Audiology Thread

[Glathannus]

I'd go for the DTS-HD.
Even though the home theater receiver in my living room doesn't support the losslessness of DTS-HD, the truncated 1536kbps which would carry through SPDIF is still really nice. There are DVD videos of not-necessarily-Japanese cinema where I bought the Japanese versions just to get 1536kps DTS (not often found in North American DVD releases anymore), because it's the best surround sound you can get with standard-definition video, and at least half of all Blu-Rays aren't bothering to surpass that.

Even if they offered 5.1 FLAC in the same complete (above 1536kbps) quality as DTS-HD, audio filetypes(s) in the Xiph family (of which FLAC is a member) have been confused about which 5.1 channels are which. Maybe it's just the Ogg container - it's been awhile since the last time I experimented with the stuff.

Don't ever pay specifically for AC3 if there are alternatives. It's synonymous with Dolby, and although they pioneered respectable advancements to analog audio, they cut too many corners on how the DVD format (and compatible equipment) was standardized, and then they started an official debate with DTS - which Dolby lost. They're this evil like MP3, which simply won't go away - because they're too deeply entrenched within our society. If the "inventor" of Dolby Digital Live were to be assassinated, I'd pitch in toward the legal fees of anyone who gets indicted for that, even if I were already near-broke (which I pretty much am sometime after February 15th).

[TurtleFu]

Thanks a lot guys!

@grandma
normally I would agree with you, but Imogen composes her music with surround sound specifically in mind. They sound good either way, but it's a treat in surround sound

It's actually a package, so MP3 is paired with AC3, FLAC with FLAC, and Apple Lossless with DTS-HD. So DTS-HD is the best but Foobar doesn't support Apple Lossless. You can't mix n match (buy the surround mixes individually)
Hm....

[Tokyo Jihad]

it can support it, google it!

[TurtleFu]

Well, I guess that solves my problem for me! Thanks again guys!

[Carlx]

Speaking of audiology, I remember reading that the lossy compression formats may cause hearing damage. I did some quick research and found that "Some listeners of MP3 music e.g. report to realize, that after longer listening of MP3 they start to perceive its typical sound flaws (so-called artifacts) also in unreduced musics, which suggests that MP3 maladjusts the hearing in a way that datareduced and unreduced sounds start to sound the same way wrongly while the hearing unlearns to recognize the difference".

Doesn't sound very convincing to me, it's not that we live hearing/listening to compressed data all day long. So, can anybody provide more reliable information? Just for the sake of curiosity.

[W3iHong]

That is very interesting. Someone please confirm it.

Plus, it is also a big difference in the kb/s in mp3 format. Carlx, do you know what kb/s the mp3 files were used in the study? And the volume it was played too if you do have the data.

[Glathannus]

To understand how/why lossy compression could be bad, first you have to understand what digital audio is. If analog audio is a solid line, then digital is a grouping of pixels attempting to mimic that solid line.

The samplerate of audio (such as 44.1kHz) is the digital resolution, which can't be rated like Width x Height. Lossy takes the dotted line, and replaces a bunch of small dots with a few larger dots (just like JPEG can do with sections of a picture) - effectively lowering the resolution of the audio - even if it's still rated at "44.1kHz". So what a lossy listener has going into their ear is even more pixelated than CD audio already is. CD audio without lossy compression can already be occasionally bad for your ears.

I could blast some lossless 96kHz/24bit really loud, and never get any listening fatigue from that - because with its "dots" being so highly concentrated, 96kHz/24bit more closely resembles the real sounds.

Not to mention that people who are okay with MP3s, are typically also okay with not-the-greatest headphones. With bad headphones (or bad speakers), the listener is subjected to frequency spikes or plateaus which less-closely-resemble the smoothness of the original sound. Compound that with how MP3s are already compromising the underlying sound, and you've got something seriously bad for your ears.

Now, it's not to say that lossy audio is automatically ear-cancer, because not all lossy is equally bad. It's just that MP3 has some fundamental design flaws you can't engineer your way out of, without changing to a different lossy format (such as Ogg Vorbis).

I've also heard about "maladjusted listeners" of MP3 (taking the coloration of MP3 sound for granted), but I'm not as confident about that as I am about some MP3s causing hearing loss which wouldn't have happened with lossless of the same recordings.

Before we had MP3 players, we had CD players. We didn't have so many CD listeners with bad hearing during the 1990s. Sure, every generation of listeners has a minority who play things louder than they should, but there wasn't as much negative consequence for doing so with CDs. With MP3s, it's not just the loud-blasting minority, who is losing their hearing.

As I've said before, I once downloaded the Cowboy Bebop CD-Box in MP3, and always got a little bit of listening fatigue (the precursor to hearing loss) if I sat through more than an hour of it at a time. When I finally bought the CDs, I thought the listening fatigue would still be there, because it must've been part of the recording/mixing. It was completely gone.

Did I experience listening fatigue with every MP3 I ever listened to? No, but a few Shiina Ringo MP3s were terrible, too. Kyoiku in MP3, is hearing loss waiting to happen - I wouldn't dream of blasting that. The CD on its own is already bad enough, but combining it with MP3 is like combining bad headphones with MP3.

[Carlx]

I understand and correspond to how bad lossy audio can be to your ears in the sense that it affects the listening experience, compromising the original timbre, smoothness, space between instruments and so on. But I still don't see how and why it may cause listening fatigue, specially if the lossy formats are theoretically supposed to omit the sound parts that the average human ear would not consciously perceive. So if the original recording wasn't already hearing damaging, then why the fact that it's compressed even more would turn it into something harmful. Following the sound/picture analogy, we'll be damaging our vision while occasionally seeing a low resolution JPG the proper way (no blinding light/no deafening loudness).

Is there a chance that your listening fatigue while listening to MP3 was exclusively a psychological response to an effort (un)consciously made to compensate for the data reduction? I mean, for example, trying to distinguish between instruments.

Originally Posted by W3iHong

Plus, it is also a big difference in the kb/s in mp3 format. Carlx, do you know what kb/s the mp3 files were used in the study? And the volume it was played too if you do have the data.

No, I don't. But I just found the whole essay here, that seems to shed light on this.

[Glathannus]

Listening fatigue doesn't have to be psychological in origin. Most of the listening fatigue anyone ever gets outside of super-loud concerts, comes from somewhere within the higher frequencies.

A good rule of thumb, is that if a cat is willing to stay and relax in your room while your stereo is blaring, then your music+equipment probably isn't fatigue-inducing. It's not that cats are hearing higher frequencies than we can hear (the recording only goes so high), but some subtle harshness which takes an hour to fatigue us, is going to be more immediately obvious to a cat.

Most studio monitors (professional speakers) are known to be fatiguing, if their higher frequencies aren't turned down. Sound mixing can be a conflict of interest, because you don't want to attack your ears, but you don't want to produce biased mixes either.

Low-detail MP3s might be unpleasant, but I don't automatically equate that with "fatiguing". Surprisingly, I haven't heard as many fatiguing 192kbps MP3s, as 320kbps MP3s. That could be because 192kbps is more likely to be first-generation lossy, while 320kbps is more likely to be a transcode (if it's something you've downloaded). It could be that transcoding has more to do with hearing loss, than any first generation of lossy. 320kbps could also be trying to preserve the highs (and ultimately ruin them), while 192kbps never tried. Not to say that 320kbps is automatically a bad thing, because I've heard plenty I never had a problem with at all.

There's so many different ways to encode an MP3 at the same bitrate, just as there are so many ways to record and mix an album at the same samplerate/bitdepth. The fatigue (or lack thereof) could ultimately boil down to how those things come together - instead of some albums or some encoding practices being universally bad.

[TeslaGuy]

I read the entire essay today and was not convinced by the author's argument that lossy audio can damage your hearing by causing your brain to fuck up the hair cells in your inner ear. There were some major flaws in his reasoning, and near the end of the paper he started to sound like a paranoid schizophrenic, claiming that his tv set was giving him tinnitus. And then after the essay's conclusion he identified himself as "CYBERYOGI Christian Oliver (=CO=) Windler (teachmaster of LOGOLOGIE - the first cyberage-religion!)". Enough said.

I found the discussion here about listening fatigue to be intriguing, although I did not agree with alot that was said. Instead of merely posting an alternative viewpoint, however, I decided to first see if I could find a legitimate definition online. The first couple of dozen hits were all contradictory "audiophile" looniness, but I finally came across a paper that was presented at the 2010 Audio Engineering Society convention. Its primary focus was on the "loudness wars", but it also discussed lossy encoding problems and listening fatigue. I've pasted a few excerpts here, but encourage all who are interested in any of the three topics to click on the link and download the PDF.

The Loudness War: Background, Speculation and Recommendations

Modern multiband compressors are carefully designed to sound as “transparent” as possible, with a minimum of audible side effects. However, there are many ways to go wrong, and the list of possible artifacts is surprisingly lengthy [24]. It is possible that some of the worst audible damage blamed on multiband compression may result primarily from equipment abuse, design flaws or artifacts such as clipping, intersample clipping [40], aliasing [41], excessive fast limiting or other forms of nonlinear distortion. Robert Dennis stated that “... two releases with an 8 dB crest factor can have remarkably different sound quality, varying between ‘not too bad’ and ‘horrible.’ A big factor behind this difference is the frequency spectrum of release. Distortion in CDs is often (or maybe primarily) odd-order harmonics caused by limiting (especially brick-wall limiting). Harmonic distortion of the bass spectrum tends to be masked by the midrange energy in the mix, something that cannot be said for harmonic distortion of the midrange and high frequency components of the spectrum.” [10]

In the digital domain, it is difficult to perform clipping without producing aliasing distortion [35]. The combination of clipping and low-bit-rate lossy encoding may be particularly unpleasant. At a recent AES workshop, Katz demonstrated the results of encoding a clipped recording using MP3 at 96 kbps (which is higher than the bit rates currently used on satellite radio). He subtracted the result from the original clipped recording and showed that the difference consisted of harsh high-frequency noise and distortion. [19] At a 2009 AES session, Marvin Caesar stated that lossy codecs respond badly to heavy compression and limiting. This effect may be especially problematic with stacked codecs in a typical FM signal chain. As a result, he claimed, “We’re foisting pretty rough audio on some of our listeners.” [42] In particular, clipping and aggressive dynamics processing may cause aliasing and intermodulation distortion when processed through lossy codecs. [35][43]

3.2. Listening Fatigue.

Even if people do not consciously notice any problems, it is possible that hypercompressed music may become physically or mentally tiring over time; listeners may gradually lose interest without knowing why. Marketing claims of “reduced listener fatigue” frequently appear in sales pitches for various audio products. In the audio engineering literature, listening fatigue has been attributed to a wide variety of causes including fast-acting compression and limiting [35], a lack of variation in loudness over the duration of a recording [35], contradictory location cues [44][45] [46], stereo image processing and active matrix upmix steering fluctuations [35][47], phantom image instability [45], phasiness [48], poor equalization [35], low frequency rumble [49], too much boost between 1-4 kHz [50], clipping and intermodulation distortion [35] [51], Doppler speaker distortion [52], low bit-rate lossy encoding [34][42], stacked codecs [42], excessive hearing aid amplification of soft background sounds [53], listening to monophonic instead of stereo recordings [54], listening to stereo instead of 3-channel recordings [46], and listening with headphones to audio mixed for speakers [55]. It appears, however, that the single most common use of the term “listener fatigue” in the engineering literature may be in regard to the need for frequent breaks during listening tests in order to avoid tiring the subjects and thereby corrupting the results (e.g., [22], [56] and [57]). This raises the not entirely facetious question of whether extended listening tests designed to measure fatigue might in fact corrupt their own results due to listening fatigue. Given the number of articles, blog postings and conference workshops on the topic, there is surprisingly little experimental evidence regarding hypercompression and listening fatigue. In one study, Stone et al reported that rapid multiband compression seemed to increase the amount of effort required to perceive independent sound sources within a complex signal, which may contribute to listening fatigue; however, this study was in the context of multiple simultaneous talkers, not music, and it did not measure fatigue as such [58].

3.2.1. Is Hypercompressed Audio More Fatiguing?

Again, there are a number of possibilities, not necessarily mutually exclusive: Hypercompression causes fatigue Despite the lack of published studies, it is widely believed that excessive compression results in listening fatigue, which may in turn discourage close, repeated listening [2]. There is a great deal of anecdotal evidence to this effect. Southall wrote, “Music is about tension and release. With very ‘hot,’ un-dynamic music there is no release because the sensory assault simply doesn’t let up.... you end up feeling like Alex at the end of A Clockwork Orange – battered, fatigued by, and disgusted with the music you love.... I very much doubt that this is just me.” [8] Mastering engineers are among those closest to the situation, with countless hours of detailed listening to varying amounts of compression. While some mastering engineers feel that the loudness war is merely a matter of finding an acceptable compromise between quality and market pressures, others are deeply concerned about listening fatigue and damaged sound quality. [12] Milner quotes mastering engineer Greg Calbi as saying that the idea that listener fatigue is a by-product of digital compression is “almost universally held.... I never heard the word ‘fatigue’ once when I was cutting vinyl.” [2] Mastering engineer Bob Weston stated that “highly compressed or limited music with no dynamic range is physically difficult to listen to for any period of time.... This ‘hearing fatigue’ doesn’t present itself as obviously aching muscles, like other forms of physical fatigue, so it’s not obvious to the listener that he or she is being affected. But if you ever wonder why you don’t like modern music as much as older recordings, or why you don’t like to listen to it for long periods of time (much less over the year), this physical and mental hearing fatigue is a big part of the reason.” [29] According to mastering engineer Joe Lambert, “Some records will work really loud; others wear on you. They sound great the first two times, but then you just stop listening because they fatigue your ear. I know that and the engineers know that, but the consumer doesn’t know why they’re not listening to those records any more.” [12] At an AES session on listener fatigue and longevity, Marvin Caesar stated that over tens of thousands of repetitions every year, when comparing competitively loud compressed recordings vs. hypercompressed songs, listeners almost always picked the less compressed one, saying “I could listen longer.” [42] Radio stations apparently adjust the amount of compression they use in an attempt to trade off a louder sound, which may initially attract listeners searching for a station, versus the danger of driving listeners away due to loudness fatigue. Stations looking for a female demographic, in particular, may avoid excessive processing because of a belief that women are particularly sensitive to overcompression. [2][14]


Types of Listening Fatigue

It is likely that the term “listening fatigue” may refer to two or more different phenomena. Johnston speculated that listener fatigue may include physical fatigue as well as central nervous system fatigue [59].

Physical fatigue

Physical fatigue, in turn, may consist of two types: mechanical cochlear fatigue and neural or biochemical cochlear fatigue. Keeping in mind that loudness (a perceptual phenomenon) and intensity (a physical phenomenon) are not perfectly correlated, Johnston suggested that mechanical cochlear fatigue may be due to high intensity, since outer hair cell damage appears to correlate to intensity. [59] High loudness, on the other hand, can cause outer hair cells to depolarize, suggesting the possibility of biochemical fatigue. In addition, the inner hair cell firing rate is largely proportional to loudness (not intensity), also suggesting biochemical fatigue. [59] Johnston mentioned that listening to something that sounds harsh, even if it’s not that loud, can make us want to run screaming in a few minutes; it’s not clear whether this is a physiological or intellectual process [59]. In a comparison of various loudspeakers, Berkow noticed that more distortion components can give speakers a fuller, richer sound, but they also cause fatigue and “wear the ears out sooner” [49].

Cognitive fatigue

Central nervous system fatigue, or “cognitive fatigue,” can result from missing, false or contradictory cues. If cues are incorrect or unavailable, the brain has to do more work to extract the information. With speech, in particular, multiband compression may flatten the signal and destroy the articulation. [59] Marvin Caesar discussed the importance of transient response: fast attack limiting crushes the leading edge of the signal, reducing intelligibility and forcing the listener to work harder to understand the content [42]. Sheffield found that digital artifacts such as echo, swirling and sibilance [61] from low bit-rate coders seemed to cause fatigue over time with speech, but not with music [34]; it is possible that dynamics compression artifacts could have a similar effect. It is unclear how cognitive fatigue might apply to music, since there may be less need to extract specific information [59]. However, to the extent that processed audio sounds “unnatural,” with a mismatch between what we hear and how things normally sound in the natural world, the brain may puzzle over various discrepancies and distractions. Thus, cognitive fatigue might result from such things as contradictory or incomplete location cues, rapidly changing frequency response [24] and various other artifacts. Slow (macrodynamic) compression causes a lack of tension and release variation, allowing the brain to regard the music as “background” and tune it out. At an AES workshop on the loudness war, Rodgers said, “When you reduce the dynamic changes, the listener can listen for a longer period of time – it’s less cognitively taxing. Attention doesn’t have to be woken up from these changes – you can tune it out.... Music becomes this background noise – it is the noise in your own personal signal-to-noise ratio. It’s not changing; it’s not engaging you cognitively.” [28] Thus, macrodynamic compression may actually reduce cognitive fatigue so long as the listener is not actively trying to attend to the information, though it may induce a related phenomenon, boredom, due to the lack of contrast and “negative space.” If the listener is trying to pay attention to the content, on the other hand, macrodynamic compression may cause increased fatigue, perhaps because the brain can have trouble efficiently decoding signals delivered at a constant level [62]. In “Over the Limit,” Rowan wrote: “WHY IS THE LOUDER IS BETTER APPROACH THE WRONG APPROACH? BECAUSE WHEN ALL OF THE SIGNAL IS AT THE MAXIMUM LEVEL, THEN THERE IS NO WAY FOR THE SIGNAL TO HAVE ANY PUNCH. THE WHOLE THING COMES SCREAMING AT YOU LIKE A MESSAGE IN ALL CAPITAL LETTERS. AS WE ALL KNOW, WHEN YOU TYPE IN ALL CAPITAL LETTERS THERE ARE NO CUES TO HELP THE BRAIN MAKE SENSE OF THE SIGNAL, AND THE MIND TIRES QUICKLY OF TRYING TO PROCESS WHAT IS, BASICALLY, WHITE NOISE. LIKEWISE, A SIGNAL THAT JUST PEGS THE METERS CAUSES THE BRAIN TO REACT AS THOUGH IT IS BEING FED WHITE NOISE. WE SIMPLY FILTER IT OUT AND QUIT TRYING TO PROCESS IT.” [63]