http://www.stylusmagazine.com/articles/ ... orever.htm

very interesting read. thoughts?

I had a skim through it - looks like a really great article. Very interesting. However it's so damn long! I just don't have the time to sit and read all that!

I had a skim through it - looks like a really great article. Very interesting. However it's so damn long! I just don't have the time to sit and read all that!

I read it from start-to-finish.

Excellent read to those missing out.

I had a skim through it - looks like a really great article. Very interesting. However it's so damn long! I just don't have the time to sit and read all that!

I read it from start-to-finish.

Excellent read to those missing out.

Yeah I skimmed it at first but then I actually read the first two paragraphs and was hooked to reading the whole thing >.<

I also had a skim, and it looked rather interesting.

I'll give it the "in-depth" later tonight.

I'm just reading through this now... I agree with lots of things he's saying, but a few things usually bug me with this kind of article-- It seems I just don't 'get' the ridiculously high standards of the audiophiles who hold this opinion, even though this guy claims not to be one. Don't get me wrong, I dislike the so-called Loudness War as much as any audiophile, but things like this:

"well, the payoff for having a million radio stations is that most of them transmit at 128kbps or less. I wouldn't listen to an MP3 encoded that badly, so I'm certainly not going to listen to radio at that bitrate."

...You'd think he's talking about a 32kbps mp3 of a song recorded with a cheap radio shack microphone from a dying 50 year-old radio or something. There are lots of noticeable differences in a 128kbps mp3, but it's so far from 'unlistenable' that it just makes this guy seem rather pretentious. Like I said, maybe I just don't get the audiophile point of view. Maybe I'm not fit to be mixing and mastering music if I can actually stand (and enjoy) listening to music at 128kbps or less, I dunno.

Here's another example of what I'm talking about:

"[...]some mastering engineers claim that a huge amount of professionally released CDs since the turn of the decade (and earlier) have been so compressed that they don’t even consider them to be “musical.” I can’t bear to play back some of my favourite records from the last few years through my hi-fi and pay them full attention, and this is upsetting."

I think that's going too far, myself. Sure, when I hear an over-compressed, ridiculously loud (or 'hot,' I suppose... ) record nowadays my reaction is basically a roll of my eyes and a slight downwards tweak of the volume knob. I acknowledge it sounds much worse than if it were 'properly' recorded, mixed and mastered, but it doesn't keep me from enjoying the music. Audiophiles have always pissed me off for these reasons, to be honest. Maybe it's just me.

Anyhow, once again, don't get the wrong impression. I very much agree with the title, 'Imperfect Sound Forever,' and most of the statements within.

EDIT: One other thing-- I do think clipping and over-compressing has its place, as crazy as that sounds. For example, a project I'm working on right now is intended to sound very ugly, semi-broken, and sort of the aural equivalent of an old junkyard. Also, for a lot of straightforward metal records, the consistent and loud production can work well, provided it's not so ridiculously overdone.

I do think clipping and over-compressing has its place, as crazy as that sounds. For example, a project I'm working on right now is intended to sound very ugly, semi-broken, and sort of the aural equivalent of an old junkyard.

That can be good alright. You hear it in the likes of Meshuggah or Fear Factory, and Dev does it sometimes too. It can sound daycent when it's done right like.

Speaking of articles, my big brother recently decided to share some of his thesis with the world. And I think he's a genius. So I'm going to post a bit of it.

"An Exploration of Multidimensional Timbre Vectors Using Synthesised Tones
The area of timbre perception research is one which has had a long, if patchy, history. As far back as von Helmholtz, there was speculation as to what made one note sound different to another if their pitches, loudnesses and durations were the same. However progress in the area has been sporadic at best, with years of inertia between bursts of advancement.

The procedure used for the second experiment was much the same as the first. Again, 27 stimulus tones produced a total of 351 pairs which were randomised before being presented to subjects, who were asked to rate how dissimilar each pair was on a nine-point Likert scale. Each subject completed the task three times; the first two were practice runs while the data collected on the third trial was used to construct the dissimilarity matrix. However at the end of this experiment, subjects were asked to write down what factors they had been utilising to determine how similar, or otherwise, the stimulus tone pairs were. This allowed the researchers something of an insight into the listening strategies employed by subjects.

Again, three dimensions were found which described the timbre space perceived by subjects in the second experiment. Dimension 1 separated tones into three groups according to how many harmonics were contained within the tone. This was the most salient dimension and had the highest eigenvalue. Dimension 2 separated tones containing five harmonics from those containing three or seven harmonics. The final dimension grouped tones by amplitude envelope type; the authors point out that the ordering of the stimuli along this axis suggests that the amount of time spent at maximum amplitude might be the critical factor in this dimension.

As well as these possible explanations provided by the authors, three more spring to mind. Firstly, although par for the course as regards multidimensional studies of timbre perception, a group of 25 subjects is not nearly big enough. This is especially true when this already small group is further subdivided into one group of seven composers and another group of eighteen non-musicians. Secondly, the fact that the two groups were presented with the stimuli in very different ways and in very different surroundings must have had an impact on responses. The fact that the non-musicians seemed to respond more predictably than the composers did seems to indicate that the optimum way to present stimuli in this paradigm is via headphones. Thirdly, and perhaps most controversially, three dimensions may not be enough to represent something as fundamentally complex as musical timbre. Krumhansl’s (1989) study, on which this work was based, was deeply compromised in that only nine subjects were used, all of whom were composers at IRCAM. Even going back as far as the early multidimensional researchers, a scan of how many subjects John Grey used gives the following underwhelming list of numbers of participants: 16, 20, 18, 15, 15, 17, 30, 13. Even the study that used thirty subjects found a possible four-dimensional solution which was rejected on the grounds of uninterpretability. Later, McAdams et al (1996), using a much more satisfactory 98 subjects, found that a six-dimensional solution was the most parsimonious available but this model was also rejected due to interpretation difficulties.

It is somewhat disconcerting that such interesting findings have been discarded due to their being difficult to understand in psychophysical terms. It appears that the status quo, in the form of a three-dimensional model of timbre perception, has remained in situ due to an unwillingness or inability on the part of researchers to carry out a sufficiently detailed multidimensional study. If such a program of research was to be carried out, perhaps a new world of possibility could be opened to electroacoustic composers and musicians in the form of synthesis packages which take scientific and multidimensional models of timbre perception into account. It is to be hoped that such a research program, employing many more subjects and a much wider range of stimulus tones, can be undertaken in the future and that new life can be infused into what currently appears to be a stagnant area.

One final idea that could perhaps be useful in future research is that studies could be broken up into smaller parts so that different kinds of timbres could be looked at individually as suggested by Wessel [ 1978]. For example, certain studies could look specifically at string timbres, others at woodwind timbres and others still at brass timbres. This would enable a more detailed understanding of these specific areas, and perhaps a situation could be arrived at whereby different types of timbres could each have their own multidimensional spaces defined by different numbers of dimensions and different parameters. Of course, real instrument tones comprise but a small subset of the unlimited number of possible timbres, and perhaps different classes of non-instrument tones could be proposed, each of which would also have its own multidimensional space. It is up to future researchers to explore all the avenues possible in search of adequate ways to explain human timbre perception."

Wow, this is a very interesting topic, nice.
The main problem with music and sound as a whole, is that it's all a matter of taste and opinions. I know a lot of 'audiophiles' via my school, who absolutely HATE Devin's productions because of the hugely compressed sound. Compression seems a big no-no in most musicians their minds. I can understand that people like the older, analogue productions where you can hear the tape-spill, and other noises in the background. It sounds a little bit warmer and 'cosier', but years went by, music changed, new genres were created, technology took over, etc. Home productions became very popular, and no one has the money for an analogue recording studio, so a lot of music is recorded digitally directly to harddrive.
I'm no audiophile, but I do think the sound of music is very important. It is absolutely true that there is a loudness war going on, and there are some records, which are unlistenable without getting a headache or getting annoyed. Unfortunately, there is even more stuff that you have to think about: people (and especially the youth) their attention span is about 30 seconds these days. How many people do you know who still listen to an entire album in one sitting? (Yes, I know YOU do, but you are Devin fan, so you rock.. ) Producers, labels, managers, promoters, ALL know these facts. Technology has brought us a whole new world of selling tracks and singles, instead of full records.
...and in the end, don't forget about money and competition. If one single on the radio sounds like a 'phat' wall of sound, and the next song sounds less in your face, the first song will be remembered.

I guess I kinda lost the topic, but anyway, this was just something which was on my mind while reading the article

"I can’t bear to play back some of my favourite records from the last few years through my hi-fi and pay them full attention, and this is upsetting."

I think that's the key point here for him complaining about mp3s and compressed music. I can stand to listen to 128kbps mp3s on my computer speakers because the speakers aren't good enough to show the flaws that come with being an mp3. But if I plug my headphones into the computer then I cringe every time the drummer hits a cymbal from the distortion on the treble... and this on 110$ Sennheiser headphones so I assume on a hifi system that costs thousands of pounds the effect is the same only much, much worse.

I suppose so, though I was referring to listening to said low-quality mp3s on a pair of Sony MDR-v6s. It's a very noticeable difference, but I have a few albums in 128 kbps that I listen to fairly often without any cringing. I suppose I can see how some couldn't stand it though.

Yeah, I guess calling it unbearable is going a little too far but I can see where he's coming from.

Some albums sound fine in 128kbps (lo-fi recordings, mainly older bands), and some don't sound good unless they're in 256kbps (modern recordings of today). It all rests on the production and mixing.

192 usually works for me. i can instantly tell when its 128 though, the sounds not as "deep".

It all rests on the production and mixing.

...And the standards of the listener as to what's acceptable as 'good.'

128 is the price I pay for downloading and not buying. Whatever I buy, I rip at 320. Whatever I download, I convert to 128.