio News

KUWS-HD went on the air over two years ago, WSCD-HD and WSCN-HD seem to have gone on some time in April or May. Neither WSCD nor WSCN transmit ANY subchannels: just a simulcast of their analog, with no real improvement in audio quality. Likewise, KUWS-HD DOES have an HD-2 channel, but it seems to be of the WPR Classical Network, which is carried on WHSA, Brule, WI, which has almost all of Superior within its 60 dBu contour is audible almost anywhere in the Duluth area on any radio (except perhap a cheap portable) worthy of the name.

What's the point of merely simulcasting perfectly-fine analog FM in digital? Why can't WSCD or WSCN import MPR's third network, The Current, to Duluth, or multicast the BBC or MPR Classical 24 as they do in Minneapolis?

Maybe they figured out that doubling the number of signals is untenable financially, and will not yield an ROI that makes it worth doing. In any case, you'll need to ask them why they're not running HD2 signals. I doubt you'll get an authoritative answer here.

 

[JohnnyElectron]

The entire Cumulus Cluster (f***) in N.Ohio went HD on FM with NO multicasting AND NO freakin' PAD - so why bother? All you get is the echo of delay between analog and deceased audio in digital and NO song titles, NO artists, -nothing- except the stations' web address.
What a waste. Throw us a bone - put on an HD-2 or two.
At least ClearChannel has complete PAD info most of the time, and a very noticeable audio switch from stereo analog to mono digital on HD-1 when the HD kicks in - plus you get an okay-sounding HD-2 channel with a terrible clone programming of the main channel (put something original on here for the love of...)

 

[Mike Walker]

I'm pretty sure mono digital is illegal when the analog is stereo. The FCC requires that the HD1 audio be at least the equal of analog. Well, that of course is subjective... your "equal" may well be different than mine. But I think we can agree that low bitrate mono digital AIN'T equal to good analog stereo!

As for "why HD when there's no multicast", at 96kbps HD CAN sound dramatically better than (most) analog FM stereo. It CAN. Obviously often it doesn't. I KNOW it can, because I've heard it. WDAV in Davidson NC, no slouch in the analog FM stereo department (classical music on 89.9), is simply stunning in HD. Through headphones, it's like a direct feed from the audio console (the HD is). When the switch takes place from analog to digital (and the analog sounds damn good), the stereo soundstage opens up, highs are crisper anc cleaner, there's more "focus"...it's easy to pinpoint individual instruments in an 80 piece orchestra, but most dramatic...there is NO noise. ZERO! And not a hint of multipath distortion (I live in a very hilly area...foothills of the Blue Ridge Mountains, Wilkes County, NC).

So the answer to "why" is the potential for better sound quality. However, there's much crossover in quality. Many analog FM stereo signals sound dramatically better than many digital signals, and vice versa. CD is more technically capable than LP (or analog tape). Full bitrate HD is more technically capable than analog FM stereo. But those are just specs. Good engineering vs. bad means that many LPs (VERY many) sound quite a bit better than the same recording on CD, and many analog FM signals sound much better than HD. Engineers who know what they're doing make all the difference. And radio has been losing engineers at an alarming rate (without replacing them) for quite some time.

 

[rbrucecarter5]

What's the point of merely simulcasting perfectly-fine analog FM in digital?

I am once again in the position of defending HD - in spite of its butchering of first adjacents.

Digital FM is indeed better, but you can only tell the difference on quality stereo component systems, not on the little squawk box speakers in trendy expensive table radios. Multicasting only serves to reduce the quality of digital FM, but the increase in frequency response alone makes it worth the effort. No more roll-offs at 50 Hz and 15 kHz. The noise floor is also much better.

 

[kc0ltv]

Yet as Bruce says, most people are only able to tell the difference with a good audio setup, and considering that 1) the bulk of listening is done on less-than-"quality stereo component systems", 2) most people don't seem to care much about audio quality or even discern the difference (unless it's significant, i.e. between AM and FM, and 3) the audio quality difference is of little importance for talk radio (in the case of WSCN), I have a hard time seeing broadcasters sell HD radio on the basis of improved audio quality vs. more channel choice. I'm not a fan of HD radio...it takes away at least six good adjacents for DXing here (91.1, 91.5, 92.7, 93.1, 100.3, 100.7), but it would be nice to have a few more options (ESPECIALLY when BOTH MPR Classical and MPR News run Prairie Home Companion simultaneously on Saturday night - who ever made that decision?).

 

[JohnnyElectron]

I've contacted the manager of the two CC stations with the 'reduced quality' HD1 signals over their main analog - he said that "there's nothing that he can do about it" as long as the have the HD2 in place and that "when they change it, it takes their whole analog transmitter out of commission too". Sounds like technical difficulties on 2 of their 3 FM's!

 

[Savage]

FWIW, we standardized on 128 kbp/s for "CD" quality - MONO - as a minimum standard at WYSL for our internet stream and for podcasts. I've gotta wonder how good 96kb sounds split up into three streams.

 

[OKCRadioGuy]

Here's the question of the week about HD: Can the HD1 take up the normal 96k and the HD2 use the extended hybrid mode? Does anyone here know if it can be split up this way? If I have the misfortune of having to go HD it would be great if I could utilize the full 96k for ourselves and give the extended area to the other guys we might have to have riding along with us...

 

[Mike Walker]

Apples and oranges, Savage. I don't know what codec you use for your internet stream, but HDC (based on AAC+) sounds quite transparent at 96kbps, damn good at 64kbps, and better than most analog FM stereo at 48kbps, when properly engineered. Below 48kbps, to these old ears, it begins to suck canal water (in stereo anyway). Older codecs like mp3 NEED higher bitrates to sound as good as (let alone better than) newer, more efficient codecs like AAC+ and HDC.

Anyone who follows progress in lossy codecs knows that newer codecs are more efficient all the time. Soon AAC+, and HDC will be "old hat" and nobody will believe that you need to go as high as 96kbps to sound "transparent" (I avoid the term "cd quality" whenever possible! By "transparent" I mean it sounds DAMN close to the audio coming off the board (or out of the audio processor).

The whole point of lossy codecs is HOW IT SOUNDS. If it sounds good, it IS good. If it doesn't, it isn't. Show me someone who draws a "line in the sand" based on some arbitrary number, and I'll show you someone who WILL NOT PASS a double-blind test! An affiliate of my oldies show INSISTS that the show arrive at 320kbps for his automation system, despite the fact that he admitted it sounded "exactly the same" when I sent it at a lower bitrate. 'Nuff said.

 

[tomservo]

Occam's Razor: The simplest answer is usually the right one.

I bet the two Duluth stations just haven't gotten all their equipment hooked up yet to put anything on the HD-2. Or they haven't finished loading a jukebox worth's songs on an old PC. ;D

HD will not sell on quality alone. Sadly Americans are quantity over quality, from the buffet table to seats in their SUVs to radio. Multicasting is what sells the radios out there now.

It's unfortunate to me that the outlets that would benefit most from the full bandwidth of just one HD channel - classical/jazz/public radio stations - are the ones most pressured into offering more choices.

Here in Mississippi MPB is advertising their HD channels heavily. MPB has gone mostly talk on the analog, relegating classical and special music programs to HD-2, and BBC World Service to HD-3. (As an aside, I tried to get some HD at a Wal-Mart the other day, with no luck. The radio stopped on scan on the lone MPB HD outlet but never decoded. It seemed sensitive enough and all the stations it hit on were clear, thanks to the Delta's insanely flat terrain. But no HD lock. I could have walked out the W-M's front door and seen the MPB tower in the distance. Maybe they took the HD off temporarily?)

 

[OKCRadioGuy]

Is it possible to use the full normal 96k for HD1 and put an HD-2 on the extended partition (assuming the station doesn't run SCAs, etc.)? If that's a possibility I'd think that would be best way to go for certain.

 

[Play Freebird]

Anyone who follows progress in lossy codecs knows that newer codecs are more efficient all the time. Soon AAC+, and HDC will be "old hat" and nobody will believe that you need to go as high as 96kbps to sound "transparent" (I avoid the term "cd quality" whenever possible! By "transparent" I mean it sounds DAMN close to the audio coming off the board (or out of the audio processor).

I dunno, Mike. I believe there's a finite limit to "efficiency" and AAC+ and HDC have practically reached it. There's just so much data that can be thrown away before your ears scream "who do you think you're fooling?"

Keep in mind that the "Nyquist limit" of frequency response is half the sample rate, so if you expect to transmit 15 kHz audio bandwidth, the sample rate must be no less than 30 kHz and preferably 32k to account for the less-than-perfect response of the low-pass filter ahead of the encoder. If a codec claims to support 15 kHz audio at less than 32 kbps, this means the higher frequencies are being replicated or synthesized. This is why the AM HD codec sounds unnatural if you try to push 15 kHz of audio though it, but much more pleasing if you stick with 10 kHz bandwidth. (Which, of course, raises the question "why bother?" when NRSC analog offers the same frequency response.)

Please note that some of the latest audio processors attempt to reduce artifacts by filtering out the higher frequencies and using a dynamic HF reduction algorithm, roughly analogous to pre-emphasis limiting in analog FM. If you want to read more about this, review the "Sensus" section beginning on Page 42 of the following PDF (numbered as 26 at the bottom of the page). Pay particular attention to the filter recommendations in the table at the bottom of Page 43:

http://www.omniaaudio.com/support/manuals/OmniaONE_MC_1.0_C_WEB.pdf

 

[tomservo]

Is it possible to use the full normal 96k for HD1 and put an HD-2 on the extended partition (assuming the station doesn't run SCAs, etc.)? If that's a possibility I'd think that would be best way to go for certain.

I'm pretty sure the 96 kbps is set. It can be dedicated to one channel or split up in various ways, but there's no way to get more.

There's just so much data that can be thrown away before your ears scream "who do you think you're fooling?"

Apparently some people's ears are less discerning than others. When data compression started making its way onto the airwaves, it stood out horribly to me. I guess engineers just have tin ears.

I remember wanting DirecTV when it first came out and then when I heard how bad the audio sounded it turned me completely off. Same thing happened with XM. Strangely, I got both now and complain about both regularly.

 

[KB1OKL]

Here in Duluth, we have three stations transmitting HD Radio:

91.3 KUWS, Wisconsin Public Radio / College Station
92.9 WSCD, Minnesota Public Radio Classical
100.5 WSCN, Minnesota Public Radio News

KUWS-HD went on the air over two years ago, WSCD-HD and WSCN-HD seem to have gone on some time in April or May. Neither WSCD nor WSCN transmit ANY subchannels: just a simulcast of their analog, with no real improvement in audio quality. Likewise, KUWS-HD DOES have an HD-2 channel, but it seems to be of the WPR Classical Network, which is carried on WHSA, Brule, WI, which has almost all of Superior within its 60 dBu contour is audible almost anywhere in the Duluth area on any radio (except perhap a cheap portable) worthy of the name.

What's the point of merely simulcasting perfectly-fine analog FM in digital? Why can't WSCD or WSCN import MPR's third network, The Current, to Duluth, or multicast the BBC or MPR Classical 24 as they do in Minneapolis?

"
What's the point of merely simulcasting perfectly-fine analog FM in digital? "

Well, they blew their money on it, are stuck with it like many other's who got taken by this lead balloon technology and goldarn it, they're going to get there money's worth whether anyone can notice any difference or not. Besides, what else can they do?

 

[tomservo]

"What's the point of merely simulcasting perfectly-fine analog FM in digital? "

Well, they blew their money on it, are stuck with it like many other's who got taken by this lead balloon technology and goldarn it, they're going to get there money's worth whether anyone can notice any difference or not. Besides, what else can they do?

Maybe they just wanted song titles and a text display. ;D

 

[Mike Walker]

Play Freebird, of course you're right that there is a practical limit to how low bitrates can go. I doubt we'll ever have transparent, stereo audio from an 8kbps stream. God I hope nobody tries. I take that back. I hope they do try, because in failing, they'll likely learn more lessons about psychoacoustics, and how to hide artifacts.

As for "lossy" codecs, remember so called "lossy" codecs aren't very lossy at all with most music, at sufficient bitrates. Take mp3 as an example. People often say "it's throwing away 90 percent of the music". Ridiculous! It's reducing the available data used to encoded the music by 90 percent. NOT THE SAME THING! Linear 16 bit, 44.1khz pulse code modulation (cd standard) can encode audio at EVERY frequency from 0hz to 20khz simultaneously. Impressive, huh? Well, only if your idea of entertainment is white noise...because audio at every frequency simultaneously is, by definition, WHITE NOISE.

Real music is a few distinct tones (notes), and a few dozen "overtones" (harmonically related tones which give instruments or voices their character), all transient (rather than continuous...as with white noise) in nature. Now THAT is far simpler to reproduce, and requires far less data to encode pretty much perfectly. And high bitrate mp3 (say 320kbps) DOES encode music pretty much prefectly, with almost no "loss" to the "lossy codec". This is easy enough to prove (as I have done here in the past) by anyone with an audio workstation program (like Adobe Audition). Simply rip a cd track uncompressed to your hard drive. Now open the uncompressed file, and save it as an mp3 (to a different file name). Now open the uncompressed file in the multitrack window of your program. Open the compressed (mp3 version) file, invert the polarity, and insert it into another pair of tracks in the multitrack window, lined up in time with the original. When you hit play, all you will hear from the speakers will be the difference between the two files (providing they're perfectly lined up). In other words, all you'll hear will be what mp3 encoding (or your preferred lossy codec) removed...almost nothing. In fact, to hear anything you'll have to turn the speakers up LOUD...then you'll hear a vague "ghost-like" whisper of the original audio.

Lower the bitrate with a particular codec, and the "null" test will show that more has been removed. But more efficient codecs like aac+ and hdc do as well at much lower bitrates (96kbps, for example) as mp3 does at 320kbps, and almost as well at 64kbps. It's not really "hocus-pocus", although it may seem like it. Today's lossy codecs sound better because we've learned to "lose less" of the original audio, even at what would have been disastrously low bitrates only a few years ago.

 

[Chuck]

Play Freebird, of course you're right that there is a practical limit to how low bitrates can go. I doubt we'll ever have transparent, stereo audio from an 8kbps stream. God I hope nobody tries. I take that back. I hope they do try, because in failing, they'll likely learn more lessons about psychoacoustics, and how to hide artifacts.

Actually, I hope they try it and figure out how to make it sound good. Think what things you could do with that!

From our perspective today, it seems like quite a reach, but I wouldn't rule it out. A lot of what was impossible yesterday is taken for granted today.

 

[tomservo]

I'll agree that codecs have gotten better, but that little bit of "ghost noise" that you get from Mike's original/compressed comparison is what adds the final 'verve' to the sound that makes it sound complete.

I know most folks are perfectly happy with 64 kbps joint stereo mp3 through their $10 earbuds, but we should never lower ourselves to the lowest standard. That'd be like making all TVs display just 256 colors because they cover the most important shades, or having cars only go 80 because few need to go faster.

Change your monitor to 256 colors down from 16.7 or 32 million colors - it ain't that much different. Now go display a nice colorful picture fullscreen. Are there obvious colorbands? Are transitions between light and dark portions choppy instead of smooth?

Of course your eyes are more adept at picking out those little missing bits than your ears - but not everyone will notice those subtleties. So why don't we all just de-colorize our pictures to 256 or 1024 colors to save space? I think that's the trend with music. Those of us who notice the subtleties missing in our music are in such a minority that no one thinks about us. I bet there isn't one single mp3 player on the market that could drive a pair of high-end headphones to distortion. Heck they probably couldn't drive them to be loud at all! Not only is the music that we put on these portable devices lossy, but the little mini-amps don't even reproduce the full spectrum of audio. The difference between my friend's iPod and my Hi-MD minidisc player through my car stereo, through the cheap-o lo-fi tape adapter, is stunningly obvious. The iPod sounds okay, actually, but it's like listening to analog FM radio: there's something missing.

Of course in the car, you're not likely to notice every little detail of the tractor trailer that just whizzed by. And you're not going to hear all the sonic subtleties on CDs or whatever. So I can kind of be okay with lossy compression for a digital transmission scheme, I just wish the codecs were a little better for us home listeners.

 

[Mike Walker]

So you're the one who's still into MD, Zach! I shoveled a lot of money down that hole! I have two working portables, and two working stand-alone md recorders. And I haven't used any of 'em in years! Once I could record uncompressed .wav files on a pocket-sized hard-disk recorder (Archos GMini-402), I never looked back. Ok, I did glance back...and weep a bit at the money spent on a losing format!

Speaking of losing formats, wanna' buy a low mileage Sony DTCA5 dat recorder, with Super Bit Mapping? Anybody?