MP3s, how and why they affect our systems, and system choices

[sandt38 448]

It has been debated rather hotly for years, and now it has become pretty much accepted that MP3 and sound quality do not mix. But why is it? Why do some people hear so much loss in their systems and others don’t? Are my old ears really better then someone 30 years my junior? Someone who still has a large amount of their hearing range left, while mine has deteriorated over the years should be able to hear these losses better then I!

I started with home tower speakers in the back seat of my old Delta88 back in the days when we had no real systems, and we still had post mounted tape decks and 8-tracks. Then times changed, and we started putting in tape decks with trunk mounted CD cartridge changers with $60 coaxials and $50 amplifiers and maybe a bazooka tube subwoofer for those of us with “discriminating” tastes. Eventually I wanted to just go boom and spent money on several thousand watts and some monster excursion subs. Then we learned about XBL2 and how these high output subwoofers could be very accurate. Finally I finished my journey with $1200 component speakers and $1000 amplifiers, all filled in at the bottom with very capable sound quality based subs dialed way back. My journey started out as just wanting to get some sounds, to wanting to get loud, to wanting to sound good.

Now that I have found a really nice sound quality install, my music seemed to really lack. I have been using MP3 for years, but eventually I started looking for higher bit-rate MP3s, and now I am at a level where only .flac or .wav or other lossless files suit my tastes. I have listened very critically to how the advancement of definition brought about by higher quality speakers and amplification, and source caused my MP3 listening experience, no matter the bit-rate to become bland, and I really felt more of the emotion brought out in music by using lossless formats in my car.

But then I asked myself, “Why does my lossless not only sound better, but seem to bring about a more emotional experience?” So I started to learn about how lossy formats work, and how they can stick so much information in such a small space.

Let’s discuss music reproduction first, so we can get an understanding of what we are listening to. All we are listening to is really a reproduction of an instrument or a voice, and to simplify this writing I will include the vocals as the instrument as well. Musical reproduction is usually characterized as fundamentals, harmonics (and as I will discuss below, low fundamentals can be included in this grouping), and finally overblow, breath, or air. The very heart of these sounds are called fundamentals. The fundamentals, are what I call meat and potatoes or music, are the essentials of the reproduction. They are what the instrument actually delivers, directly. But then we have carry over, which are termed harmonics. These are not as much a base of the delivery of the instrument, as they are an extension of it. You might also see the lower frequency carry overs dubbed as a “low fundamental”, as they are more apparent in low frequencies, where they tend to be less noticed in higher pitches. I like to call this the emotion of the music. Finally you have overblow, which is usually only noticed in wind instruments and vocals. I personally like to classify this as emotion as well, and therefore will include them as a group when I say emotion or harmonics.

While it takes a trained ear to listen for image and stage presentation, younger people with a naturally better range can hear losses in the “emotional” frequencies or the harmonics of the sound easier than us old fogies. So why is it that those of us who are older tend to hear these losses better then the young folks? I would dare to venture that it is simply because we tend to have more money invested in our component systems (and when I say system, I mean everything from source material, source equipment, sound processing, amplifiers, speakers, and installation techniques) with a higher definition then the younger folks who are running a simple inexpensive component system install. This is not to say our systems are better for all purposes, as I will discuss in this writing. In fact, for the vast majority of folks, a lower definition system makes more sense, as many are more concerned with the ease of storage media and they tend to use MP3 or other lossy formats. When we have these high definition systems, many of the weaknesses inherent in lossy encoding really jump out glaringly.

So, sandt38, what is it that we really listen to? Well, the human brain perceives and retains everything it sees, hears, thinks, etc. It is actually an amazing organ that captures far more then we are able to drag out of it. We simply do not have all the doorways opened up in the brain, so we forget, or don’t notice many of the things that we actually perceive. The mind actually only directly notices the loudest things we hear. So while there may be several instruments may be reproducing the same frequency at the same time, we only perceive the loudest of this information we actually hear. We ignore the other lower volumes of the same frequencies, but we still hear it. Our mind processes this information but determines that it is not that important so it will file it away in that subconscious part of the brain we do not actually consciously notice.

So let’s look at how these MP3 encoders work. In order to save space, an MP3 encoder usually selects certain information to discard, and other information to retain. The most common method of MP3 encoding functions similarly to how we perceive the music. Just like our brain, it ignores the lower volumes of the same frequencies, and all it retains it the higher volume material at that frequency. We fill in the rest subconsciously. Please be aware that many encoders eliminate only certain lower frequencies altogether, but they variably retain the higher volume information up to a certain point. For example, the encoder may retain ½ of the information at a certain frequency. But it will retain the higher volumes at a greater rate. While these are not the only methods of encoding MP3s, they are the most common.

So you see, by considering the way the encoder works we can get an idea of what happens to the music we listen to. The fundamentals tend to get retained, while the harmonics tend to be lost. So we lose the emotion, which is why MP3s sound so bland on systems with very high quality components.

Now, there are other things to consider when it comes to MP3 compression. Bit-rates, and not only the rate, but variable or constant bit-rates, are the biggest consideration here. Very quickly, the higher the listed bit-rate, the more information that is saved, and therefore the better the quality. But to be frank, these don’t alter what occurs with the music. Yes, these higher bitrates do sound much better, but there is still something that is lost.

So what can we gather about speaker selection from our source material? Well, we can help decide whether it is really worth it for us to spend $1000 on components, or if a $200 set will suffice. I am not trying to say that if you spend a grand on a set of components it will automatically be a better set then a $500 set. My mention of pricing basically assumes a basic cost/quality ratio. Clearly a $200 set of Sony Xplod coaxial speakers is nowhere near as good as a $180 Image Dynamics CTX65CS component set. So take these generalizations to assume a similar cost/quality ratio. The high dollar sets that you see out there are far more true to source then the lower priced units. This is a good thing if the source material is pristine. But if your source is not that great, these more accurate speakers will start to sound like crap. It isn’t the speakers that suck, it is the material you are putting through it. But when we get down into the lower dollar sets, these weaknesses are not quite so apparent.

I do want to point out this article, no matter how poorly the whole test was conducted the results still prove my point. The guy used some of the most awful source and speaker combinations. Why you would ask people to listen for nuances using a $200 receiver and a $200 subwoofer and satellite set using 5 3.5 inch full range units is beyond me. Sheer idiocy. But the end results are very significant, the differences between MP3s and lossless were detected by all listeners in the test.

All this considered, please keep a few other things in mind. I cannot stress the importance the rest of your signal chain has on the final product. The higher end speakers will also show off weaknesses in them as well. Be prepared to make some really high dollar signal chain investments when you decide to buy these high end components.

I also think it should be noted that while amplifiers of the same class may sound different while sitting still with little ambient noise, but once we get moving these nuances really disappear. I do feel that differing class amplifiers can have a different overall sound, and can be distinguished while moving.

[Tirefryr 1,742]

I can't tell a difference between 320K and a standard recording on CD. I can tell you that many producers are recording at higher than normal levels and some even adding distortion. The quality of original recordings today is horrible anyways, so I don't see the issue with a move to mp3 or otherwise aside from "audiophile" recordings.

[manrossdamn 57]

That, was, an, amazing, read.

[Bob345 39]

That, was, an, amazing, read.

X2

[Sir-Lancelot 216]

Nice post. Thanks!

I use mostly MP3's now mostly for the convenience of being able to put a shit load of them on the thumb drive. With my current set up I cant tell much difference, but when I had the Dynaudio installed I could.

Was not warm and detailed like a good recording was. Psychoacoustics? Maybe/Maybe not.

[Shogen 325]

I use highest bit wav I can with my flash drives... Using a 32gb flash currently... Good times...

Good post as well!

[diablo_331 23]

Great post!

[j-roadtatts 1,129]

Good write up Seth. You describe my musical journey to a tee, starting with the house towers. lol.

I think the reason elders have a more refined ear is, once our brain hears the music reproduced so accuratly that it envokes emotions, we don't want to except anything less.

Ignorance is bliss I guess.

[sandt38 448]

I can't tell a difference between 320K and a standard recording on CD. I can tell you that many producers are recording at higher than normal levels and some even adding distortion. The quality of original recordings today is horrible anyways, so I don't see the issue with a move to mp3 or otherwise aside from "audiophile" recordings.

I did a blind test with my wife in my car. I made 2 copies of the same track, same media, 1 in 44.1khz .wav and the other in 320VBR MP3. She perceived a difference between the .wav and MP3, but she couldn't really pinpoint the difference. I didn't offer any thoughts on the matter at first, just asked her which sounded better and why. Her preference was for the .wav file. I asked her after she made her observation if she just felt more emotion in one rather than the other. She replied that the .wav just felt more "alive".

My wife is no audiophile by any means, but she does have a lot of nice gear around her. She even has a killer system in her vehicle, and of course my home system and car systems are no slouches either, so she knows what sounds good.

I agree that most of today's music is bloated during mastering. I hate those "remasters" that sound like ass. If I want it louder, I will turn up my stereo, I don't need some studio engineer to bloat the sound and call it better. But alas, I listen to older music, so it isn't really an issue for me. Aside from TOOL or APC I don't listen any music from this decade unless someone else wants to hear it. IMO, not only is the music tainted by poor recording technique and mastering, but it just plain sucks bawls.

[sandt38 448]

Nice post. Thanks!

I use mostly MP3's now mostly for the convenience of being able to put a shit load of them on the thumb drive. With my current set up I cant tell much difference, but when I had the Dynaudio installed I could.

Was not warm and detailed like a good recording was. Psychoacoustics? Maybe/Maybe not.

Ironic you would specify psychoacoustics... that is exactly what MP3 plays on.

[Aaron Clinton 2,372]

Very nice topic, got me thinking.

[helotaxi 46]

I think that the main reason that older people are more likely to notice the subtle differences it that we have actually heard good recordings of real musicians that were well mastered. The younger generations have only ever had shit from performance to recording to mastering. Rap, the current garbage that is passed off as R&B, top 40, basically all of what passes for mainstream music these days is so generally atrocious that listening from a CD or a low bit-rate mp3 is no different. The master recording is flat, utterly lacking in dynamic range. The "music" is electronically produced, the vocals get a lot of "engineering" work done to them as well. The overall result is that there is almost no emotion in music because there is so little actual human element left. While a fantastic recording of a great performance by a true artist is the dream, I'd rather have a mediocre reproduction of an OK recording of a great performance by a true artist. I'll listen to good musicians on satellite radio or my iPod on cheap speakers over what my wife listens to regardless of medium or reproduction equipment.

[Kanaworm$ 292]

Very interesting and informative read!

[kirill007 52]

Can you actually hear the difference between a 1000kb/s mp3 and just straight from a cd? ( don't forget to copy both, the original so you need to burn it again on you're own cd's and then burn mp3. That's maybe why you lose quality because you can't burn the cd's properly)

I have never heared difference between types of speaker cable or like this a 1000+kb/s mp3 and just cd quality.

And i've listened it on B&W 802 speakers and a pair of monoblock's.(can't remember brand)

What music do you listen to that u can hear the difference?

[Tirefryr 1,742]

There's no way to answer that. What one person hears or discerns will be different from others.

[sandt38 448]

I think that the main reason that older people are more likely to notice the subtle differences it that we have actually heard good recordings of real musicians that were well mastered. The younger generations have only ever had shit from performance to recording to mastering. Rap, the current garbage that is passed off as R&B, top 40, basically all of what passes for mainstream music these days is so generally atrocious that listening from a CD or a low bit-rate mp3 is no different. The master recording is flat, utterly lacking in dynamic range. The "music" is electronically produced, the vocals get a lot of "engineering" work done to them as well. The overall result is that there is almost no emotion in music because there is so little actual human element left. While a fantastic recording of a great performance by a true artist is the dream, I'd rather have a mediocre reproduction of an OK recording of a great performance by a true artist. I'll listen to good musicians on satellite radio or my iPod on cheap speakers over what my wife listens to regardless of medium or reproduction equipment.

I think you bring up some very valid points. My wife likes Enrique Iglesias, she thinks his voice is so great. I told her it was engineered, you can hear it. Of course, she get's bent. But you are right. Even lossless it sounds like ass.

Can you actually hear the difference between a 1000kb/s mp3 and just straight from a cd? ( don't forget to copy both, the original so you need to burn it again on you're own cd's and then burn mp3. That's maybe why you lose quality because you can't burn the cd's properly)

I have never heared difference between types of speaker cable or like this a 1000+kb/s mp3 and just cd quality.

And i've listened it on B&W 802 speakers and a pair of monoblock's.(can't remember brand)

What music do you listen to that u can hear the difference?

1000kbps? Sorry, this is not a real level of MP3. Some mp3 generators may claim high bit rates, but that is greatly exaggerated. Non-standard bit rates up to 640 kbit/s can be achieved with the LAME encoder and the freeformat option, although few MP3 players can play those files. According to the ISO standard, decoders are only required to be able to decode streams up to 320 kbit/s

I am not usually burning CDs, I use HDDs, but yes I can burn them properly. I use only high quality optical media (Taiyo Yuden), and i use only quality ripping and burning programs on a solid Lite-On burner. I am an avid burner (I have almost 2000 DVDs) and I also design and print all my own labels and covers. People here who know me will tel you, I don't do anything 1/2 way.

Anyways, as i mentioned, I placed both a 320MP3 and a .wav on the same media, same song, and played it blind for my wife... read above.

I listen to everything just about. From old school blues, to classical, to R&B, rock, metal, death metal, alternative, you name it. I can say though, outside of TOOL, APC and maybe Norah Jones, I have no music that I personally listen to from this decade.

[kirill007 52]

1000kbps? Sorry, this is not a real level of MP3. Some mp3 generators may claim high bit rates, but that is greatly exaggerated. Non-standard bit rates up to 640 kbit/s can be achieved with the LAME encoder and the freeformat option, although few MP3 players can play those files. According to the ISO standard, decoders are only required to be able to decode streams up to 320 kbit/s

I am not usually burning CDs, I use HDDs, but yes I can burn them properly. I use only high quality optical media (Taiyo Yuden), and i use only quality ripping and burning programs on a solid Lite-On burner. I am an avid burner (I have almost 2000 DVDs) and I also design and print all my own labels and covers. People here who know me will tel you, I don't do anything 1/2 way.

Anyways, as i mentioned, I placed both a 320MP3 and a .wav on the same media, same song, and played it blind for my wife... read above.

I listen to everything just about. From old school blues, to classical, to R&B, rock, metal, death metal, alternative, you name it. I can say though, outside of TOOL, APC and maybe Norah Jones, I have no music that I personally listen to from this decade.

Oh, i'm sorry i always use wma files,not mp3 so i'm sorry for the misunderstanding.(so not mp3, i'm wrong, but my mp3 can read it, i forgot my settings)

And you can hear a difference in metal,death metal between a mp3 and wav? (because most metal,death metal i've heared isn't mastered properly in the studio.)

How do you get you're mp3?

Because the difference between when you rip a cd and just rip it in 320mp3, or ripped it in full quality and then used a converting program, well the last one is the worst. (or even waaayyy worse if u download songs. )

Actually i haven't really tried out the 320mp3 and a normal wma lossles , so i'll try it tomorrow.(do you have any songs in particular for that in mind? i thought some micheal jackson songs, because those were mastered properly.)

(i'm sorry because i can be quite sceptical, and if i can't hear it myself or when a bunch of people tell me it's right i only agree with things.)

Edited December 18, 2010 by kirill007

[Randal Johnson 544]

Interesting thread ...

[sandt38 448]

Oh, i'm sorry i always use wma files,not mp3 so i'm sorry for the misunderstanding.(so not mp3, i'm wrong, but my mp3 can read it, i forgot my settings)

And you can hear a difference in metal,death metal between a mp3 and wav? (because most metal,death metal i've heared isn't mastered properly in the studio.)

How do you get you're mp3?

Because the difference between when you rip a cd and just rip it in 320mp3, or ripped it in full quality and then used a converting program, well the last one is the worst. (or even waaayyy worse if u download songs. )

Actually i haven't really tried out the 320mp3 and a normal wma lossles , so i'll try it tomorrow.(do you have any songs in particular for that in mind? i thought some micheal jackson songs, because those were mastered properly.)

(i'm sorry because i can be quite sceptical, and if i can't hear it myself or when a bunch of people tell me it's right i only agree with things.)

Most metal and death metal isn't mastered properly? Due to lower budgeting they tend to be less manipulated then mainstream music like Michael Jackson. IMO it is far better raw then it is overly manipulated. I prefer live recordings and older school recordings where "mastering" isn't total manipulation. See helotaxi's commentary above, which i agree with completely.

I might suggest looking into Alice In Chains Unplugged in NY. Great staging, imaging, and because there is no real manipulation, it is very clean and realistic. Other discs worth looking into would be Pink Floyd Wish You Were Here or Dark Side of the Moon (not a remastered version either), Norah Jones Come Away With Me

You can be skeptical all you want. If you do some research yourself you can see how compression works yourself. There is very little to be skeptical about, it is just a fact. As i suggested in the article, if you have average gear, you might not hear much at all. If you have nice gear in a nice install, it will be apparent.

[kirill007 52]

I couldn't notice a difference between 320kpbs MP3 and cd-quality.

But i could notice a difference between 256kps and cd-quality, but i think i used a bad song.

I haven't found the cd from alice in chains unplugged, nor Dark Side of the moon or that other one.

I think i have the pink floyd cd, but didn't find it yet.

My father nor mother could hear any difference between 128kbps MP3 and cd-quality...Nor that guy with his high quality 'expensive" Bose equipment..

[sandt38 448]

Uncompressed audio as stored on an audio-CD has a bit rate of 1,411.2 kbit/s, so the bitrates 128, 160, 192 and 320 kbit/s represent compression ratios of approximately 11:1, 9:1, 7:1 and 4.4:1 respectively. So you are hearing less than 1/4 of the information at 320. Still doubt that it is audible?

[Aaron Clinton 2,372]

Just shared this topic via our twitter.

[sandt38 448]

Just shared this topic via our twitter.

I do expect a deposit in my paypal account, of course

[Tirefryr 1,742]

Uncompressed audio as stored on an audio-CD has a bit rate of 1,411.2 kbit/s, so the bitrates 128, 160, 192 and 320 kbit/s represent compression ratios of approximately 11:1, 9:1, 7:1 and 4.4:1 respectively. So you are hearing less than 1/4 of the information at 320. Still doubt that it is audible?

It may be audible, but I can tell you my brain can't distinguish a difference.

[kirill007 52]

Uncompressed audio as stored on an audio-CD has a bit rate of 1,411.2 kbit/s, so the bitrates 128, 160, 192 and 320 kbit/s represent compression ratios of approximately 11:1, 9:1, 7:1 and 4.4:1 respectively. So you are hearing less than 1/4 of the information at 320. Still doubt that it is audible?

Yes, and 1400kbps? (i saw from where u got it, but experience says different)

Highest bitrate that i have is 1106kpbs,most of it is about 950, and it depends on the song what bitrate it has.

So if u use the most common bitrate of 960 it is only a third that you hear.

And you don't need the freqeuncies above 20khz or below 20hz, so those freqeuncies are gone.

That's what they do, use the loudest freqeuncies and remove everything else.

I've noticed that instruments do sound a tiny bit different(piano), and that there seems to be less dynamics.

The freqeuncies that are really high are gone so it doesn't sounds as 'clear'.(i don't like to use subjective words.)

The 'sparkling" highs aren't there, couldn't notice no difference in bass nor midbass.

Haven't tried out a song with voices so i don't know if voices are affected but according to my theory there shouldn't be any change with how voices sound.

What do you hear that is different between a mp3 and cd-quality?

(not sceptical anymore, there is a small difference not worth it if you don't have good equipment or ears, and not needed for everyday listening.)

EDIT: i compared a 256kps MP3 to cd-quality.

Edited December 21, 2010 by kirill007