Audio is such a great subject; everyone has a different opinion! I may as well add mine, as it's a subject I've been paying some attention to lately. I would also consider myself to have pretty good hearing as well as musical appreciation. I've been involved with many classical music performances over the years, and many people I know consider my hearing to be quite accurate. I would not consider myself an "audiophile", however I do spend some time perusing forums such as diyaudio. Now, here's the point: although I expect to be ridiculed by some, I simply can't tell the difference between a properly encoded MP3 and the "real" thing. I've tried over and over with different tracks and bitrates. I've got a pretty nice M-Audio card in my system and some Sennheiser phones, and it's only when I listen to bit rates of around 160 or below that I can tell any difference. As bit rates decrease I first notice a loss of depth and liveliness, then a loss of detail, and finally the introduction of artifacts. Why am I telling you this? For the sake of a some perspective, as there is basically no limit on what you can spend on audio in the quest for "perfection". The reality of the situation is that you are at the mercy of your recordings. There is massive variation in the quality of my collection of around 200 CDs. When I first got my soundcard I realised just how badly produced some of these recordings are. Obviously we don't want to make matters worse by using crappy equipment and high compression, but let's be practical.
Because of your requirements, in particular the requirement for surround sound, I recommend the a multichannel digital receiver like the Panasonic previously mentioned (the Yamaha you're interested in seems to be similar). You'll get pretty good quality with no fuss. Because of the digital input you won't need to worry about messing around with soundcards, and there are less hassles when you upgrade your PC. Motherboards these days usually have digital out, and a single cable keeps things neat too. If on the other hand you were only interested in stereo, I would recommend a decent USB DAC and amplifier.
I like your idea of using two good quality speakers for music, then adding other speakers later; it's a great compromise. As far as bookshelf vs floorstanding is concerned, some bookshelf speakers can be quite good. For example I've got a pair of B&Ws and the bass is actually quite impressive for their size. On the other hand, all speakers need to placed appropriately for good performance, and the criteria include having them at around listening height and away from walls, so in many cases a bookshelf is not ideal. To get around this many people put their speakers on stands. I figure that you might as well get floor standing speakers for the better sound and ease of placement, as they are usually designed to be at listening height. They are of course larger and heavier, and this is something you will also need to think about if you are planning on moving around a lot. Practicality is a wonderful guide when making decisions.
Audio(phile?) / Hi-Fi Questions...
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
The listening tests I've seen, such as the various blind tests at the Hydrogen Audio forum, indicate that to the vast majority of people can't hear any difference between high quality MP3/Ogg and the uncompressed source. Maybe that's their ears or their equipment, but I doubt that everyone contributing is either half deaf or listening on $20 computer speakers.aef110 wrote:The difference between any MP3 (or LAME, or OGG, etc.) and the original is typically astronomical. If you're not hearing a difference, let alone a large one, it's exposing the limited resolution of your equipment and / or your setup.
I'm not saying that you aren't hearing it, just that to most people the differences are much more subtle than you indicate, and that the loss of quality is generally imperceptible. In my experience the difference in sound between multiple releases of the same CD can be vastly greater than the difference between MP3 (even at a fairly low bitrate) and the CD audio.
Having said all that, hard drive space is cheaper than ever; it's no longer that expensive to store 1000+ CDs in a lossless format. Rather than carrying out listening tests and worrying about whether a particular MP3 bitrate will be transparent, it makes sense to simply use FLAC and avoid any quality issues. Converting from FLAC to MP3 is quick and easy when required for portable use.
Actually the LAME developers have been looking for people who can hear encoding issues in high-bitrate files. If you've got some free time maybe you could send them some test results. I'm not being facetious; with the improvements to the encoder they're finding it hard to find anyone who can hear any loss of quality or audio artifacts, even on problem samples that are difficult to encode. People who can actually hear any improvements are very useful when trying to improve an encoder.aef110 wrote:I've done extensive testing of all audio compression formats with various source material, and even others listening blindly have commented that the difference is like night and day, even with the highest bit rates compressed using the highest quality settings. (On very old, mono recordings I was surprised to have heard no difference, though it's surely due to the limited resolution of the originals.)
The same can be said of any product; that's simply a fact of running a business. In the case of quality hi-fi, however, most of it is by smaller companies that aren't in the consumer spotlight and which don't have enormous advertising budgets -- as do the publicly traded companies that sell to the mass market. Hi-fi companies do, however, care about creating quality products at their target price points (which is once in a while cost-no-object). They sacrifice the bells and whistles that attract the mass market, in favor of better quality components and ultimately better sound.The Gangrel wrote:My problem with the "hi-fi industry" is that an £1000 amp will only have £150 worth of electronics inside it, you are paying for brand/lifestyle/boasting prowess etc. That's why I suggest buying as cheap as sounds good enough, because basically you're getting ripped off.
Regardless, we're not talking about balls-to-the-wall mortgage-your-house gear. We're simply talking about reasonably good-sounding entry-level hi-fi. We're well below the point of diminishing returns (which exists in audio just as it does in the PC world), and none of your dollars is going to prestige value here. Many of these components are nothing to look at (especially most NAD gear), and some are downright unattractive. An entry-level NAD amp with small but quality bookshelf speakers like I recommend isn't getting you any boasting power. It will get you a surprising amount of musical enjoyment for the price, though. I've heard excellent budget amps from Rotel, Creek and Cambridge Audio, as well as the older harman/kardon, and PSB makes surprisingly good speakers for very little money.
I've learned an enormous amount from all of you in these forums and couldn't have built my new quiet audio workstation without your considerable help. I learned along the way that if you want quality, for instance a very quiet and efficient power supply or a quiet case, you do need to pay for it. We're always working under real-world constraints, and it's just not possible to get quad-core and the fastest, hottest-running graphics in SLI mode and four hard drives in a micro ATX case and passively cool the thing, and it's certainly not possible to do so for only a few hundred bucks. Audio is no different in that respect, which I why I recommend not trying to get a full-fledged surround system for such a tiny outlay as 1000 euros. Audio gear also needs to have high-quality power supplies to deliver sufficient outlays of current as well as voltage to meet the demands of musical transients. All 100-watt amplifiers are by no means the same, from a purely electrical perspective, and they by no means sound the same, either. A quality 35-watt amp can thus easily outperform a mediocre 100-watt amp.
For this money it's better to stick with a good entry-level two-channel amp and bookshelf speakers. Good luck with your purchase and enjoy whatever you get.
a. Yes, given the low cost of storage FLAC makes sense. I've certainly heard quality losses on MP3s; they're most evident on acoustic recordings that start with a lot of ambient information; least evident on pop music.
b. Gangrel, aef110's the "listening to music is a skill that develops over time" point is entirely true; in fact I think I would say obviously true if you give it any thought. Even coming to grips with a new genre of music can change the way you hear. It's another reason for keeping your initial expenditures modest (and considering used gear) because the experience of listening to a good system for a year or so will change your tastes. Let's try to listen to different folks' contributions with respect.
c. And as should be clear to anyone who reads carefully, aef110 has been making precisely the point that Gangrel believes he's making, that audio is a mixed bag and that you can get good results by listening carefully and going for brands that are not flashy and lack large ad budgets. Creek is a good example. "Buying as cheap as sounds good enough" is *precisely* the point, and many of us have done that by buying used gear (50% off!), DIYing, and now moving into PC audio which has the potential to be a high-quality digital source for less than what we've spent in the past on CD players. A few summers ago a lot of us experimented with a $20 digital amp (the Sonic Impact -- quite good but low-powered). So while there are snob audiophiles and woo-woo audiophiles, there are also quite a few of us who are penny-pinching and empirically-minded, and who have also found that with a little patience we can get stunning sound.
For people who have DIY inclinations, incidentally, there's a whole world of possibility out there. Doing a little DIYing whill quickly disabuse you of the idea that an audio component is just a bunch of parts, but it's educational as hell and if you like building stuff, entertaining. Passdiy is one interestingly quirky site around one line of amplifiers, http://sound.westhost.com/ is a no-bullshit Australian site, http://www.t-linespeakers.org/ is loads of fun. diyaudio.com is a nice board.
b. Gangrel, aef110's the "listening to music is a skill that develops over time" point is entirely true; in fact I think I would say obviously true if you give it any thought. Even coming to grips with a new genre of music can change the way you hear. It's another reason for keeping your initial expenditures modest (and considering used gear) because the experience of listening to a good system for a year or so will change your tastes. Let's try to listen to different folks' contributions with respect.
c. And as should be clear to anyone who reads carefully, aef110 has been making precisely the point that Gangrel believes he's making, that audio is a mixed bag and that you can get good results by listening carefully and going for brands that are not flashy and lack large ad budgets. Creek is a good example. "Buying as cheap as sounds good enough" is *precisely* the point, and many of us have done that by buying used gear (50% off!), DIYing, and now moving into PC audio which has the potential to be a high-quality digital source for less than what we've spent in the past on CD players. A few summers ago a lot of us experimented with a $20 digital amp (the Sonic Impact -- quite good but low-powered). So while there are snob audiophiles and woo-woo audiophiles, there are also quite a few of us who are penny-pinching and empirically-minded, and who have also found that with a little patience we can get stunning sound.
For people who have DIY inclinations, incidentally, there's a whole world of possibility out there. Doing a little DIYing whill quickly disabuse you of the idea that an audio component is just a bunch of parts, but it's educational as hell and if you like building stuff, entertaining. Passdiy is one interestingly quirky site around one line of amplifiers, http://sound.westhost.com/ is a no-bullshit Australian site, http://www.t-linespeakers.org/ is loads of fun. diyaudio.com is a nice board.
Can you give specific examples of where you and your friends heard this "night and day" difference (in blind testing)? If the difference is as "typically astronomical" as you claim, you should have plenty to choose from, and I'd like to post the clips on the HA forums for further analysis.aef110 wrote: The difference between any MP3 (or LAME, or OGG, etc.) and the original is typically astronomical. If you're not hearing a difference, let alone a large one, it's exposing the limited resolution of your equipment and / or your setup. I've done extensive testing of all audio compression formats with various source material, and even others listening blindly have commented that the difference is like night and day, even with the highest bit rates compressed using the highest quality settings.
Probably placebo 
That said, the differences are most easily spotted on "noise", such as crowd applause and cymbals. If you know what to look for, you'll hear even the slightest occurance.
I remember encoding a Cranberries album in 1998 or 1999... Back then, there were only two real choices: Fraunhofer and the horrible Xing. Surprisingly, Xing handled the Irish very well even on 128 kbps (and by "very well", I mean "listenable"). Fraunhofer, on the other hand, was a flanging artefactor of doom, even on 192 kbps! Encoders have improved since, so they'd probably produce superior results to those two nowadays.
That said, the differences are most easily spotted on "noise", such as crowd applause and cymbals. If you know what to look for, you'll hear even the slightest occurance.
I remember encoding a Cranberries album in 1998 or 1999... Back then, there were only two real choices: Fraunhofer and the horrible Xing. Surprisingly, Xing handled the Irish very well even on 128 kbps (and by "very well", I mean "listenable"). Fraunhofer, on the other hand, was a flanging artefactor of doom, even on 192 kbps! Encoders have improved since, so they'd probably produce superior results to those two nowadays.
I just dug up several test CDs I made a few years ago of classical music and jazz. It's been some time and encoding has improved since then; further, I lost the track list and don't know which encoding is which. I'm pretty impressed with my (very) cursory listening right now, at least with the very high bit-rate samples. Where there are differences, they're admittedly subtleties such as soundstage, imaging, three-dimensionality, the timbre of a violin or female voice, etc.nick705 wrote:Can you give specific examples of where you and your friends heard this "night and day" difference (in blind testing)? If the difference is as "typically astronomical" as you claim, you should have plenty to choose from, and I'd like to post the clips on the HA forums for further analysis.
I compressed a couple files more recently, but they were for projects that haven't yet been released commercially and I don't have the authority to post them online. One soprano immediately identified a 192 kB/s MP3 (LAME-encoded using the highest quality settings and the newest LAME codec) as lacking the presence of the original. The other was originally a DVD-Audio disc recorded at 24 bits and 192 kHz -- there's an obvious and large loss of quality converting to CD (16-bit, 44.1 kHz) and certainly compressing it further. Keep in mind I'm dealing with professional classical musicians and we're trained (hopefully!) to listen extremely carefully for the tiniest of subtleties. That doesn't mean we're all audiophiles -- the opposite is often true, in fact.
Re: blind testing, that's something I'm not in favor of. Blind testing would seem to be the ideal method for "proving" something one way or the other, but it's fatiguing and hence confusing and it doesn't reflect the way we actually listen to music, which is indeed subjective. I have no doubt that the placebo effect plays a role in hi-fi, and I won't claim to be an exception. That doesn't mean, however, that there aren't often very large differences. I've been shocked by the differences between even very expensive CD players, for instance, or cables, for that matter (although that's an area where profit margins are disproportionately large and you really are often paying for marketing).
The difficulty is that the better your equipment, the more it reveals differences elsewhere in the chain. It's like a "silent" PC in this regard: One person, whose components are louder, might hear no difference after switching to a quieter CPU fan, for instance, whereas someone else, with a quieter system in a quieter room might hear a large difference. It also depends on our hearing -- we all hear differently, after all, and by definition I can't be sure that what I'm hearing is what you're hearing any more than I can be certain that the blue I see is the exact same blue you see.
In the end, I'm definitely impressed with high bit-rate encoding -- enough to agree that my "astronomical" comment earlier is perhaps presumptuous -- but the differences that exist depend on the source material as well as on the playback equipment and its setup. (I'm sure if I upgraded to some of the ultra-expensive gear I've heard, such as the incredible JM Lab Utopia speakers at ca. $80,000 a pair (!), I'd never be able to listen to compressed audio again.) There are recordings on which I've simply heard no difference even if I tried to convince myself I must have! I second the suggestion to use lossless compression such as FLAC since storage space is cheap nowadays. That way we can be sure that we're not throwing away musical information.
I recommend resources such as Stereophile (www.stereophile.com) for further info. This has been an interesting discussion and I've learned from everyone's input.
A ground current can interfere with your audio signal only once it has been converted to analog, not when digital! If you want to test if this is a problem in your case, first fully disconnect your pc from your audio rig, turn on the volume to max and, using any electrical wire at your disposal, connect the chassis of your amplifier to the chassis of your PC. If the background noise does not increase then your have no ground loop problems. If your motherboard has an optical out then it should be fine (I don't think they all re-sample), although one could worry about potential clock stability issues if it is built too cheaply.autoboy wrote:I'm not sure I follow this. There is only one reason to use a separate soundcard with toslink when you motherboard supports toslink also, and that is that some soundcards are able to bypass windows audio and do not resample the 44khz signal to 48khz. This is a pretty rare thing and i've not seen too many do it. The chaintek is one of those boards that do it but the software setup is more involved than standard toslink.fgp wrote:One more point, I recommand to use a cheap sound card with optical out (not coax) such as the chaintech AV710 to connect your PC to your audio system. This breaks a ground loop with the amplifier (this eliminated a small but annoying hum for me) and external DACs are usually better (cleaner power...).
As for the ground loop issue, i'm not sure how a groundloop problem would cause hum with a digital signal but moving to a optical cable would eliminate ground loop problems, though I have never heard of it with a toslink. Optical cables IMO are less robust, have bad connections, and are prone to failure, while the copper cable is robust, effective, and free because you always have some spare video cables lying around. They use the same specs (75ohm) so they can be used interchangably
Concerning MP3 encoding. Honestly when encoded with Lame and adequate settings (say variable bit rate -V0 ~200kbs) I essentially can not tell the difference with the corresponding FLAC files. Despite that I go safe and use FLAC since storage is so cheap. In particular this is future proof as one can re-encode in an other format without loss of information. I don't want to rip once again my entire collection, no way!
So no specific examples then?aef110 wrote: I just dug up several test CDs I made a few years ago of classical music and jazz. It's been some time and encoding has improved since then; further, I lost the track list and don't know which encoding is which. I'm pretty impressed with my (very) cursory listening right now, at least with the very high bit-rate samples. Where there are differences, they're admittedly subtleties such as soundstage, imaging, three-dimensionality, the timbre of a violin or female voice, etc.
I compressed a couple files more recently, but they were for projects that haven't yet been released commercially and I don't have the authority to post them online. One soprano immediately identified a 192 kB/s MP3 (LAME-encoded using the highest quality settings and the newest LAME codec) as lacking the presence of the original. The other was originally a DVD-Audio disc recorded at 24 bits and 192 kHz -- there's an obvious and large loss of quality converting to CD (16-bit, 44.1 kHz) and certainly compressing it further. Keep in mind I'm dealing with professional classical musicians and we're trained (hopefully!) to listen extremely carefully for the tiniest of subtleties. That doesn't mean we're all audiophiles -- the opposite is often true, in fact.
Re: blind testing, that's something I'm not in favor of. Blind testing would seem to be the ideal method for "proving" something one way or the other, but it's fatiguing and hence confusing and it doesn't reflect the way we actually listen to music, which is indeed subjective. I have no doubt that the placebo effect plays a role in hi-fi, and I won't claim to be an exception. That doesn't mean, however, that there aren't often very large differences. I've been shocked by the differences between even very expensive CD players, for instance, or cables, for that matter (although that's an area where profit margins are disproportionately large and you really are often paying for marketing).
The difficulty is that the better your equipment, the more it reveals differences elsewhere in the chain. It's like a "silent" PC in this regard: One person, whose components are louder, might hear no difference after switching to a quieter CPU fan, for instance, whereas someone else, with a quieter system in a quieter room might hear a large difference. It also depends on our hearing -- we all hear differently, after all, and by definition I can't be sure that what I'm hearing is what you're hearing any more than I can be certain that the blue I see is the exact same blue you see.
In the end, I'm definitely impressed with high bit-rate encoding -- enough to agree that my "astronomical" comment earlier is perhaps presumptuous -- but the differences that exist depend on the source material as well as on the playback equipment and its setup. (I'm sure if I upgraded to some of the ultra-expensive gear I've heard, such as the incredible JM Lab Utopia speakers at ca. $80,000 a pair (!), I'd never be able to listen to compressed audio again.) There are recordings on which I've simply heard no difference even if I tried to convince myself I must have! I second the suggestion to use lossless compression such as FLAC since storage space is cheap nowadays. That way we can be sure that we're not throwing away musical information.
I recommend resources such as Stereophile (www.stereophile.com) for further info. This has been an interesting discussion and I've learned from everyone's input.
-
NeilBlanchard
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
Hello,
I have participated in "tests" comparing 44.1mHz/16bit PCM ("CD" quality) with 96mHz/24bit PCM and the differences were quite obvious to me, and to just about everybody else in the room (a meeting of the Boston Audio Society). I have made 320kb MP3's from both CD's and LP's that I own, and there are obvious losses in quality.
I have made 44.1/16 and 96/32 float transfers of the same recording (7 1/2 IPS stereo analog from my Tandberg TD-20A SE), and the quality differences between those digital transfers is glaringly obvious. YMMV, of course.
I have participated in "tests" comparing 44.1mHz/16bit PCM ("CD" quality) with 96mHz/24bit PCM and the differences were quite obvious to me, and to just about everybody else in the room (a meeting of the Boston Audio Society). I have made 320kb MP3's from both CD's and LP's that I own, and there are obvious losses in quality.
I have made 44.1/16 and 96/32 float transfers of the same recording (7 1/2 IPS stereo analog from my Tandberg TD-20A SE), and the quality differences between those digital transfers is glaringly obvious. YMMV, of course.
-
TheAtomicKid
- Posts: 95
- Joined: Sat Aug 20, 2005 3:01 am
Bland, unqualified statements like "obvious losses in quality" aren't very helpful.NeilBlanchard wrote:I have made 320kb MP3's from both CD's and LP's that I own, and there are obvious losses in quality.
What encoder were you using? LAME, FhG, Xing, Helix, Blade? What command-line arguments? What exactly is the nature of the "obvious" quality loss? What piece of music? What CD/catalog number/track? At what time location in the given track is the difference apparent?
If the differences are "obvious," a specific example should be very easy to find, and I'm sure the codec developers would appreciate it.
I don't really understand this. Surely the only significant difference between a blind and a non-blind test is whether you know what you're listening to? I don't see why that would add any extra fatigue and confusion to the process. This seems more like an argument against any kind of testing, but without that how do you decide what audio hardware/software to use?aef110 wrote: Re: blind testing, that's something I'm not in favor of. Blind testing would seem to be the ideal method for "proving" something one way or the other, but it's fatiguing and hence confusing and it doesn't reflect the way we actually listen to music, which is indeed subjective. I have no doubt that the placebo effect plays a role in hi-fi, and I won't claim to be an exception. That doesn't mean, however, that there aren't often very large differences.
Test conditions may not be the same as normal listening, but I don't see why that would make a huge difference to the sound you hear. Maybe it could lead you to miss subtleties, but if there's a large difference, 'like night and day' as you put it, then I can't see it being totally eliminated by the testing process.
I used to read Stereophile, but after the snake oil products they've promoted over the years, and the inconsistent and highly subjective reviews, I find it hard to take them seriously as a good source of information.aef110 wrote: I recommend resources such as Stereophile (www.stereophile.com) for further info. This has been an interesting discussion and I've learned from everyone's input.
How do you decide which music to listen to?Steve_Y wrote:I don't really understand this. Surely the only significant difference between a blind and a non-blind test is whether you know what you're listening to? I don't see why that would add any extra fatigue and confusion to the process. This seems more like an argument against any kind of testing, but without that how do you decide what audio hardware/software to use?
Blind testing isn't the object here. Blind testing is fatiguing in that its purpose is to identify which component is which rather than to describe (and enjoy) the sound you hear. It's like sipping wines ad infinitum until you can no longer taste them. It's best instead to listen extensively to the components in your price range, paying attention to component interaction, and settle on those that give you the most satisfaction with the music you play. After all, you're going to be listening to the thing all those hours.
For that matter, how do you decide which car to drive? Hopefully not through blind testing!
-
NeilBlanchard
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
Greetings,
How does one listen to music? If you get a shiver in your spine, or something takes your breath away, or a chord change makes you relax into a pool of tranquility -- how do you quantify that?!
If you hear recordings of different performances of the same music -- how do you "know" that they were different? How do you recognize your mother's voice on the telephone?
Our brains are amazingly good at picking apart sound, and using very subtle differences to differentiate between them. You get into the flow of the music, and the timing changes are quite obvious, between different performances.
I have participated in ABX tests (at the BAS) and they are horribly flawed. Someone else (another tester) was flipping the switches at will, and just keeping track of whether we were listening to A or B was frustratingly difficult. And if you spend all/most of your time keeping track of which you are listening to -- it hardly makes it possible to listen to the music in a state equivalent to listening to the music as you would for the pleasure of listening to it.
My opinions are my opinions, and if you trust me, then trust my judgments.
How does one listen to music? If you get a shiver in your spine, or something takes your breath away, or a chord change makes you relax into a pool of tranquility -- how do you quantify that?!
If you hear recordings of different performances of the same music -- how do you "know" that they were different? How do you recognize your mother's voice on the telephone?
Our brains are amazingly good at picking apart sound, and using very subtle differences to differentiate between them. You get into the flow of the music, and the timing changes are quite obvious, between different performances.
I have participated in ABX tests (at the BAS) and they are horribly flawed. Someone else (another tester) was flipping the switches at will, and just keeping track of whether we were listening to A or B was frustratingly difficult. And if you spend all/most of your time keeping track of which you are listening to -- it hardly makes it possible to listen to the music in a state equivalent to listening to the music as you would for the pleasure of listening to it.
My opinions are my opinions, and if you trust me, then trust my judgments.
-
The Gangrel
- Posts: 87
- Joined: Sun Apr 29, 2007 1:20 am
Any passage where you claim to have heard differences, variously described as "night and day" "astronomical" and "subtle" will do.aef110 wrote: How do you decide which music to listen to?
I'm not denying any of that - I'm trying to identify specific examples of where a well-encoded MP3 is clearly (or even slightly) distinguishable from the source material.NeilBlanchard wrote: How does one listen to music? If you get a shiver in your spine, or something takes your breath away, or a chord change makes you relax into a pool of tranquility -- how do you quantify that?!
If you hear recordings of different performances of the same music -- how do you "know" that they were different? How do you recognize your mother's voice on the telephone?
Our brains are amazingly good at picking apart sound, and using very subtle differences to differentiate between them. You get into the flow of the music, and the timing changes are quite obvious, between different performances.
The whole subjective nature of listening to music is precisely the problem, and precisely what ABX testing is designed to counteract, so that an objective claim of MP3 encoding causing audible degradation can be independently repeated and verified.
Probably, but in all honesty I doubt if it will go anywhere useful.The Gangrel wrote: I thought I'd be the last to say this, but maybe we've hijacked the OP's thread enough.
Time to split?
Sure. Listening is inherently subjective, as are all our five senses. It's important to understand that there's a difference between our subjective impression of sound and our ability to articulate it. We might be able to come up with a handful of adjectives for why we're attracted to a particular woman (she's beautiful, charming, intelligent), but at the end of the day there will just be "something about her" that attracts us and that we simply can't consciously identify.nick705 wrote:The whole subjective nature of listening to music is precisely the problem, and precisely what ABX testing is designed to counteract, so that an objective claim of MP3 encoding causing audible degradation can be independently repeated and verified.
I think listening to music is like that. I recommend going to the top hi-fi dealer in your area, bringing CDs you know and like, and just listening to really good gear. Don't try to analyze everything at the outset, just enjoy the music. While many differences will be obvious, it's only after extensive experience on high-quality equipment that you'll be able to perceive the more subtle differences, such as sharp sibilants in female vocals, a tubby midbass, phase distortions, scratchy violins, or a silky baritone, smooth top, tight bass (I almost wrote "tight ass"...). These differences will likely become more important to you over time.
Only after your ears have been opened up through extensive experience is it wise to experiment with SACD vs. CD or CD vs. MP3 and such, although you can certainly do so anytime. After all, a newcomer to classical music might not notice much if any differences between two different performances of a Mozart sonata, whereas professionals might become infuriated over one and ecstatic over the other.
In any case, happy listening!
That's hardly the same as trying to determine differences in sound quality. It's like the difference between deciding which computer game is more fun to play, and determining which graphics card gives you a higher frame rate and visual quality. While the former is purely subjective, the latter can also be tested and compared with other products in a more scientific way.aef110 wrote:How do you decide which music to listen to?
The differences between two pieces of music aren't subtle, they can't be compared like compression formats or pieces of audio hardware. On the other hand, if I was comparing two different releases of a CD, then I definitely would do a blind test to ensure that my expectations aren't getting in the way.
That's usually the first stage; if there's no detectable difference between two files or two components then there's not much point in going any further. Once someone can determine that there actually is a difference they can move on, describe that difference in sound, and comment on whether it's positive or negative.aef110 wrote:Blind testing isn't the object here. Blind testing is fatiguing in that its purpose is to identify which component is which rather than to describe (and enjoy) the sound you hear.
Surely that depends on exactly how the blind testing is set up? There's no reason why a blind test couldn't be run in a way that's pretty similar to normal listening. Rather than rapidly swapping between short samples as in a typical ABX test, you could still have long and careful listening sessions. All that's required is that you don't know what you're listening to, so that the placebo effect is eliminated.aef110 wrote:It's like sipping wines ad infinitum until you can no longer taste them.
Personally I find it very difficult to believe that a genuine and easily noticeable difference in sound could be eliminated by that testing.
When looking for a complete audio system it can be hard to test components in any other way. Setting up a blind test of a variety of different component combinations probably isn't a very practical option. Personally I think that's very unfortunate; it's probably one of the reasons why there's such a high quantity of rip off products and snake oil nonsense in the audio market.aef110 wrote:It's best instead to listen extensively to the components in your price range, paying attention to component interaction, and settle on those that give you the most satisfaction with the music you play. After all, you're going to be listening to the thing all those hours.
You don't generally seem to see the same psuedo-science junk in the computer market. No $1000 PSU cables, $100 'quantum' marker pens to colour the edge of your data CDs, or $230 magic stones claiming to speed up your PC. I guess the benefits of such products would be too easy to test, people are used to benchmarking their computers and would expect to see a measurable improvement. It's much harder to expose overpriced rubbish when you can only carry out subjective listening tests.
Thankfully blind testing of different compression formats on a computer is much easier to carry out. There's even software that'll quickly set up a configurable test using the files you select. They'll let you listen to as much of the music as you like, for as long as you like, and simply hide the file details until the test is over. In my opinion that's much more reliable and useful than a listening test where you know exactly what's playing.
I suppose ultimately it is the satisfaction you get from your system that's the most important thing; even if that satisfaction comes from placebo effect rather than real audio improvements. However, it's something I'd like to avoid as much as possible, especially when spending my money on components, or spending a lot of time encoding files.
Steve,
Your comments are immanently reasonable and I personally agree it's wise to regard some of the more esoteric products with suspicion, although not all products are snake oil and there's nothing to preclude some of them from working as claimed, even some of the more unusual ones. There have been cases where people wrote off, a priori, the possibility of something making an audible difference, only to be later countered by scientific evidence that proves the difference. Let's not forget the psychologist and philosopher William James' famous proclamation: "Belief creates the actual fact."
As for listening, it's worth commenting that two things are necessary: You need to have high-quality equipment that can reveal differences and you need to have highly developed ears in order to register those differences.
Again, listening is inherently subjective, and blind testing is not the way we should shop for a music system, even if it were practical to set up such tests. I've even found components that are audibly, and measurably, more accurate but that I like less than more "colored" gear. Therefore, it's not always about pure accuracy and delineating differences: You might like some of those differences. If this weren't the case, we would all just measure equipment for accuracy and buy accordingly, without having to listen to it. Some manufacturers have tried this approach with disastrous musical results, although they still impress the uninformed with meaningless specs.
As regards compression, why not switch to FLAC or other lossless formats as has been suggested here? Storage is cheap nowadays, plus you'll spare yourself the definite signal degradation that comes from compressing twice, which all compression inventors insist is a no-no. We don't know what compression schemes will be available in the future or what our portable players of tomorrow will read. Surely it's best to keep a lossless scheme on hand and to listen to the originals wherever possible.
Your comments are immanently reasonable and I personally agree it's wise to regard some of the more esoteric products with suspicion, although not all products are snake oil and there's nothing to preclude some of them from working as claimed, even some of the more unusual ones. There have been cases where people wrote off, a priori, the possibility of something making an audible difference, only to be later countered by scientific evidence that proves the difference. Let's not forget the psychologist and philosopher William James' famous proclamation: "Belief creates the actual fact."
As for listening, it's worth commenting that two things are necessary: You need to have high-quality equipment that can reveal differences and you need to have highly developed ears in order to register those differences.
Again, listening is inherently subjective, and blind testing is not the way we should shop for a music system, even if it were practical to set up such tests. I've even found components that are audibly, and measurably, more accurate but that I like less than more "colored" gear. Therefore, it's not always about pure accuracy and delineating differences: You might like some of those differences. If this weren't the case, we would all just measure equipment for accuracy and buy accordingly, without having to listen to it. Some manufacturers have tried this approach with disastrous musical results, although they still impress the uninformed with meaningless specs.
As regards compression, why not switch to FLAC or other lossless formats as has been suggested here? Storage is cheap nowadays, plus you'll spare yourself the definite signal degradation that comes from compressing twice, which all compression inventors insist is a no-no. We don't know what compression schemes will be available in the future or what our portable players of tomorrow will read. Surely it's best to keep a lossless scheme on hand and to listen to the originals wherever possible.
Can you not give a single specific example, in the entire immeasurable catalog of commercially released music, where you personally have heard these differences (with your high quality equipment and highly developed ears)?aef110 wrote: As for listening, it's worth commenting that two things are necessary: You need to have high-quality equipment that can reveal differences and you need to have highly developed ears in order to register those differences.
Let's not forget the farmer and writer Thomas Tusser's famous proclamation: "A fool and his money are soon parted."aef110 wrote: Let's not forget the psychologist and philosopher William James' famous proclamation: "Belief creates the actual fact."
That at least we can agree on.aef110 wrote:As regards compression, why not switch to FLAC or other lossless formats as has been suggested here? Storage is cheap nowadays, plus you'll spare yourself the definite signal degradation that comes from compressing twice, which all compression inventors insist is a no-no. We don't know what compression schemes will be available in the future or what our portable players of tomorrow will read. Surely it's best to keep a lossless scheme on hand and to listen to the originals wherever possible.
I've done my best to express myself with a respectful tone and would appreciate the same courtesy. Mere differences of opinion about audio equipment are hardly justification for disrespecting one another. This will be my last post on the subject.nick705 wrote:Can you not give a single specific example, in the entire immeasurable catalog of commercially released music, where you personally have heard these differences (with your high quality equipment and highly developed ears)?
I have heard differences in the audiophile label Telarc's recordings of the Beethoven String Quartets with the Cleveland Quartet, in Decca's release of Cecilia Bartoli singing Mozart arias, in Hilary Hahn's performances of the Bach Sonatas and Partitas for Violin, and in some Herbie Hancock whose title I don't recall. I have also heard differences in numerous private or other recordings not yet released commercially. I have not heard differences in historical classical recordings (1930s and earlier), presumably because the original medium contained limited information due to the technology of the time.
wow... that's not the reaction I was expecting, but I intended no disrespect and I apologise unreservedly.aef110 wrote: I've done my best to express myself with a respectful tone and would appreciate the same courtesy. Mere differences of opinion about audio equipment are hardly justification for disrespecting one another. This will be my last post on the subject.
I guess if you've withdrawn from the thread there's no point asking, but what differences are apparent in those recordings, to those with "highly developed ears" (your words, not mine)? At what points are the artefacts audible (CD/track number/timestamp would be helpful), and what form do they take?
If you don't want to contribute any more then fair enough, but I'm genuinely interested in what I'm apparently missing.
I would have to say there's at least as much snake oil and rip-offs in the PC market; 1kw PSUs, case fans that supposedly blow 70CFM at 8dBA, etc etc etc. These are the kinds of scams we are very familiar with on SPCR.You don't generally seem to see the same psuedo-science junk in the computer market. No $1000 PSU cables, $100 'quantum' marker pens to colour the edge of your data CDs, or $230 magic stones claiming to speed up your PC.
I think the difference is though, any such claims are easily debunked. It's harder with "audiophile" baubles and trinkets, when the vendor will simply turn round and say there must be something wrong with your hearing or equipment if you can't hear the difference, and they reject blind testing out of hand.jaganath wrote: I would have to say there's at least as much snake oil and rip-offs in the PC market; 1kw PSUs, case fans that supposedly blow 70CFM at 8dBA, etc etc etc. These are the kinds of scams we are very familiar with on SPCR.
The audio marketplace contains some unbelievable woo and there are audiophiles every bit as reality-averse as any new ager or follower of the occult. Google "machina dynamica" or "peter belt" for examples. So far the PC market has little of this (though for all I know there's someone out there selling magic oil to anoint your hard drive and chase away viruses). It's also true that a typical defense of the woo-audiophiles is hermetic relativism: I heard it, it's my reality.
One of the things I like about this forum is that it's empirical; people post data and talk about the physical mechanisms that produce it. We come here for measurable results.
So back to audio. First, a lot of important and interesting stuff *is* measured by instruments. What I recommend in _Stereophile_ are John Atkinson's "measurements" sections at the end of most reviews, which apply a consistent methodology and have some smart discussion. There's a pretty good archive online. This gives you the same kind of useful sense that this forum can provide of technical and engineering challenges, tradeoffs and bottlenecks.
What is so far not measured with instruments is the kind of test I use when I buy gear: in complex musical passages, can I hear each instrument distinctly? In a capella gospel music, do all the voices sound like real, breathing, human voices? Does a piano sound like a real piano or does it just remind me of a piano? I have no doubt that with good enough microphones and good enough signal-processing and whatnot, you could get instruments to measure these things and give you numbers, but we don't have them. (These are the same kinds of tests that will show up the limits of lossy audio compression.)
Wine is a fair analogy. In principle it's possible to measure everything in wine, and entrepreneurial chemists are doing more of that in recent years, but at this point we rely on human tasting for making most distinctions. (We're also used to the idea that tasting is an ability that you develop.)
So Nick I appreciate and share your empiricism. I just want to resist putting all differences that are not currently well-measured in the same category as pseudoscience and woo.
And yeah, apologies for thread highjacking, but I did make an effort earlier to figure out what differences mattered to the original poster and suggest how to spend money only on things he was gonna notice. PSB is a nice example of a sober, value-for-money, Canadian audio company. They have good engineering -- they're not making the kind of boom-and-tizz speakers that sound impressive on the floor at Best Buy but have no decent midrange, they make a nice array of modestly-priced gear, and they're great rock and roll speakers.
One of the things I like about this forum is that it's empirical; people post data and talk about the physical mechanisms that produce it. We come here for measurable results.
So back to audio. First, a lot of important and interesting stuff *is* measured by instruments. What I recommend in _Stereophile_ are John Atkinson's "measurements" sections at the end of most reviews, which apply a consistent methodology and have some smart discussion. There's a pretty good archive online. This gives you the same kind of useful sense that this forum can provide of technical and engineering challenges, tradeoffs and bottlenecks.
What is so far not measured with instruments is the kind of test I use when I buy gear: in complex musical passages, can I hear each instrument distinctly? In a capella gospel music, do all the voices sound like real, breathing, human voices? Does a piano sound like a real piano or does it just remind me of a piano? I have no doubt that with good enough microphones and good enough signal-processing and whatnot, you could get instruments to measure these things and give you numbers, but we don't have them. (These are the same kinds of tests that will show up the limits of lossy audio compression.)
Wine is a fair analogy. In principle it's possible to measure everything in wine, and entrepreneurial chemists are doing more of that in recent years, but at this point we rely on human tasting for making most distinctions. (We're also used to the idea that tasting is an ability that you develop.)
So Nick I appreciate and share your empiricism. I just want to resist putting all differences that are not currently well-measured in the same category as pseudoscience and woo.
And yeah, apologies for thread highjacking, but I did make an effort earlier to figure out what differences mattered to the original poster and suggest how to spend money only on things he was gonna notice. PSB is a nice example of a sober, value-for-money, Canadian audio company. They have good engineering -- they're not making the kind of boom-and-tizz speakers that sound impressive on the floor at Best Buy but have no decent midrange, they make a nice array of modestly-priced gear, and they're great rock and roll speakers.
colin2, I wouldn't argue with any of that, and if anything it supports the concept of ABX testing.colin2 wrote:What is so far not measured with instruments is the kind of test I use when I buy gear: in complex musical passages, can I hear each instrument distinctly? In a capella gospel music, do all the voices sound like real, breathing, human voices? Does a piano sound like a real piano or does it just remind me of a piano? I have no doubt that with good enough microphones and good enough signal-processing and whatnot, you could get instruments to measure these things and give you numbers, but we don't have them. (These are the same kinds of tests that will show up the limits of lossy audio compression.)
Properly conducted ABXing relies *entirely* on human perceptions, not scientific instruments - the objective is simply to remove (as far as possible) any suggestion of placebo. Sure, you can call it "measuring" if you like, but not in the sense of, say, comparing an MP3 file to the PCM source in a wav editor and "measuring" the visible difference (which would completely miss the point of a "perceptual" lossy encoder).
It's not a case of ruthlessly trying to reduce someone's musical experience to bits of scientific data - it's the simple and singular purpose of establishing whether a claimed difference is genuinely audible, or if it arises from the powerful postprocessor located between the ears (which I fully appreciate may be every bit as "real" to the individual concerned).
Very much so, and there's usually a sort of underlying suggestion that subjecting *any* kind of musical experience to objective analysis is somehow inherently invalid, or at least in poor taste, but if we're not going to retreat into mysticism/religion (in which case there's logically nowhere to go) I don't see an alternative.colin2 wrote:It's also true that a typical defense of the woo-audiophiles is hermetic relativism: I heard it, it's my reality.
Mine too, it just rattles my cage when people make sweeping statements like "MP3 is obviously inferior" and decline to provide any data to back up their assertion. If I were to make a post in the storage forum to the effect that "suspending hard drives will cause them to die early" without any evidence to support my claim, I'd be jumped on from a great height and justifiably so.colin2 wrote:And yeah, apologies for thread highjacking
Anyway, enough of this tomfoolery, the pub beckons...
I agree that those products shouldn't be dismissed out of hand. The problem is how to tell the difference between the unusual and the utterly non-functional. Personally I think blind testing is the best way to do that, maybe even the only reliable way.aef110 wrote:Your comments are immanently reasonable and I personally agree it's wise to regard some of the more esoteric products with suspicion, although not all products are snake oil and there's nothing to preclude some of them from working as claimed, even some of the more unusual ones. There have been cases where people wrote off, a priori, the possibility of something making an audible difference, only to be later countered by scientific evidence that proves the difference. Let's not forget the psychologist and philosopher William James' famous proclamation: "Belief creates the actual fact."
Just about all the 'snake oil' products have received rave reviews from one magazine or another, even the ones that run on quantum-state-gobbledygook, or cheap products that are simply relabeled and resold at a higher price. Just read the reviews of Shakti Stones for example: http://www.shakti-innovations.com/reviews.htm
Maybe they work and maybe they don't, but that kind of gushing review isn't uncommon for all kinds of weird products. I doubt very much that all these little audio companies have made amazing new scientific discoveries. In my opinion it shows how untrustworthy personal opinion can be, even when it's the opinion of people with decent equipment and well trained ears, like those Hi-Fi magazine reviewers.
I think we'll have to agree to disagree about that.aef110 wrote:Again, listening is inherently subjective, and blind testing is not the way we should shop for a music system, even if it were practical to set up such tests.
I'm not sure I see the relevance of this to the issue of blind testing. Blind testing removes any effect our preconceptions have on what we hear, such as an expectation that a more expensive option should sound better, but which actually does sound better to you is still your subjective opinion. To me blind testing is just a way of solving a very real problem with the human perception of audio in typical listening tests.aef110 wrote:I've even found components that are audibly, and measurably, more accurate but that I like less than more "colored" gear. Therefore, it's not always about pure accuracy and delineating differences: You might like some of those differences. If this weren't the case, we would all just measure equipment for accuracy and buy accordingly, without having to listen to it. Some manufacturers have tried this approach with disastrous musical results, although they still impress the uninformed with meaningless specs.
These days all the music on my computer is stored as FLAC, actually I think I was one of the people who suggested it in this thread. When I first started encoding music, lossless simply wasn't practical (I think I had a 10Gb hard drive at the time), so I carried out listening tests to find the best balance of file size and quality. I carried out more tests using the latest MP3 encoding options when I bought an MP3 player a few years back. Every so often I'll try new settings or a new lossy file format, just to see for myself if it matches the hype. Storing my music as FLAC made that much more convenient; I can easily transcode with different encoder settings without re-ripping.aef110 wrote:As regards compression, why not switch to FLAC or other lossless formats as has been suggested here? Storage is cheap nowadays, plus you'll spare yourself the definite signal degradation that comes from compressing twice, which all compression inventors insist is a no-no. We don't know what compression schemes will be available in the future or what our portable players of tomorrow will read. Surely it's best to keep a lossless scheme on hand and to listen to the originals wherever possible.
My point was that blind testing is very easy in those circumstances, especially when compared with the difficulty of testing hardware. I found blind testing essential when trying to find the point where lossy audio becomes transparent to me.
Part of the problem with my initial non-blind testing was that I was carefully listening out for sound quality, and expecting a particular file to sound worse than another. That makes little defects in the source recording jump out far more than in normal listening. It's easy to assume that they're defects in the encoding and perceive a drop in sound quality that doesn't actually exist. Comparing the files blindly removed that potential for self deception; my first few ABX tests were a bit of a shock.
I will make one brief comment, with the remark that, while I am a musician, I have a highly scientific mind and very nearly became a scientist. I appreciate the quest for objectivity. But we are dealing with a subjective phenomenon here, and there is no single objective correct answer. Listening to music is not a science, and buying audio equipment by measurement is a recipe for unsatisfactory sound. That's one of the well-known paradoxes of high-end audio.nick705 wrote:... it just rattles my cage when people make sweeping statements like "MP3 is obviously inferior" and decline to provide any data to back up their assertion. If I were to make a post in the storage forum to the effect that "suspending hard drives will cause them to die early" without any evidence to support my claim, I'd be jumped on from a great height and justifiably so.
I have the remarkable benefit here of having many people who are vastly more experienced with PC equipment. You've been an invaluable resource. Although it is off-topic, you have the benefit of having several highly experienced audiophiles and at least one professional classical musician whose business it is to have extremely sensitive ears. That's what we do most of the day, every day: listen. (Sadly, classical music seems to have lost its values and some would-be musicians have blunted their ears.) No one is obligated to listen to my input, but I do hope the little bit of my experience that I can encapsulate into a few sentences will speak to some people and help them make reasonable, practical decisions.
We agree much more than we differ. We're not talking about the esoteric here, we're simply trying to help someone with a small budget get a decent sound system for music. I recommended, as have others, no-frills brands such as NAD, Creek, Rotel, PSB, Cambridge and Vienna Acoustics. They all make good-sounding products at reasonable price points. It's not reasonable, however, to expect to assemble a whole 5.1 channel surround system for 1000 bucks and expect it to sound good. The money would be much better spent on modest "low end of the high end" gear. Equipment at this entry-level price point cannot begin to touch the "voodoo" and snake oil that you are rightly arguing against. It also will most likely not reveal many of the differences between compressed music and the original. That will have to wait until the next upgrade!
Steve,
You got me thinking about blind listening. I think one useful application might be to go to a hi-fi shop and have the dealer play various combinations of equipment within a given price range. Listen blindly, buy what you like best, and you wouldn't be influenced by other factors.
It's blind tests that simply might not be as definitive as we might like. I'm certain that if you took unfamiliar music and played it on unfamiliar equipment in an unfamiliar acoustic and administered a blind test, listeners may very easily fail. The results, however, would be dubious.
You got me thinking about blind listening. I think one useful application might be to go to a hi-fi shop and have the dealer play various combinations of equipment within a given price range. Listen blindly, buy what you like best, and you wouldn't be influenced by other factors.
It's blind tests that simply might not be as definitive as we might like. I'm certain that if you took unfamiliar music and played it on unfamiliar equipment in an unfamiliar acoustic and administered a blind test, listeners may very easily fail. The results, however, would be dubious.