What I hear

[Mike Walker]

After a few months listening to HD Radio (since November when I got my Accurian), I have to say there's a real downside to it. Analog fm stereo can sound really grating to me. I'm not sure if it's the (way too) aggressive processing that most stations use on their analog signals these days, or what...but there is a "cleanness" a "thereness" to HD that is most obvious by it's absence when I'm forced to return to analog fm stereo (as in my car, or on my Walkman). I've got to think that it's mostly processing artifacts from stations having their boxes on "blast-off", because I certainly have some warm-fuzzy memories of how fm stereo used to sound!

Interestingly, even though AM is often more highly processed than FM, I don't hear this "edginess" on AM...and I'm talking about on GOOD AM radios, not the response challenged "typical" units most people suffer through. Here in the foothills and mountains of northwest NC there are still quite a few AM stations that play music, and more and more I find that the smoothness (with limited highs) on AM is preferable to the "edginess" (for want of a better description of the distortion I'm hearing) of analog FM stereo. Best of all is HD...particularly on the types of acoustic music I prefer (classical, acoustic jazz, bluegrass/folk, etc.)

The best experience I can relate this (hearing previously unnoticed artifacts in analog material after exposure to digital) is how I felt about analog lp after early exposure to cd back in the early 80s. I've always been particularly sensitive to pitch (make fun if you like, I'm blessed/cursed with nearly "perfect pitch". A speed error of .5 percent, unnoticable to most people, is TORTURE to me!) YES I heard the piercing brightness of early digital, but I accepted it because of a benefit that was a Godsend...PERFECT PITCH ACCURACY, and rock solid stability. Suddenly I could TRULY enjoy music...free of the most objectionable (to me) artifacts...SPEED ERROR AND INSTABILITY. Well HD has made me keenly aware of distortion/grittiness/"edge" in analog fm stereo...even on well engineered stations like WDAV Davidson NC. When my radio switches to HD, the heavens open up, and I am IN THE ROOM with the musicians.

Anyone who has spent a few hours of listening to HD through headphones, or really good speakers...and who truly adores music...knows what I'm talking about. Either that, or they're suffering some SERIOUS hearing loss!

 

[radiopilot]

After a few months listening to HD Radio (since November when I got my Accurian), I have to say there's a real downside to it. Analog fm stereo can sound really grating to me. I'm not sure if it's the (way too) aggressive processing that most stations use on their analog signals these days, or what...but there is a "cleanness" a "thereness" to HD that is most obvious by it's absence when I'm forced to return to analog fm stereo (as in my car, or on my Walkman). I've got to think that it's mostly processing artifacts from stations having their boxes on "blast-off", because I certainly have some warm-fuzzy memories of how fm stereo used to sound!

Interestingly, even though AM is often more highly processed than FM, I don't hear this "edginess" on AM...and I'm talking about on GOOD AM radios, not the response challenged "typical" units most people suffer through. Here in the foothills and mountains of northwest NC there are still quite a few AM stations that play music, and more and more I find that the smoothness (with limited highs) on AM is preferable to the "edginess" (for want of a better description of the distortion I'm hearing) of analog FM stereo. Best of all is HD...particularly on the types of acoustic music I prefer (classical, acoustic jazz, bluegrass/folk, etc.)

The best experience I can relate this (hearing previously unnoticed artifacts in analog material after exposure to digital) is how I felt about analog lp after early exposure to cd back in the early 80s. I've always been particularly sensitive to pitch (make fun if you like, I'm blessed/cursed with nearly "perfect pitch". A speed error of .5 percent, unnoticable to most people, is TORTURE to me!) YES I heard the piercing brightness of early digital, but I accepted it because of a benefit that was a Godsend...PERFECT PITCH ACCURACY, and rock solid stability. Suddenly I could TRULY enjoy music...free of the most objectionable (to me) artifacts...SPEED ERROR AND INSTABILITY. Well HD has made me keenly aware of distortion/grittiness/"edge" in analog fm stereo...even on well engineered stations like WDAV Davidson NC. When my radio switches to HD, the heavens open up, and I am IN THE ROOM with the musicians.

Anyone who has spent a few hours of listening to HD through headphones, or really good speakers...and who truly adores music...knows what I'm talking about. Either that, or they're suffering some SERIOUS hearing loss!

Perfect Pitch? What instrument do you play? Piano, guitar? Are you tuned to A440? I'll await your answer...

Hearing music is personal to everyone on a personal basis, one person may hear highs to a certain point while others it's unlimited, some people hear lows to a certain point and others can't, so no two people will listen to the same degree.

Yeah I can tell the difference between lp's, cd's and live music because my ears are trained to it, others can't, if you can great you should be able to pick up a sheet of piano music and sing the notes out in perfect pitch!

I'm not saying digital music isn't great or anything, just the delivery system of HD. I'd rather hear the music live if I can, but we all can't do that, and anything that comes close is great, but at what costs?

Radiopilot

 

[PocketRadio]

When my radio switches to HD, the heavens open up, and I am IN THE ROOM with the musicians.

With our new in-dash Sirius Radio (we didn't have a choice), I can't tell any difference compared to analog FM. What IS annoying, as with HD Radio, are the "moments of silence", as one goes under power lines, into tunnels, or pulls into the garage. But, it is amusing to hear uncensored radio.

 

[tomservo]

If you're losing your satellite radio going under power lines, somethin' ain't right. The garage I can understand if there's living space above it, or it's metal/brick directly above. Then again Sirius' polar-elliptical orbiting birds make reception interesting sometimes. :

Mike, if you're as sensitive to pitch as you claim then I really feel for you. It's annoying to notice those little details that everyone else either doesn't notice, or easily ignores. I'm sensitive that way to codec artifacts; the early years of internet streaming and mp2/3 technology was a painful time for me.

Ironically now, the vast, vast majority of my radio listening is via satellite radio or internet streaming. I went from the ancient mp2-based radio stations in Birmgham (ex-Dick owned) to minidiscs (Sony's ATRAC codecs still set the bar for lossy audio compression, IMHO) to XM and internet feeds. I'm actually "used to" XM since it's almost all I listen to, musically. The shock comes when I go back to MDs or listen to the XM feed on DirecTV, or listen to public radio. Then the codec-fatigue sets in all over again.

OK, I guess I need to make this relevant.

At least the HD radio codec is decent on FM. They sure got satellite beat, SQ-wise (even in analog, heck even AM stereo) but ya can't beat the coast-to-coast coverage.

 

[Mike Walker]

With respect Radiopilot, "Perfect pitch" ain't the same thing as being able to sight-read music, although most musicians can do that to some degree. I play trumpet.

What I mean by "perfect pitch" as it relates to recordings is that if I've heard a recording just once at the correct speed (which I used to verify with a strobe disc...today that's unnecessary...digital may have flaws, but speed instability ain't one of 'em), I can ALWAYS tell INSTANTLY if I later hear it too fast, or too slow (pitch too high, or too low). I used to have fun demonstrating this to listeners on the air by having them name a popular song. I would hum the first note of the song...at exactly the right pitch, then start the song. It didn't matter if I hadn't heard the song in three weeks, or fifteen years...if I've heard it, and remember it, I that memory includes the EXACT pitch at which I learned it. Call it a curse, or whatever, but I can EASILY demonstrate it. Hand me a cd of music I'm familiar with. Tell me the song. I'll hum the first note, then play the cd. The note will be EXACTLY "on pitch". A completely useless ability (curse) that kept me from ever getting any enjoyment out of cassette decks, which were always notoriously off-speed...hence off-pitch. The average speed error on cassette decks was/is somewhere around 1 percent...WAY above my "pitch threshold".

 

[Mike Walker]

I tried Sirius for a while, because there's some programming on there I really wanted to hear. But while I can also hear digital artifacts on XM, they're of a nature that I find much less bothersome. There is a crakly distortion during complex passages on Sirius that drives me NUTS. It's 'fingernails-on-chalkboard' obvious...sounds like something somewhere in the chain clipping (being overdriven into distortion), but it's on EVERY music channel...more obvious on some than others (acoustic music seems to suffer less than rock...probably because there's usually less high frequency content).

HD was to have used the same codec as Sirius, but listening tests had participants screaming bloody murder about sound quality issues, especially at the very low bitrates used for AM HD. HDC is quite a bit better (but I have yet to hear an example of artifact-free AM HD! Even the WOR samples I've heard, while sounding quite nice in some ways, have that weird, "comb-filtery" thing goin' on...the same thing we used to endure with audio on dialup modems. I'm not saying it sounds as bad as dialup modems! But the effect is similar (to my ears).

Of course you may find XM's artifacts more annoying than Sirius. Ditto HD. We all listen differently, and have different thresholds for discomfort with artifacts.

 

[Tom Wells]

With respect Radiopilot, "Perfect pitch" ain't the same thing as being able to sight-read music, although most musicians can do that to some degree. I play trumpet.

What I mean by "perfect pitch" as it relates to recordings is that if I've heard a recording just once at the correct speed (which I used to verify with a strobe disc...today that's unnecessary...digital may have flaws, but speed instability ain't one of 'em), I can ALWAYS tell INSTANTLY if I later hear it too fast, or too slow (pitch too high, or too low). I used to have fun demonstrating this to listeners on the air by having them name a popular song. I would hum the first note of the song...at exactly the right pitch, then start the song. It didn't matter if I hadn't heard the song in three weeks, or fifteen years...if I've heard it, and remember it, I that memory includes the EXACT pitch at which I learned it. Call it a curse, or whatever, but I can EASILY demonstrate it. Hand me a cd of music I'm familiar with. Tell me the song. I'll hum the first note, then play the cd. The note will be EXACTLY "on pitch". A completely useless ability (curse) that kept me from ever getting any enjoyment out of cassette decks, which were always notoriously off-speed...hence off-pitch. The average speed error on cassette decks was/is somewhere around 1 percent...WAY above my "pitch threshold".

I don't find this a curse. I replaced the drive motors in 2 tape decks, and you can't buy a timing tape anymore.
I set them by pitch with known good recordings. If I put in a tape recorded on one particular 'other' deck, I immediately know it.
I don't read music, but can close my eyes, and with some effort, play piano along with music. (right hand) Opening my eyes makes it damn near impossible.
My memory of music also includes pitch, and if there is anything that's a curse about it, it's listening to people try to sing who seem to have no pitch
abilities at all.

I have one or two 45s where it's obvious that one of them was a dub at off-speed.


Regarding MP2 and MP3, I never even was tempted to try them. I heard them being auditioned on-air on WLS back in the 90's, and
I was horrified that such a garbled sounding product would be brought to market.

I CAN listen to 1920s-30's music OK on a low bitrate stream, but anything new seems zizzly if it less than 96k.
128k I can accept as "transparent".
I wish more streaming audio broadcasters would listen to the stream, and roll off the high frequency info before encoding.
I'd much rather have LESS high frequency info, if the resulting sound had fewer artifacts.

 

[tomservo]

I like that idea. I'm not sure how well it works in the real world, but I always thought the best thing to do encoding low bitrate music streams/files would be to roll off the treble a bit, then allow the end user to turn up the treble to compensate if desired.

You can bring the (slightly less?) watery treble up if it don't bother you, but the rest of us can't eliminate that same problem by cutting treble. Or at least that's been my experience.

 

[R.F. Burns]

I like that idea. I'm not sure how well it works in the real world, but I always thought the best thing to do encoding low bitrate music streams/files would be to roll off the treble a bit, then allow the end user to turn up the treble to compensate if desired.

You can bring the (slightly less?) watery treble up if it don't bother you, but the rest of us can't eliminate that same problem by cutting treble. Or at least that's been my experience.

Conversly you can always turn down the treble (if there's a tone control) if you feel the high end is filled with artifacts (and on AM, and some FM Hd2's there are disturbing artifacts). What I've found using my Receptor is that audio tayloring makes all the difference in how listenable some of these low bit rate audio streams are.

 

[Mike Walker]

Actually RF though it's the attempt at extended treble at low bitrates that really results in artifacts, the most audible artifacts from this occur in the midrange. If you have a copy of Adobe Audition, open an uncompressed .wav file, and save it to an mp3 at 32kbps stereo with a 44.1khz sample rate, and 20khz frequency response. What a freakin' mess! And the artifacts are in the midrange. Now go back to the uncompressed .wav file, and save it with exactly the same settings, EXCEPT restrict the frequency response to 8khz. Yes there are audible artifacts, but it's completely tolerable now, isn't it?

There are no musical fundamentals (notes) above about 3khz. Extension of frequency response beyond 10khz aids "air", "sheen", "presence", "clarity", "crispness" to music...but remember...the ear is most sensitive at midrange frequencies. These MUST be reproduced clearly, or nothing else matters. If you MUST encode at a low bitrate, the first thing to go should be "extreme" highs...above 10khz. If you're trying to do stereo at bitrates as low as 32kbps, then use a lower sample rate (16khz is good) and restrict highs to 6-8khz. If you insist on using 44.1khz sample rate, then still restrict the highs. This allows the codec to allocate those precious bits to the midrange, where it counts...where music lives, and where the ear is most sensitive.

Bass is relatively unimportant to this equation, by the way, as bass is simple to encode, and requires very few bits. There is no comparable gain in clarity or quality to be had by restricting bass extension. Some things I've learned preparing low bitrate audio for the "interweb" (it's not a truck, but a series of tubes) for about 12 years.

 

[R.F. Burns]

Actually RF though it's the attempt at extended treble at low bitrates that really results in artifacts, the most audible artifacts from this occur in the midrange. If you have a copy of Adobe Audition, open an uncompressed .wav file, and save it to an mp3 at 32kbps stereo with a 44.1khz sample rate, and 20khz frequency response. What a freakin' mess! And the artifacts are in the midrange. Now go back to the uncompressed .wav file, and save it with exactly the same settings, EXCEPT restrict the frequency response to 8khz. Yes there are audible artifacts, but it's completely tolerable now, isn't it?

There are no musical fundamentals (notes) above about 3khz. Extension of frequency response beyond 10khz aids "air", "sheen", "presence", "clarity", "crispness" to music...but remember...the ear is most sensitive at midrange frequencies. These MUST be reproduced clearly, or nothing else matters. If you MUST encode at a low bitrate, the first thing to go should be "extreme" highs...above 10khz. If you're trying to do stereo at bitrates as low as 32kbps, then use a lower sample rate (16khz is good) and restrict highs to 6-8khz. If you insist on using 44.1khz sample rate, then still restrict the highs. This allows the codec to allocate those precious bits to the midrange, where it counts...where music lives, and where the ear is most sensitive.

Bass is relatively unimportant to this equation, by the way, as bass is simple to encode, and requires very few bits. There is no comparable gain in clarity or quality to be had by restricting bass extension. Some things I've learned preparing low bitrate audio for the "interweb" (it's not a truck, but a series of tubes) for about 12 years.

I do have both Sound Forge 6 & 9 with noise reduction and other plug-ins. I'll try to look at a compressed file and see what happens when I notch out the offending frequncies. If it works, maybe the same could be done at the receiver end.

 

[Mike Walker]

RF I'm talking about using telling the codec to not encode the higher frequencies. You can define a high frequency cutoff for the codec in Adobe Audition. I'm not sure if you can acomplish the same thing by just filtering all the highs prior to encoding. Perhaps. Let me/us know.