Neil Young talks about music quality

[rbrucecarter5]

http://www.cnn.com/2012/01/31/tech/web/neil-young-apple-high-def-music/index.html?hpt=hp_bn6

So what does this have to do with HD radio? Plenty. This is another situation where consumers are offered quality better than MP3, and have "little interest" as the article said. Unless you are willing to pay extra for premium earbuds or headphones, you will not notice the difference between an MP3 and vinyl. There just isn't enough quality in the signal chain to make up for the miserably small drivers in the earbuds.

There was also no mention of HD radio. If this higher quality download initiative had any consumer acceptance, then it could be coupled with HD radio just as Blu Ray and HD TV are coupled. But there is a problem there, too. Ordinary DVD's look so good on Blu Ray players and HD TV's the Blu Ray format is stagnant with consumers. It has picked up a bit, but the Blu Ray and DVD products are still coupled in one package so DVD's can be played in the car. Or more likely given away / sold.

 

[spunker88]

Bitrate matters more than codec at least for me. In MP3 anything encoded at less than 100 kb/s sounds noticeably low quality. For me MP3s encoded around 320 kb/s sound pretty darn good.

The problem with FM HD Radio is it only provides up to 150 kb/s of bandwidth in hybrid mode. HD Radio uses HE-AAC which is a derivative of AAC, the preferred format of many online streams. AAC does better at lower bitrates than MP3, compare a 32 kb/s MP3 stream and a 32 kb/s AAC stream and the AAC sounds better. But there is still heavy compression going on with anything lower than around 100 kb/s. I find stations running a single HD-1 bearable to listen to in HD, but its pointless to only have 1 HD channel. Add another HD stream and the bandwidth gets divided in half.

But if people are okay with listening to a 32 kb/s AAC stream on their smartphones then they would have no problem with HD Radio if they want the content. And to be honest if HD Radio gave a reliable signal that covered the same area as the analog signal (I'm talking a lot more than just 60dbu) I wouldn't mind listening if they exclusively had a format I liked.

 

[pianoplayer88key]

Bitrate matters more than codec at least for me. In MP3 anything encoded at less than 100 kb/s sounds noticeably low quality. For me MP3s encoded around 320 kb/s sound pretty darn good.

The problem with FM HD Radio is it only provides up to 150 kb/s of bandwidth in hybrid mode. HD Radio uses HE-AAC which is a derivative of AAC, the preferred format of many online streams. AAC does better at lower bitrates than MP3, compare a 32 kb/s MP3 stream and a 32 kb/s AAC stream and the AAC sounds better. But there is still heavy compression going on with anything lower than around 100 kb/s. I find stations running a single HD-1 bearable to listen to in HD, but its pointless to only have 1 HD channel. Add another HD stream and the bandwidth gets divided in half.

But if people are okay with listening to a 32 kb/s AAC stream on their smartphones then they would have no problem with HD Radio if they want the content. And to be honest if HD Radio gave a reliable signal that covered the same area as the analog signal (I'm talking a lot more than just 60dbu) I wouldn't mind listening if they exclusively had a format I liked.

I typically use mp3 for my music (no copy protection, more widely compatible than ogg vorbis (another format I'd otherwise consider which I understand is more efficient), and I know how to encode to get the results I want) and it works for me. I usually will force the lowpass filter a bit lower, or force the bitrate a bit higher than what my LAME program recommends - this helps cut down on the tinkly effects on the highs, except maybe at very low bit rates.

As for HD vs analog coverage, I'd go WAY beyond the 60 dBu for the goalpost coverage to match. If, using a high-end communications receiver (one in which just the materials to make ONE IF filter costs the factory upwards of US$500k in 1800s dollars) plus a comparably more elaborate antenna than a ferrite sleeve loop with a full-wavelength coil diameter, you can barely detect a QRSS CW carrier from the analog side of the broadcast (even if it's too weak too actually ID), you should have perfect decode from the HD broadcast on a $1 Coby iPod shuffle-sized radio using only its built-in antenna. Also the HD should be spectrally efficient (a fraction of the analog RF bandwidth for better audio quality, as well as steep enough bandpass filters so a crystal set's antenna can get DX on the next channel over even when touching the HD's transmit antenna (and I don't mean a xtal set with elaborate tuned circuits)) and not interfere with analog stations on its channel. For example re: range, rbrucecarter5 should be able to get a perfect quality signal from KFI 24/7, possibly even when his receive antenna is physically touching WSM's transmit antenna. (I seem to remember he posted somewhere about hearing KFI midday one winter, but I can't remember if it was here or on ABDX. I do realize your reception was in TX somewhere, but based on the quality and the equipment you used, would you think a setup like I described above might have a chance at daytime KFI reception in Nashville?)

 

[tomservo]

I prefer the way AAC alters the sound to comparable-quality mp3s, but HE-AAC just doesn't cut it for serious listening. I know if I bought my HD radios for something I really wanted to hear and not just geek factor, I would be sorely disappointed. But then I'm sorely disappointed in the way FM stereo sounds. (But I strangely love C-QUAM stereo, go figure.)

Pretty much all music I have collected on my PC is in FLAC format, a lossless compressed scheme that is about half the size of the raw wave file while not altering the sound at all. With today's huge storage devices there is no excuse to use ANY compressive encoding at all. Even though my ancient beloved iRiver has only a 20 GB HDD (big for its day!) I still manage to cram plenty of FLAC files on it for the more demanding music.

And yes, I can blind-test tell a difference between a 320 kbps mp3 and FLAC/uncompressed. Once you know what to listen for it's pretty obvious in my opinion. I'm less able with my preferred compressed format which is .m4a, which I guess is some mp4/aac type deal. I use FAAC MP4/AAC Encoder v1.28 via BoncENC, with the following settings: 100% quality, maximum quality 22 kHz, joint stereo on, temporal noise shaping on, mp4 type LC. None of the other encoder settings for type (LHX, MAIN, LC) or file extension (m4a or aac) would work, but it sounds good for music and all my podcasts I re-encode. Much of my music is transparent enough for portable listening in the car in this format, but some music still shines light on lossy compression's foibles. It also yields a noticeably smaller file than equivalent mp3. Up until I embraced FLAC, I have always had a soft spot in my heart for what I thought was the best lossy compression, that being ATRAC+, used by Sony in their minidisc players. It wasn't necessarily the truest to the original, but the way the artifacts were hidden and the system optimized for music by people who know both the depths of music AND how the mind perceives sound made it the best overall choice for compressed music, imho. But Sony being Sony they kept it to themselves for their proprietary hardware so it never got into the mainstream.

Of course, all this is so far over and beyond the encoding technology used for HD, that IBOC just doesn't compare, even when a station is running just the one main HD channel

 

[SirRoxalot]

No matter how you encode it, today's recordings that have been compressed into a brick-like waveform still sound like crap.

 

[K6JHU]

I have new modern CD's that I have put into MP3 (192 Kbps). I also have black vinyl that I did the same to. And you know, the black vinyl still sounds better. I believe that in the days of black vinyl the engineers CARED about what went on the tracks. With a CD, it seems that it is so easy, the engineers don't give a d*** about the sound.

 

[tomservo]

K6JHU I get the impression that many engineers *do* care, they are just overruled by the record executives or recording artists. But then I've also heard some recording artists decry the poor sound quality of their own releases, so who knows where the blame really lies.

Well no, I know where it lies: with consumers, who tolerate this brick walled garbage.

The nice thing about transferring vinyl over to a digital format is you aren't constrained by the CD's 44.1 kHz, 16-bit sampling rate. Even though I have a sub par cartridge on my old dilapidated turntable, I record everything at 48 kHz, 24-bit. And it does seem to make a subtle difference, especially when downconverting to a lossy format, where the bitrate and sampling rates are still preserved.

Of course, that requires a sound card of good quality and a mp3 player that can handle those specs.

If I had the turntable of my dreams and a really low noise pro sound card, I would probably record everything at 96/24 just for the heck of it. My iRiver has no trouble playing that back (although it may downconvert to 48 or 88 kHz.)

 

[K6JHU]

BTW, I also took one of my old 78's and processed the h*** out of it (including some manual edits) to get rid of the surface noise, clicks, and pops. The amazing thing that since a 78 was spinning so fast and the grooves were so wide, the sound quality was incedible once the surface noise was eliminated. I am not planning on doing that again since it took me about forever

 

[KB1OKL]

I have new modern CD's that I have put into MP3 (192 Kbps). I also have black vinyl that I did the same to. And you know, the black vinyl still sounds better. I believe that in the days of black vinyl the engineers CARED about what went on the tracks. With a CD, it seems that it is so easy, the engineers don't give a d*** about the sound.

I bought a pristine mono high fidelity WB Everly Brothers album from the mid 60's last week and with no surface noise the sound of it is fantastic. When you listen to something like this on a good Turn Table and system it makes obvious just how distorted, compressed, loud and unmusical most new digital recordings are.

 

[Darth_vader]

@K6--

WHY are you encoding your records at 192, and in MP3 yet? (Yuck...) That bitrate/codec may be fine for speech or music-on-hold content for the f0ne, but for serious listening you shouldn't use anything less than 48000/384 MP2!

And Zach, CD's 44100 Hz sampling rate/16-bit sample depth *is* a digital format, just not one that's as flexible as others are. Of course, once it hits the amplifier and loudspeakers it's just as analogue as anything else.

 

[scanman1]

I encode MP3s at 320K/48Khz for vinyl (to keep it compatible with my MP3 player) and it sounds pretty freakin good!

 

[spunker88]

I bought a pristine mono high fidelity WB Everly Brothers album from the mid 60's last week and with no surface noise the sound of it is fantastic. When you listen to something like this on a good Turn Table and system it makes obvious just how distorted, compressed, loud and unmusical most new digital recordings are.

The loudness problem is one that really annoys me, I was shocked at how overly loud modern MP3 music looks when you analyze it in Audacity. There are sacrifices we have to make in audio quality in order to fit thousands of songs on a tiny MP3 player, but overly loud and distorted music shouldn't be one of them.

 

[Darth_vader]

That's just it--it isn't. Loud and distorted is an indicator of sloppy technique on the part of the operator doing the encoding, nothing else. A 3-minute discrete stereo CBR (dare I say it?) MP3 at 48000/320 with all samples compacted to full-scale uses the same amount of disk/disc space as one encoded to the same parametres but with peaks at -50 dB!

Hence, there is no technical function whatsoever to excessive loudness. All it does is make the recording more grating to have to listen to.

 

[hubcity]

Amplifying Darth's comment (heh), the loudness wars are really an artifact of labels not wanting to release records that were the quietest in the CD changer. Now, I'd say some are trying to tweak loudness on a track-by-track basis, since that's how music's bought by-and-large nowadays. Waveforms of those files just make you want to cry.

On my iPod, their efforts are for naught. I run MP3Gain and all those files play back at roughly the same volume (and thanks to the smart engineers that developed the MP3 format, it's a change that only changes the gain setting in each MP3 frame, thus adding no generational quality loss to the file. Neat trick.)

Still, if a vinyl version of a release is available, I like to encode off that. It's slower, sure, but it sounds way better, and the waveform looks better, too.

 

[pianoplayer88key]

I'd also like to know what happened to dynamic range in recordings. A lot of songs I hear are pretty much the same volume throughout the entire song. Some classical music recordings, for example, will have a lot of dynamic range. I've heard songs that had music so loud it almost hurt my ears (in this case because it was loud, not because it was overdriven), and another part of the same song was so quiet I could barely hear it (and if I looked at the waveform I probably would have had to zoom in for it to not look like a flat line).
Are any popular music artists in the last 5 years producing and releasing music with extensive dynamic range like that?
Another thing I don't like about a lot of modern popular music I hear is the monotony of the chord structures/progressions, among other things. Sure, some songs are better than others, but for the most part you have the same progression of 2 or 4 chords repeated throughout the entire song. Some popular music from the 50s or around that time (especially the slower ballads in 12/8 time I think) I've heard have more variety on the music behind the song, but even then they can't touch the variety in some classical music. (BTW I'm not generally a classical music fan per se, but there are some pieces of music that I do enjoy.)

 

[tomservo]

So which came first: overly loud music recordings, or weak amplifiers that can't play more dynamic music at appropriately loud levels?

I can't help but think music will continue to get louder as portable devices like mp3 players and cell phones get smaller and less high quality amps in them.

 

[JohnnyElectron]

In the 70's, we really cared about high quality audio - big ass speakers, reel to reel decks with a ton of headroom and quality at 7-1/2IPS, turntables with a strobe and low wow & flutter; most kids nowadays aren't like us relics that like this HQ stuff. I too really like the sound that CQuam stereo AM puts out, and I've always considered Sony's last version of ATRAC and excellent 'cassette replacer' for home recordings. I'll take all of the above over MP3's and ipad anything.
Wish Neil Young was able to work with Steve Jobs (who had tons of vinyl, yet sold crappy MP3's) on making the 2nd coming of HQ analog-sounding audio again!

 

[Darth_vader]

*High-quality* audio, or analogue anything? From Apple?

Keep dreaming. You're better off just encoding your own content in LAME (another little gem Cupertino would rather you didn't know about) or even better yet, just using straight PCM.

[size=8pt](Which "cassette replacer" would that be? Minidisc or Elcaset?)

 

[SirRoxalot]

One of the big differences between vinyl & digital recording is that vinyl has a maximum volume level that can be recorded. You can only cut the plastic so deep. Great care must be taken during mastering to work within the parameters of the medium. Dynamic range and accurate recording of instrument sounds was more highly prized. Processing came after the recording was played. These days, once music is digitized, it can be manipulated in ways undreamed of during the vinyl era, and loudness is simply one of the manipulations that are possible with digital.

 

[Darth_vader]

Not necessarily. First off, by "vinyl era" I'll assume you mean the "analogue era" which actually never ended, nor will it ever end. (Regardless of the storage and encoding method, once the audio signal gets to the amplifier and loudspeaker, it's all analogue.)

That said, I've heard tapes that were recorded so loudly (i.e. so oversaturated) they would put the player's amplifier into clipping. Their waveforms, when sampled, appeared very much like some of the crap CDs you get these days, only slightly more "ragged" and warmer (no waveform is perfectly square in the analogue domain.) Such manipulations were (still are) possible in the analogue domain with certain recording formats if you push them far enough. It wasn't until digital-domain recording formats came along that people really started to rape and abuse such techniques.

[size=8pt](Even on "vinyl" [contrary to popular belief, my generation actually prefers to call them "records", too] hypothetically you could get unlimited volume levels by using [1] a thicker disc to make a deeper cut into and [2] consequently, a longer and wider needle to actually play back such a record. Just something to ponder...)