Study: iPod generation prefer MP3 fidelity to CD

[Bernie]

Just another sign that the younger generation has no concept of good sound...

Jonathan Berger, Professor of Music at Stanford University, California, has conducted an eight-year study in which students have rated various audio formats while listening to the same song.
...
The Professor found over time the preference was for MP3 encoded songs, with those listening failing to establish any loss in audio quality normally associated with compressed digital music.

Read the whole report and weep:
iPod generation prefer MP3 fidelity to CD says study - iPod/iPhone - Macworld UK

[BrunoRepublic]

Sadly, it doesn't surprise me.

I come across people who are holding out for lossless downloads (or even hi-rez), who point out that there's no reason why these can't be offered.

Seems there's no significant demand for them (much like SACD and DVD-A). Most people can't hear the difference, or simply don't care.

I can't hear above 17KHz, which is where the lossy codecs do the most damage, so I jumped onboard when iTunes went 256kbps AAC without DRM. I figure that's about as good as things are going to get. You can't hold out for physical media anymore... too much stuff is download-only these days.

[Noman]

Worse than the Mp3 is the extreme remastering with limiters and dynamic compression etc.

And the most popular earphones are those that enhance the bass so "the classic V" (EQ curve) is like a Himalayan valley.

Those who listen to most modern music probably like eating a synthetic steak with bubblegum flavour and not knowing anything better :icon_confused:

[BrunoRepublic]

I recently did some experiments with different encoders to see exactly what these lossy encoders are taking out. Basically, I took an uncompressed WAV file and then encoded it to 256kbps AAC, and 256kbps MP3 twice: one with the Fraunhofer encoder, once with the much-touted LAME encoder. I then inverted the encoded files and pasted them over the original WAV, which let me hear exactly what was being taken out.

The results were rather surprising. As I expected, the 256kbps AAC fared best, leaving a low-level "shadow" of the original material, devoid of any identifiable harmonics, and sounding like slightly musical tape hiss. It's actually very similar to what I would be removing with digital noise reduction. To my surprise, it doesn't wipe out everything above 17KHz like MP3 does. Still, it took out more than I thought it would.

The 256kbps Fraunhofer MP3 faired considerably worse, taking out much more material. The "minus" file for this was much louder and brighter, indicating far greater signal loss.

The results for the 256kbps LAME MP3 were most surprising of all. First, it took me ages to line it up correctly, since the encoder pads the beginning of the file with several hundred samples of silence. Second of all, there was a very significant amount left in the "minus" file. Unlike the Fraunhofer encoder, the signal was more focused in the mid-range than the highs. This really shocked me because LAME is supposed to be so much better than Fraunhofer. My best guess is that it seems to be a lower bitrates because it doesn't suck out the highs as much, and apparently people notice that more than sucked-out midrange.

So, in conclusion, they all suck, but AAC sucks significantly less.

[kdavid13]

I recently did some experiments with different encoders to see exactly what these lossy encoders are taking out. Basically, I took an uncompressed WAV file and then encoded it to 256kbps AAC, and 256kbps MP3 twice: one with the Fraunhofer encoder, once with the much-touted LAME encoder. I then inverted the encoded files and pasted them over the original WAV, which let me hear exactly what was being taken out.

The results were rather surprising. As I expected, the 256kbps AAC fared best, leaving a low-level "shadow" of the original material, devoid of any identifiable harmonics, and sounding like slightly musical tape hiss. It's actually very similar to what I would be removing with digital noise reduction. To my surprise, it doesn't wipe out everything above 17KHz like MP3 does. Still, it took out more than I thought it would.

The 256kbps Fraunhofer MP3 faired considerably worse, taking out much more material. The "minus" file for this was much louder and brighter, indicating far greater signal loss.

The results for the 256kbps LAME MP3 were most surprising of all. First, it took me ages to line it up correctly, since the encoder pads the beginning of the file with several hundred samples of silence. Second of all, there was a very significant amount left in the "minus" file. Unlike the Fraunhofer encoder, the signal was more focused in the mid-range than the highs. This really shocked me because LAME is supposed to be so much better than Fraunhofer. My best guess is that it seems to be a lower bitrates because it doesn't suck out the highs as much, and apparently people notice that more than sucked-out midrange.

So, in conclusion, they all suck, but AAC sucks significantly less.

Thanks Bruno for this very interesting post.
:icon_cool:

[Bernie]

Bruno,

Thanks for looking into this and posting your observations. Of all people, I knew you would get to the bottom of things

I recently did some experiments with different encoders to see exactly what these lossy encoders are taking out. Basically, I took an uncompressed WAV file and then encoded it to 256kbps AAC, and 256kbps MP3 twice: one with the Fraunhofer encoder, once with the much-touted LAME encoder. I then inverted the encoded files and pasted them over the original WAV, which let me hear exactly what was being taken out.

The results were rather surprising. As I expected, the 256kbps AAC fared best, leaving a low-level "shadow" of the original material, devoid of any identifiable harmonics, and sounding like slightly musical tape hiss. It's actually very similar to what I would be removing with digital noise reduction. To my surprise, it doesn't wipe out everything above 17KHz like MP3 does. Still, it took out more than I thought it would.

The 256kbps Fraunhofer MP3 faired considerably worse, taking out much more material. The "minus" file for this was much louder and brighter, indicating far greater signal loss.

The results for the 256kbps LAME MP3 were most surprising of all. First, it took me ages to line it up correctly, since the encoder pads the beginning of the file with several hundred samples of silence. Second of all, there was a very significant amount left in the "minus" file. Unlike the Fraunhofer encoder, the signal was more focused in the mid-range than the highs. This really shocked me because LAME is supposed to be so much better than Fraunhofer. My best guess is that it seems to be a lower bitrates because it doesn't suck out the highs as much, and apparently people notice that more than sucked-out midrange.

So, in conclusion, they all suck, but AAC sucks significantly less.

[paul]

Interesting and I'm with you guys on this.
It took a while for me to warm up to cds and now I find most that I want acceptable enough to get them over vinyl.
What I find especially disturbing on some cds is the compression and limiting as Noman mentioned. I've look at the dynamic range on some of these cds and it seems some of 'em are running in a 10-15 db window for the most part. They make some of my vinyl originals look like the advanced format whose dynamic ranges were much greater.

Getting back to format preference, I blame it on the excessive greed we've seen in the last 30 years. My understanding is with SACD you can't make exact copies the way you can with cd. Also, they were expensive and you couldn't play them on your existing cd players. I believe DVD had some similar issues.
Compare this to when stereo was coming on in the 50s and 60s. You could still play them on your existing record players. In fact (here comes a rant:icon_mad:) I have a lot of problems with what seems to be greed driven formatting. Blu-ray is another one. It will probably make it but to use it I have buy a new player to play it, and the nice ones aren't cheap. Buying a bluray disc is useless if you don't have the player though I heard a month ago the Japanese came up with a dual format disc.
Compare this with the introduction of color. You could still use your old black and white because the new color format was put in a sub-carrier so it was no issue for older B&W technology.
OK, I'm done bitchin'

[BrunoRepublic]

Interesting and I'm with you guys on this.
It took a while for me to warm up to cds and now I find most that I want acceptable enough to get them over vinyl.
What I find especially disturbing on some cds is the compression and limiting as Noman mentioned. I've look at the dynamic range on some of these cds and it seems some of 'em are running in a 10-15 db window for the most part.

10-15db? Where are you finding these incredibly dynamic CDs? Most anything I've seen mastered in the past ten years is about half of that. :/

My understanding is with SACD you can't make exact copies the way you can with cd.

I think that's more because nobody's developed any equipment to do so. You couldn't make exact copies of CDs for the first 10 or so years they were around either. :)

Also, they were expensive and you couldn't play them on your existing cd players.

Most SACDs are "dual layer", meaning that they have a hi-res layer readable by the SACD player, and a plain-old red-book layer that any CD player can use. In fact, the first edition of the last round of Rolling Stones remasters were hybrid SACDs and they deliberately didn't mark them as such so people wouldn't be confused.

But yeah, aside from a small jazz and classical niche, the format is dead. Most people can't hear the improvement in fidelity or don't care. The winning format has always been the most convenient, and you can't get much more convenient than MP3.

[paul]

Thanks for certain clarifications Bruno