I don't know if this has been posted here before (knowing this forum, it probably has), but a pretty good write up about how difficult it is to tell the difference between lossy and lossless audio:

http://www.maximumpc.com/article/do_higher_mp3_bit_rates_pay_off?page=0%2C0

What I found most interesting is that most of the testers could differentiate lossy from lossless on their "favorite" or familiar tracks, but basically had less than coin-flip accuracy on tracks they were not familliar with . . .

Begins to shed some light on why (perhaps?) the HD-DVD camp does not consider lossy audio compression to be an issue.

Indeed... very interesting.

And I'd be willing to bet that in actual rooms (as opposed to headphones), and may HT installations, the ability to discern the differences would be even less.

So, some perspective:

A 160kbps MP3 = 80kbps per channel
A 320kbps MP3 = 160kbps per channel

A 640kbps DD(+) = 107kbps per channel
A 1.5Mbps DD(+)/DTS = 250kbps per channel

Now here's a question: Is the bitrate allocated to a multichannel mix statically allocated amongst the channels?

I tend to suspect the encoder can allocate bits the the channel that needs it most for a given passage. This means that given the fact that surrounds and subs often have far less information being encoded, that the net effect is that the main L/C/R channels may have effective bitrates approaching or better than 500kbps for a majority of the time for a 1.5Mb mix...

Now here's a question: Is the bitrate allocated to a multichannel mix statically allocated amongst the channels?

I tend to suspect the encoder can allocate bits the the channel that needs it most for a given passage. This means that given the fact that surrounds and subs often have far less information being encoded, that the net effect is that the main L/C/R channels may have effective bitrates approaching or better than 500kbps for a majority of the time for a 1.5Mb mix...
You are correct, modern codecs will allocate bits between channels to provide the best possible overall audio quality. The .1 LFE track certainly requires many fewer bits than FL or FR, for example.

WOW! After all the intense debates I see in other threads about audio quality, I'm really surprised by the lack of comments in this thread . . . everyone's probably busy commenting on cheap Wal-Mart / Chineses HD DVD players though . . . .

Using a creative soundcard????? WTF?? Try testing on a real hi-fi system with real speakers.

WOW! After all the intense debates I see in other threads about audio quality, I'm really surprised by the lack of comments in this thread . . . everyone's probably busy commenting on cheap Wal-Mart / Chineses HD DVD players though . . . .
Loud and clear. People ignore what they don't want to hear, or read in this case. The silence is deafening. We do use a lot of auditory expressions don't we?

Many people apparently find it extremely difficult to admit that differences the believe they're hearing may, and in fact, quite often do, exist only in their imagination. Perhaps they weren't paying attention in psychology class. *shrug *

What I found most interesting is that most of the testers could differentiate lossy from lossless on their "favorite" or familiar tracks, but basically had less than coin-flip accuracy on tracks they were not familliar with . . .

d'uh

next you will tell us that because some guy from mid west US that does not know any Asian languages has less then a coin-flip accuracy to tell us which is Korean and which is Japanese that those two languages must be the same one.

Read that and then think what the perceivable differences between 16/48 PCM (most likely lossy when compared to the original audio) and 24/48 DD or DTS. Unless night mode is engaged (noticable by the lack of dynamics that most here complain about) the DD/DTS audio has a good shot at being at least slightly better.

d'uh

next you will tell us that because some guy from mid west US that does not know any Asian languages has less then a coin-flip accuracy to tell us which is Korean and which is Japanese that those two languages must be the same one.


In the long history of goofy, nonsensical analogies, this one might take the cake. :rolleyes:

Read that and then think what the perceivable differences between 16/48 PCM (most likely lossy when compared to the original audio) and 24/48 DD or DTS. Unless night mode is engaged (noticable by the lack of dynamics that most here complain about) the DD/DTS audio has a good shot at being at least slightly better.

Do you mean 24/48 DD+ ? Then there is a chance that you are right. DD is just bad. I think that properly downsampled 16-bit is difficult to distinguish from 24-bit, as the noise floor is still pretty low (yeah, I know I have a post somewhere where I advocate 24-bit, but really an accurate 18-19 is likely enough). You might hear a difference in music, but not likley in a multi-channel soundtrack. The problem with lossy audio is that it doesn't just reduce some insignificant bits -- it causes distortions to the waveform. Sometimes these distortions don't matter, and sometimes they do.

The OP's link is an interesting read, but I do question the quality of the sound-card, and the use or these specific headphones.

A quote on amazon about these headphone:

"The 580 cannot be adequately driven by most portable music players and PC soundcards, which can be another reason for disappointing performance."

A better test would have used an audiophile playback DAC (with good jitter insensitivity), with the computer generating only the digital stream. Also, a high-end stereo might be better than headphones. If you do use headphones, at least drive them with a high-end headphone amp!

I have no doubt that variable bitrate encoding is a step-up from fixed rate. It is clear that at some point, on some songs, you won't be able to tell the difference. The problem is that the point will depend on the song, the listener, and the conditions. It's pretty difficult to make a blanket statement that "this amount of compression is perceptably identical to the original" on all data. That's why the audiophile wants lossless.

You are correct, modern codecs will allocate bits between channels to provide the best possible overall audio quality. The .1 LFE track certainly requires many fewer bits than FL or FR, for example.

Oh but theres more ;)

Modern codecs also use the redundancy between channels to save a lot of bandwidth. So if your going to compare bandwidth between 2 channel and 5.1 channel, then its closer to just twice the bandwidth for similar quality. Thus we have:

160k/s stereo <=> 320k/s 5.1
320k/s stereo <=> 640k/s 5.1

And were not even taking into account that MP3 is somewhat antiquated and easily outperformed by DD+.

Hope that gives some perspective on why 1500k/s DD+ sounds so good. :)

Maybe I misunderstand the results of the posted test. What does it mean when a test labeled something 'pass'? Did the music track pass his test? Or did the tester correctly label something?

- Tom

Maybe I misunderstand the results of the posted test. What does it mean when a test labeled something 'pass'? Did the music track pass his test? Or did the tester correctly label something?

- Tom

The tester was able to correctly label something.

So the tester had to say exactly which encoding it was? Not just whether it was the uncompressed version?

So the tester had to say exactly which encoding it was? Not just whether it was the uncompressed version?

Yes, they were each given twelve extensive trials:

The participants listened to not only the three versions of their own track, but also the three tracks from each of the other participants, for a total of 12 tracks in all. Each participant was allowed to listen to each track as long as he or she could stand it, and was allowed to repeat portions of the track and do A/B testing with the other tracks.

If you look at the results, they're all within what you would expect from choosing by chance.

Yes, they were each given twelve extensive trials:



If you look at the results, they're all within what you would expect from choosing by chance.

Well, there were 48 trials total. Imagine this done with deaf testers who each carried a 3 sided fair coin they could flip. You would expect they would get 1/3 of the trials right, or pass 16 of the total 48 trials. But they got 19.

This is where I've forgotten too much math and statistics. If the null hypothesis is they were using these deaf coin toting testers, what is the probability we would actually see the observed results? Is it significant? Anybody remember how to do this?

Likewise if you only look at the total 12 trials where the testers were listening only to their own well known track. The coin flippers would have been expected to get a total of 4 right but the actual testers got 7 in this case. Can anybody calc the results from this? It intuitively seems more significant. Is it?

- Tom

In the long history of goofy, nonsensical analogies, this one might take the cake. :rolleyes:

Agreed. I was going to reply, and then I thought . . . where do I start? I'm not sure what the comment had to do with the price of tea in China (or the article posted, for that matter).

Do you mean 24/48 DD+ ? Then there is a chance that you are right. DD is just bad. I think that properly downsampled 16-bit is difficult to distinguish from 24-bit, as the noise floor is still pretty low (yeah, I know I have a post somewhere where I advocate 24-bit, but really an accurate 18-19 is likely enough). You might hear a difference in music, but not likley in a multi-channel soundtrack. The problem with lossy audio is that it doesn't just reduce some insignificant bits -- it causes distortions to the waveform. Sometimes these distortions don't matter, and sometimes they do.

The OP's link is an interesting read, but I do question the quality of the sound-card, and the use or these specific headphones.

A quote on amazon about these headphone:

"The 580 cannot be adequately driven by most portable music players and PC soundcards, which can be another reason for disappointing performance."

A better test would have used an audiophile playback DAC (with good jitter insensitivity), with the computer generating only the digital stream. Also, a high-end stereo might be better than headphones. If you do use headphones, at least drive them with a high-end headphone amp!

I have no doubt that variable bitrate encoding is a step-up from fixed rate. It is clear that at some point, on some songs, you won't be able to tell the difference. The problem is that the point will depend on the song, the listener, and the conditions. It's pretty difficult to make a blanket statement that "this amount of compression is perceptably identical to the original" on all data. That's why the audiophile wants lossless.

I wouldn't disagree with anything you've written here -- mostly because I'm not an audiophile (although there was one in the test - and he/she fared no better than the other three testers), and I cannot debate you about what you've written in an informed manner (you likely have forgotten more than I'll know about things audio-related). Having said that, however, while your points may be (likely?) quite valid for audiophiles, I guess I would counter with "How many people fall into the audiophile category?" 3%? 5%? 10%? I don't have any hard data, but I suspect it's towards the low end of those numbers. If so (even at the high end, frankly), I can't see how having 10% of the consuming populace having an "issue" with lossy audio on the next HD disc format being a compelling product design decision for an OEM or studio, or if it were (i.e. BR having available PCM, even though I understand the lossless-ness of PCM is a bit in question due to the bit rate of encode vs. the master), I can't see it necessarily being a "winning" attribute in the BR vs. HD DVD format war. Or said differently, I believe that most mass market consumers (i.e., people who shop at Wal-Mart) when presented with two players -- one that has lossless audio (BR) price at $400 / player and one that has lossy audio (HD DVD) priced at $200/ player -- will chose the less expensive alternative. Or said differently, I believe significant price differentials (assuming BR cannot respond in kind to the recently announced cheap Chinese players) will be a larger factor in affecting adoption rates than relatively minute differences in sound quality of lossless vs. lossy.

I also think that from a very pragmatic point of view -- how most people will be using most of the equipment most of the time (e.g., mass market consumers using their HD equipment to watch a new or at least not intimately familiar movie in a house with ambient noise on the speakers that camp with their HTiB and probably not set up or tuned properly) that the differences betweeen lossless and lossy are probably nil. Obviously, for people who are really into their sound / music (audiophiles), this sucks.

...the HD-DVD camp does not consider lossy audio compression to be an issue.
I would disagree with that part.

I don't know that this is an accurate statement, even if there is a "(perhaps)" in front, to soften ithe statement.

The heart of the debate between the "two camps" on lossless audio is that most people feel that lossless audio should be present on the discs using the advanced codecs' lossless compression, rather than ruining the video by squandering the disc space by using uncompressed L-PCM tracks - such as many Bluray titles did, with the resultant poor PQ.

Of course, one realizes that the only reason this is a debate in the first place is because the Bluray player specs do not require any lossless advanced codecs to be in the players, whereas HD DVD does require TrueHD.

This is why more HD DVD titles have TrueHD lossless tracks than the Bluray titles.

All of that said, I think that it is quite hard to tell the difference between some codecs, -however- I do know that my subjective choice in the "good ol' days" of DVD was always DTS tracks over DD tracks.

There was a very distinct difference to me. Perhaps because 448 DD was handicapped by the bit rate, or perhaps because DTS was able to use 1.5 megabits - but the DTS tracks always sound much more open and dynamic, with faster bass and less "mash" which is the best way I can decribe some of the effects of the lower quality DD tracks I listened to, which muddled the sound.

However, the first time I watched Phantom of the Opera on HD DVD, and switched between the DD+ and TrueHD soundtracks - the difference was INSTANTLY noticeable. It was so much like night and day as to make identification easy.

Having a choice between only two tracks would have made the above project easier, I think, based on my experience with POTO. If any of you have setups where you can compare the lossless tracks, and a copy of POTO, give it a try.

Well, there were 48 trials total. Imagine this done with deaf testers who each carried a 3 sided fair coin they could flip. You would expect they would get 1/3 of the trials right, or pass 16 of the total 48 trials. But they got 19.

This is where I've forgotten too much math and statistics. If the null hypothesis is they were using these deaf coin toting testers, what is the probability we would actually see the observed results? Is it significant? Anybody remember how to do this?

Likewise if you only look at the total 12 trials where the testers were listening only to their own well known track. The coin flippers would have been expected to get a total of 4 right but the actual testers got 7 in this case. Can anybody calc the results from this? It intuitively seems more significant. Is it?

- Tom


The aggregate results are within the range of what might be expected by chance, but I wouldn't read much into this test either way. There were only four participants, and they brought favorite specific recordings with which they were intimately familiar. They were free to perform as many A/B comparisons as they wished, and listen intently as long as they could stand. Though it might seem counter-intuitive, comparative results of exposure to different stimuli for extended periods are usually less accurate, than those where a brief exposure is employed. Conversely, it's hardly scientifically appropriate to allow test subjects to choose a specific stimulus which they've already been exposed to on an unknown, but presumably enormous number of occasions.

d'uh

next you will tell us that because some guy from mid west US that does not know any Asian languages has less then a coin-flip accuracy to tell us which is Korean and which is Japanese that those two languages must be the same one.

In the long history of goofy, nonsensical analogies, this one might take the cake.

Actually, he's quite correct and that's why I have problems with these "test" methodologies.

The acticle asks the participants to SUBJECTIVELY identify which track is lossless, 320kbps and 160kbps. However, this relies completely on the listener's memory and taste (ie which one sounds "better") and has nothing to do about which track is the most ACCURATE REPRESENTATION of the original CD.

The test should only be:
Track A - Original track used for baseline comparison and identified as such;
Track X, Y & Z - randomly uncompressed WAV, compressed at 320kbps and at 160kbps

The only question should be "Can you hear the differences between Track X/Y/Z and Track A?" Unlike visual comparisons (like Xylon's wonderful thread), audio comparisons are more difficult because you still have to rely on your memory a bit to recall what you heard on the original track. I guess the perfect test scenario would be to have the tracks OVERLAP the original.

In that sense, the analogy is correct in that unless you knew exactly what to listen for, most Westerners can't distinguish when a language spoken is japanese and korean unless you had a baseline comparison. I can easily tell you when something is missing; it's much more difficult to choose which one sounds right. Hence, the "tests" by these articles always leads to failure. Heck, wasn't that the big problem with the Traffic HD DVD? Most professional reviewers gave it so-so HD picture quality rating, even though it was most likely an upconverted DVD! We only know that now because Xylon posted a better 720p broadcast version as a direct comparison.

I mean, isn't that what we strive for here is the truthful duplicate of the original? Downrezzing mp3 usually has a smoothing/blending effect, which many listeners take to mean better sounding, but it reality is not how the artist intended it.

Like most AV enthusiasts, we don't want the studios to remove film grain just to make a more pleasing & pretty HD picture, we don't want the audio to be toyed with either. Give us the perfect audio track the way it was intended to be heard.

The only question should be "Can you hear the differences between Track X/Y/Z and Track A?" Unlike visual comparisons (like Xylon's wonderful thread), audio comparisons are more difficult because you still have to rely on your memory a bit to recall what you heard on the original track. I guess the perfect test scenario would be to have the tracks OVERLAP the original.

....

Like most AV enthusiasts, we don't want the studios to remove film grain just to make a more pleasing & pretty HD picture, we don't want the audio to be toyed with either. Give us the perfect audio track the way it was intended to be heard.
I like your testing point, but for the last part, is this the Audio/Video Science Forum or Audiophile/Videophile and where's the Science Forum? It's the equivalent of standing 6" in front of a 100" projection to find artifacts. It doesn't matter. In fact, the science part is that you won't see them under normal viewing conditions. Same with audio codecs - where's the beef?
,

Actually, he's quite correct and that's why I have problems with these "test" methodologies.

The acticle asks the participants to SUBJECTIVELY identify which track is lossless, 320kbps and 160kbps. However, this relies completely on the listener's memory and taste (ie which one sounds "better") and has nothing to do about which track is the most ACCURATE REPRESENTATION of the original CD.

The test should only be:
Track A - Original track used for baseline comparison and identified as such;
Track X, Y & Z - randomly uncompressed WAV, compressed at 320kbps and at 160kbps

I agree that the methodology of this test was decidedly unscientific, but you're also subjectively characterizing how the participants made their choices.

The only question should be "Can you hear the differences between Track X/Y/Z and Track A?"

No, clearly identifying a single reference sample wouldn't be appropriate, because it would introduce a subjective component of certain knowledge that other choices are not the reference sample, and that they must be measured against that particular sample. That's why someone came up with the concept of an ABX comparator.

Unlike visual comparisons (like Xylon's wonderful thread), audio comparisons are more difficult because you still have to rely on your memory a bit to recall what you heard on the original track. I guess the perfect test scenario would be to have the tracks OVERLAP the original.

Level matched, temporally synchronized comparisons between two samples are the protocol used in an ABX procedure. A is known, B is known, and X is a random choice between the two, unknown even to the administrator of the test, until the procedure is complete.

In that sense, the analogy is correct in that unless you knew exactly what to listen for, most Westerners can't distinguish when a language spoken is japanese and korean unless you had a baseline comparison. I can easily tell you when something is missing; it's much more difficult to choose which one sounds right. Hence, the "tests" by these articles always leads to failure. Heck, wasn't that the big problem with the Traffic HD DVD? Most professional reviewers gave it so-so HD picture quality rating, even though it was most likely an upconverted DVD! We only know that now because Xylon posted a better 720p broadcast version as a direct comparison.

It's completely useless analogy, because the test subject is already intimately familiar with one of the musical samples ( or the analogous language, let's say Chinese ) Under that circumstance, distinguishing the familiar sound of Chinese from Japanese and/or Korean would be childs-play for anyone, because the characteristic tonality, phonetic cadence, etc. of each language differs in such a clearly identifiable manner.

I mean, isn't that what we strive for here is the truthful duplicate of the original? Downrezzing mp3 usually has a smoothing/blending effect, which many listeners take to mean better sounding, but it reality is not how the artist intended it.

Like most AV enthusiasts, we don't want the studios to remove film grain just to make a more pleasing & pretty HD picture, we don't want the audio to be toyed with either. Give us the perfect audio track the way it was intended to be heard.

The proper employment of modern data compression can deliver an AV presentation which is perceptibly transparent to the original source material by the most efficient practicable means within the context of available consumer electronics, and it's here to stay, like it or not.

Having only four participants in a study is far from sufficient to do any meaningful statistical analysis of any kind.

BTW, I don't mean to say that good equipment isn't needed to enjoy the best of any audio or video codec. What is coming into play is the limit of our perception. Even at 1080p, you have to avoid seeing the pixels, or be close enough the increased resolution is of value. If you sit at 12 ', stick with 720 for non-FP sizes. The audio limits are more subjective, but there is a reason everything isn't audiophile.

Agreed. I was going to reply, and then I thought . . . where do I start? I'm not sure what the comment had to do with the price of tea in China (or the article posted, for that matter). [/quuote]

it had to do with the flaw in the test. Like you pointed out

[quote]What I found most interesting is that most of the testers could differentiate lossy from lossless on their "favorite" or familiar tracks, but basically had less than coin-flip accuracy on tracks they were not familliar with . . .



my point is how do you want them to? Let me try again. The people were hearing a differnce, they knew the difference, but how do you want them to know if the answer is A, B or C how are they suppposed to know how it is supposed to sound in the first place when it is a song they don't know what it is sup[posed to sound like?

YTou need to undersatand how compression and encoders work. They work by simplyfing the charactersistics of what is encoded. It can change a sound slightly to fit better to the math, it can eliminate some others or add stuff that is not there. To know if it should sound like A, B or C in the first place would be paramount.

PS that is why each person did well on his own songs but the other people did not do well on the others songs.

I like your testing point, but for the last part, is this the Audio/Video Science Forum or Audiophile/Videophile and where's the Science Forum? It's the equivalent of standing 6" in front of a 100" projection to find artifacts. It doesn't matter. In fact, the science part is that you won't see them under normal viewing conditions. Same with audio codecs - where's the beef?
we are AVS we want our science and the quality. Everyone that has been here for years is someone that came here to get something better then what he had. The test should be done at the most exigent specs and if it has a difference at 6" from a 100" screen then go for it even if you sit 150" from that 100" screen. Now if at 6" or 2" or 1" you don't see a difference then that is where science stops and religion starts.

.. how are they suppposed to know how it is supposed to sound in the first place when it is a song they don't know what it is sup[posed to sound like?

I'm not sure what they did here - but they should usually let the testers hear the identified sounds before starting the test.

Likewise if you only look at the total 12 trials where the testers were listening only to their own well known track. The coin flippers would have been expected to get a total of 4 right but the actual testers got 7 in this case. Can anybody calc the results from this? It intuitively seems more significant. Is it?


If you look at the test of first person, he got all the ones he knew correct. That is definitely not insignificant. Chances of that are quite low (3/12c3=1.4%).

we are AVS we want our science and the quality. Everyone that has been here for years is someone that came here to get something better then what he had. The test should be done at the most exigent specs and if it has a difference at 6" from a 100" screen then go for it even if you sit 150" from that 100" screen. Now if at 6" or 2" or 1" you don't see a difference then that is where science stops and religion starts.

So your approach would be to run visual acuity tests that wouldn't be replicated in real world viewing (i.e., 6" from a, for example, 50" screen)? While I understand that approach would help to derive what the differences actually are, I'm not sure of what use such differences would be in given normal usage patterns. Or said differently, I don't think that most people would argue that lossless doesn't provide a more accurate reproduction of the original bits (if it is indeed lossless, that is), nor do I think that most people would argue that 1080p doesn't provide more detail than 720p -- either would be silly to argue. However, I think that many people would argue that in most (average, non-audiophile) conditions, most of the time, good lossy audio (i.e., high variable bit rates) or 720p viewed from an average distance of >8' would be perceived as just as good by most people. Will there be a very small subset of people who are either enthusiasts, obsessive, or downright dellusional (increasing order there, not implying these are all the same) that believe in a given condition or environment or situation that they can perceive a difference? Absolutely! Most A/V equipment manufacturers (not all) don't build equipment for that 1%, 3%, 5% of people because the market size is too small for them.

So, if this is the "best scientific test" forum / discussion area, so be it -- but it has always been my perception (maybe wrongly so) that most discussion here has more to do about more typical conditions. Again, that's just my take, as with all things, YMMV.

I'm not sure what they did here - but they should usually let the testers hear the identified sounds before starting the test.

Nataraj : agree, but does not look like they tried to make a good test?

If you look at the test of first person, he got all the ones he knew correct. That is definitely not insignificant. Chances of that are quite low (3/12c3=1.4%).

that is the funniest part the testers and the fanboys think it proves the opposite of what it does. It shows that anyone that knows what he is looking for can tell the difference. They don't even realize the 4 people in the test 1 is the test object and the other 3 are the control.

So your approach would be to run visual acuity tests that wouldn't be replicated in real world viewing (i.e., 6" from a, for example, 50" screen)?

depends on ewhat you want to test. 6"/50" no, but then showing a difference at that does not mean it is the only difference (is it also at 12"/50"? 20/50?50/50?...) saying there is a difference at A but no one will use A so we will assume there is no difference at B is just stupid and is what this whole thread is about.

The problem is what size screen do you use? how far do you sit. how good your TV would be. If you want a hard proof "no one will" then you need to test the inner limits not the other. For example proving that on a cheap/bad projector at 2x SW and upscaled DVD you can't see a difference won't prove the rest of us that want 1xsw with a decent projector won't see it.

However, I think that many people would argue that in most (average, non-audiophile) conditions, most of the time, good lossy audio (i.e., high variable bit rates) or 720p viewed from an average distance of >8' would be perceived as just as good by most people

I am sure many will argue that snowy over the air SD and bad audio uis good enough, what does that prove?

[quote]So, if this is the "best scientific test" forum / discussion area, so be it -- but it has always been my perception (maybe wrongly so) that most discussion here has more to do about more typical conditions. Again, that's just my take, as with all things, YMMV. [/quopte]

unfortunetly the HDOM section is not. It is filled with insane rabid fanboys. The other forums (projector.....) tend to be where the more "best scientific test" discussions happen.

Now if at 6" or 2" or 1" you don't see a difference then that is where science stops and religion starts.
Not for me. I don't confuse religion and engineering.

Nataraj : agree, but does not look like they tried to make a good test?

Right. I don't think their test will pass AES's rigor ...

depends on ewhat you want to test. 6"/50" no, but then showing a difference at that does not mean it is the only difference (is it also at 12"/50"? 20/50?50/50?...) saying there is a difference at A but no one will use A so we will assume there is no difference at B is just stupid and is what this whole thread is about.

The problem is what size screen do you use? how far do you sit. how good your TV would be. If you want a hard proof "no one will" then you need to test the inner limits not the other. For example proving that on a cheap/bad projector at 2x SW and upscaled DVD you can't see a difference won't prove the rest of us that want 1xsw with a decent projector won't see it.
How do you think they came up with ATSC? They had tons and tons of testing by experts and non-experts under all kinds of conditions. They cleared out your problem of A and B. They knew when they decided on the standard.

Now, why is the audio industry afraid of testing?

Not for me. I don't confuse religion and engineering.

no you don't. You just have religion. You decided math was biased because it sais BD has 67% more space then HD DVD and that is too big a number

How do you think they came up with ATSC? They had tons and tons of testing by experts and non-experts under all kinds of conditions. They cleared out your problem of A and B. They knew when they decided on the standard.

Now, why is the audio industry afraid of testing?

And incidentally, there are resolution/distance charts based on just that research. The ability of a human to perceive resolution for a given arc was determined as part of determing screen sizes and resolutions for a given viewing distance.

Imagine that!

The other day I tested a dvd audio disc (A.I. movie soundtrack) and could tell a noticeable difference between the dolby digital track and the high resolution(lossless) track. The voices were so much clearer.

But I also tried POTO on hddvd and compared the DTS downmix to the truehd track, and had the hardest time telling the difference. On some voices, the truehd track sounded a tad clearer, but for the majority of the movie, it was the same to my ears. Sometimes I even chose the DTS track over the truehd one.

I'm no audiophile, but the fact that I(also my wife) could tell a noticeable difference between DD and lossless shows I'm not completely clueless about lossless. But now from DTS(at least the Downmixed DTS on hddvd) to trueHD? Trust me, I want to hear the difference, but it just isn't there most of the time.

I think Toshiba made this DTS downmix so good, that most people won't really need lossless.

Despite these conclusions I've reached, I'll still use the truHD track when possible, just in case I miss something. But then again what you don't know can't hurt, or help you for that matter, right?