Ok, I asked this over in Head-Fi without much results sadly.  I probably won't see much here either, but I have to at least ask.  Oversampling DACs have always been described as trading an added harshness for more clarity and I have to agree on the harshness.  I've tried a few sound systems, including some external DACs and just recently got the E-MU 0404 PCI which uses an AKM DAC (and, I must say that of all the DACs I've heard, I do like the AKM best overall) but the result has always been the same.  It's not a HUGE difference, but I'm finding that it's enough to not sit well with my current headphones (which I will admit are a bit on the bright side.)  Even on my speakers it can be a little bit harsh at times.  When I compare with something that uses a non-oversampling DAC such as my SCPH-1001 PSX system among other things I have to say that I truly do prefer a DAC that doesn't oversample.  Unfortunately, getting truly good sound without oversampling can be a bit hard whereas oversampling is a pretty easy method to do it, so almost everything has switched over to this method.

What I would like to somehow do is find a way to listen to my music at home at least with a non-oversampling DAC.  Preferably without it costing me an arm and a leg given that I'm pretty broke ATM.  Now, it doesn't necessarily actually have to be a soundcard.  In fact, I was thinking I wouldn't mind it at all if it were some sort of receiver even.  Obviously I don't need some gigantic surround receiver or something though.  My 2.1 speakers have their own amp and work just fine for my needs (and if I were to change I'd go to 2.0, so I'd be even further away from the target of surround receivers.)  My headphone amp is just about perfect I find when I plug it into things like that PSX.  (It's known to be relatively neutral with just a hint of warmth in the mids, which suits me just fine) so I DEFINITELY don't want it to be amped since it's highly unlikely it would happen to sound quite as close to perfect as this and frankly isn't even that likely to be oriented at headphones at all.  In fact, it's not impossible that some might not be able to handle my headphones since they use a low impedance at the cost of a higher current.  (In other words, I need an unamped line output.)  Rather than having to be a soundcard, if it can take digital output (coax or optical -- either is fine as my gaming/etc soundcard can do either one as can even my onboard SPU and if worst comes to worst, I can get a converter later) it would work just fine for my needs.

The key problem, and one of the biggest reasons I can't just use any old receiver is that I'm looking for one with a good sound quality to it.  Ideally maybe even something AKM (though I realize that's probably asking a bit much -- but then again, maybe this whole thing is asking a bit much...)  Really though, it's pretty hard to find many things without an oversampling DAC.  I remember a while ago I found an external DAC+amp that didn't use oversampling, but, I really couldn't afford one that expensive now and obviously I don't need the amp section (which basically means just that more waste of money even if I were to buy that one...)

Any thoughts?

Good NOS DAC? Isn't that an oxymoron?

I recall seeing one or two DAC kits that did not cost an arm and a leg, eg $200. But NOS DACs never really were all that common to begin with. Even some first-generation CD players used 2x oversampling IIRC.

But if you asked most people on this forum (including me) they would probably tell you that an NOS DAC is going to be more harsh, not less, and that whatever preference you do have for the sound can be reproduced 100% in an oversampling design with judicious digital filtering.

 I said it on there and I'll say it here too.  I've found the sound of things like that SCPH-1001 PSX to be nicer and less harsh compared to each oversampling DAC I've tried so far.  I would add that I even get something of a direct comparison as the SCPH-1001 uses a 1-bit non-oversampling AKM DAC and the E-MU 0404 PCI uses a 128x oversampling AKM DAC...  (Specifically, the SCPH-1001 uses the AK4309AVM and the E-MU 0404 PCI uses the AK4396.)  It's DEFINITELY not more harsh in comparison.  (Note that by harsh I primarily refer to what oversampling does -- specifically, to the highs.)

Oh, and note that I'm not asking for highest level audiophile quality or anything like that either. It needs to be GOOD quality, but, not necessarily the best.

Indeed. Do we really have to discuss this? This isn't the snake oil enthusiast's resource, after all. NOS DACs either add imaging or heavy ringing, that were both not part of the original signal. That kind of distortion should be easy to add for those with a strange taste.

Isn't the PSX' non-oversampling property a false myth, anyway?

That's what I want to know.  I didn't ask "do you think NOS sounds better" or anything like that.



How so?  The AK4309 is a 1-bit DAC and doesn't oversample.  It's pretty much as simple as that.  Where is there room for myth or not myth? Or are people saying that the CD player software itself oversamples?

1-bit DACs oversample by definition.

 Assuming the information I read about them (which, btw, is pretty hard to find as something better than a highly technical report) was indeed wrong, then I guess I am looking for a 1-bit DAC perhaps.  Basically, a DAC without as much of that upper end harshness but enough precision at least to still enjoy.

The bullets in the first couple lines of the datasheet you linked mention that it oversamples 256x.



It sounds like you just want something similar to that DAC you linked. Looking at that datasheet, its a pretty typical low end DAC. Have you tried your system's onboard? Its probably very similar.

 Right, I didn't really expect to have any luck here, but I thought I'd try.  Nevermind then.

BTW, the datasheet does not say that it is 256x oversampling -- or at least not in any way that I can see. Maybe the 8x interpolation means this, but that I didn't know about. The only instances I see of the number 256 being mentioned is one of the two supported master clocks is 256fs.

Anyway, just forget it. At least I've cleared up that one thing and will later be able to ask better elsewhere.

I've previously owned a Sigtone (www.sigtone.net) Nos DAC which sounded pretty much as you are describing. As I understand this was basically about the designer aiming for a certain frequency response that gave that perception.

You will probably achieve exactly the same thing by a subtle EQ.


Try something like the following:

2khz: -0,5dB
4khz: -1dB
8khz: -1.5dB
16khz: -2dB

And possibly increase the lower end slightly (maybe +3dB at 60hz?) if you feel the bass is lacking.

It ought to sound less harsh and more "easy" on the ear.

Oh you're right. Its just 8x oversampling. I didn't realize theres a clock divider.

But still, if the main clock is defined to be 256*sampling rate, I think its a pretty big clue the system is oversampling

I invoke TOS 8 and ask that you prove that two different DACs sound different with a level-matched time-synched, level-matched bias-controlled listening test. Feel free to pick one that is oversampled, and the other not. The only other requirement is that the two DACs are rated "excellent" by the Audio Rightmark test. You get points but no cigar if one or both of the DACs come out "very good".

Yeah, uh, there's no way I can or want to do that. I plead no contest.

My point was more that if a difference does exist, it would favor oversampling from a distortion standpoint. That much does not require listening.

I would put it ths way: if the hypothesis being tested was " Typical CDPs will sound the same in well-controlled listening tests", I would disallow a NOS DAC CDP as one of the devices under test, as being markedly atypical. This doesn't mean a NOS DAC might, in fact, be indistinguishable from a typical CDP, just that one has more reason to predict it wouldn't, than normal.

Hm. The NOS dac I previously owned sounded good, and less harsh (possibly due to laidback high frequencies) than the other DACS (onboard cd player, squeezebox and onboard denon surround receiver) I had at the time.

I have no comment on wether it could be reproduced with an oversampling dac or wether a nos dac should by default sound more harsh than an oversampling dac, but that wasn't my experience. I experienced less listening fatigue with that DAC than what I had previously.

EDIT: Small TOS#8 disclaimer, it could of course all have been in my head.

My qualification about passing a Rightmark test with an excellent score eliminates that possible difference.

The CDP 101 had a non-oversampled DAC, and AFAIK it won't score excellent on a Rightmark test because there is a marginally audible high end roll-off.

I think that there were better non-oversampled DACs that had better analog brick wall filtersm and would do better on a Rightmark.

Nevertheless, if anybody tries to ABX regular music on a CDP 101 with a regular 2-channel monitoring system, I predict quick humiliation for the listner.

IME, people who say that one good DAC sounds harsh and another good DAC sounds better, are pretty much admitting that they have limited or zero experience with proper listening tests involving these kinds of differences, because it just ain't that easy.

You're entering the world of diminishing returns. IE solutions now become exponentially more expensive, and results decrease by the same amount.

Before I tell you what I think would be the best solution though, check a few of the very noticeable and easily fixable (free) tweeks. First, I don't know what OS you use, but you need to be certain the info on your drive gets to the DAC unaltered. In the case of VISTA, do a search on WASAPI. For XP, it's ASIO. I never believed XP did it, in spite of my massive attempts. Vista native-ly sounds better, and is better still using a WASAPI plug-in via Foobar2000.

Second, make sure you are using lossless (FLAC) or at least very high quality lossy compression for your files. (320 or above).

Now, as far as non free options, the first thing to remember is audio is a chain, and the chain is as strong as the weakest link. Spend $10K on a DAC if you want to, but put it through your $200 home receiver, and it's going to sound like $200.

Secondly, Your headphones have the single largest cost / change ratio of anything in your chain. Get a good set used on eBay. You can find Sennheiser HD550's and HD600's for less than 200.

So if you have a half decent set of headphones, I'd say the DAC is your next step, and you want to go outboard. all soundcards and internals sound like garbage. For little more than the amount you'll spend on the best Creative X-blaster or whatever, you could get THIS There's also a company called Audio-GD making some good value stuff. Don't be afraid of the overseas part, both companies are well regarded. after that, you can step up your amp(s). Then if your tastes continue to refine, you can start the process over again.

Ultimately for DAC's you want to get away from Op-Amps and use those which have a discrete output section. Then comes better internals, clocking and synching, etc. I'm using a PS audio 3 with Cullen level 3 mods.
I can tell you that the sound produced is worth every penny, but to bring my longwinded post full circle The law of diminishing returns is nowhere more important than in audio. It is far more apparent to step down in quality from your current level and hear incredible differences (IE "How the hell did I listen to this") than it is to step up.

I've been at it for a few years now, and I'm now several levels apart from where I started, which was wanting to get a little better sound than my gaming headphones and soundcard to listen to music.

So far so good.



Here's a TOS 8 challenge for you - prove via a level matched, bias-controlled, time-synched listening test that a $10K DAC necessarily sounds different, let alone better, than a good $200 receiver. Otherwise, cease and desist!



No problem with the idea that transducers are among the weakest links. However the room that the recording is made in, as well as the room that it is played back in if not headphones, are very significant weak links.



Here's the same TOS-8 challenge back at you again, Cease and desist with this sort of noise unless you can properly back it up.



Now you're just talking trash.

Aside to moderator, if this guy won't play by the rules, please ban him!

We won't ban him without warning him first, though that is not to say that said warning won't come without penalties.

@oatmeal769, please review the terms of service to which you agreed as a conditions on participation upon registering. Check this one in particular:
http://www.hydrogenaudio.org/forums/index....974#entry149481
Arny's response to you was spot on, though it should have also extended to cover your unsupported claims about Vista/XP/WASAPI/ASIO.

You determined that how?




You determined that how?




It's a shame you mix good advice, e.g., diminishing returns, with so much that's dubious. There may be no audiophile advice more insidious than that given by people who get stuff half right .

Fair enough. Apologies for assuming I knew the rules without actually reading them. A/B/X comparisons such that are accepted here are generally regarded as subjective, controversial or bogus in other forums to which I belong. I assumed this applied here as well, and was incorrect. Nevertheless, assuming the time is long enough, I think any of what I posted is easily discernible as posted in the rules. Did I actually DO IT with each and every card or DAC? Of course not. I will refrain from making comments and generalizations (all soundcards sound like trash, WASAPI / VISTA, going outboard, etc.) to which I have not personally witnessed.

Also, it seems to me that the second to last word, "listening" of the phrase "level matched, bias-controlled, time-synched listening test" renders said test entirely subjective. Thereby negating any objectivity of said test, and making any statement arrived upon by said means an unsubstantiated opinion. Anyone can hear whatever they want if they really want to. The flip-side would be if one used a pocket transistor radio to A/B two sources. I doubt anyone could hear even the most obvious change.

Again I mean not to buck rules, but I have questions reflecting the seeming contradictions of TOS #1 and TOS #8: Does not TOS #1 say that posts are the opinion of the author? It goes without saying that my opinion is no more valid than anyone else's. I'm sure this all has been discussed elsewhere in the forum ad-nauseum, it just seems that these objections are more semantic than legitimate. If I did as hybris did a few posts earlier and made a "TOS#8 disclaimer", Would I then not catch any heat?

Apologies again if I've offended, or gone outside the rules and spirit of this forum. I'm definitely NOT here to troll or argue.



After all this, if I'm "just not getting it", and need to be banned I would respectfully ask the moderators to simply tell me not to post anything again other than simple questions in the foobar forum. (I'm still on the steep end of the learning curve with foobar, and need all the help I can get.) I can be taken at my word, which would be that I won't post opinion in the forum again.

...and comments regarding sound quality about which you have personally witnessed shall be compliant with TOS #8 or else they are not welcome here.


It seems that you would be completely wrong.


If people truly "hear" whatever they want then how would taking away the ability to see or otherwise know what's creating the sound they hear beyond what is received by the ears alone make any difference?


Yes, but it most certainly doesn't grant you the permission to say anything you wish to say; otherwise many of the rules would also be deemed invalid, not just TOS #8.


You would likely avoid heat, yes, though hybris should have refrained from making such a claim as well, even with the disclaimer. Got that hybris?

You don't seem to understand what ABX means. It's a protocol to eliminate exactly that side effect of subjective testing that you call "anyone can hear whatever they want". For example, Foobar will output a high 'probability of guessing' score when your answers don't correlate with a real audible difference and are just imagination.

ABX can't tell you which gear is better, but it is a perfect tool to evaluate if your "opinion" is even worth considering. ABX can separate imagined audible differences from actually perceived ones. HA is not an opinion petting zoo. If you don't like the TOS8 requirement, the place where you came from might suit you better.

ABX results are not only accepted here, but many times required. Think about it a bit though. How in the world can an ABX comparison be subjective? They're done precisely in order to remove subject bias. How can it be more bogus that someone expecting people believe them only because they heard something or read it from some audio "authority"? Especially considering that it's the same species of animal that see Jesus and the virgin mary, and Elvis, in their toasts every day.

Someone is being dishonest there. The other forums, or science?

You're mistaken. The purpose of a blind listening test is not to somehow make the listening and evaluation experience itself, objective. It doesn't turn the listener into an infallibly accurate perceiver and recorder of reality. What it does is test the listener's subjective experience against a very objective measure: how many times the listener correctly identified 'X'.

That number in turn is compared with statistical probabilities, to come to conclusion as to whether the listener was likely to have been hearing a real difference , versus 'just guessing'.

This is all quite necessary because we Homo sapiens are replete with psychological quirks and biases, conscious and not, that probably have or had some evolutionary advantage, but which keep us from being perfectly 'objective'. Recognition of this fundamental aspect of human nature, is why scientific method demands 'controls' in an experiment.

Obviously not every post here gets subjected to the TOS8 rule - its importance is related to the objectivity of psychoacoustics. That is, nobody's going to ask one to ABX a volume change (that's above 0.1db at least). Instead the focus is, rightly, on subjects which tend to be in conflict with accepted science, and on flat-out statistically unlikely or impossible stuff ("I need this level of quality because 0.001% of the population can hear it!"). It does get confused for hypocrisy often, but it's not.

Bad DBTs do get run, but they almost universally can be identified from the protocol description and the test taking notes. Moreover, one DBT alone does not constitute proof; multiple DBTs are necessary. Bad subjective tests, on the other hand, are impossible to identify...

We're pretty well aware of the objections raised against DBTs in audio, but unfortunately (and this is perhaps something which should be rectified in the wiki), there's no real authoritative collection of all of our counterobjections. I guess Doug Self's site is a good start.

Yes, I know what an A/B/X/ test is, thanks all for sending me the P.M.'s with links.

I used to do them (unknowing that was what they were called, or that it was 'scientific', we just knew that if someone could see the equipment who also had an opinion, he could manipulate the result) with co-workers a lot regarding equipment. I remember particular arguments about how microphones sound different than one another despite being the exact same mic, same year, same mfr. I have even been using the snap in (?) with foobar for a year now , which I think is a great little tool.

Please don't take this as me being stubborn or obtuse, I hear what you all are saying. The way you folks do things seems different, and I'd like to understand better, because this forums opinions will be an excellent place for me to look for counterpoints when I'm doing my own research on audio.

I'm not sure, but I think we're all sort of saying the same thing, I just have a different way of seeing it. I'm just wondering how you know those results are valid, unless you personally witnessed the test along with the person making the statement? Unless each of us here are in the same room at the same time and witness the same test (even if said test takes several listening sessions before each 'pick' is made) How do we all know someone isn't full of cr&% if they come in here and say "When I A/B/X/'ed these two CD Players I scored 90% in favor of there being a difference between the two." ? What if that guy didn't actually do a test, or - probably more commonly - thinks he knows beyond doubt, but hasn't truthfully done the test? What if I'd said in my first post I A/B/X'd my soundcard vs. my outboard DAC. I was right 90% of the time, and to me my DAC sounded better. Would that have been acceptable? I still have my soundcard, and to me the difference is so large, there's no need for a test. I nearly instantly can tell the difference. Similarly, at what point does one need to present such substantiation? I'm pretty sure the general consensus here is that a lossless FLAC file will sound different and subjectively 'better' to just about anyone than a 64k MP3 of the same recording. However, My Mom couldn't tell the difference. At what point is it necessary to deviate from 'consensus' to A/B/X? Even if the difference is there and quite large, again I ask what if one used a pocket transistor radio to A/B two sources. I doubt anyone could hear even the most obvious change.
Accepted and understood - but what if someone is lying about the test, also something our species does often.
Ok, I think I've got it. I guess I'm coming over from two other forums where the validity of A/B/X/ is considered suspect. I'm glad to know, and will act accordingly in the future.

That's why replication is encouraged. You'll see many people here sharing their samples for other people to ABX, and sharing of every little detail of the test is also encouraged.

HA users have a long history with bold, false statements as yours regarding, for example, the relevance of WASAPI exclusive mode or plainly audible differences between DACs. They have been refuted here many times (and repeated with high class equipment). Most turn out as slight volume differences or simply imagination when put to test in a correct ABX setup.

Lying would be kind of pointless, because ABX results aren't accepted as proof. When you have found something opposing HA 'consensus', you don't just post an ABX log and that's it, but include as much documentation as possible for those interested in repeating the phenomenon. The more groundbreaking (regarding the former 'consensus') the claims are the more likely it is that people will repeat your test. Missing documentation or no ABX at all usually only results in some barking at the person and soon after it just gets ignored as the usual noise spilling over from the audiophool boards. And look through some former threads, HA doesn't usually defend its 'consensus' in an ignorant fashion. When protocols and documentation are provided especially those claims that oppose current consensus get a lot of positive attention and motivate repetition.

A while ago I thought myself, that I had found a sample allowing to ABX my high end DAC vs. my Macbook Pro's onboard sound. I posted every detail possible and it turned out that my setup was flawed and the ADC used a digital volume control which corrupted results. After this was eliminated it was impossible to ABX the difference again.

Okay, well thanks guys, I stand corrected RE protocol.

Regarding my newfound knowledge, a few questions instantly emerge:
1.) I believe the difference between XP and VISTA, both 'unmodified' is plainly audible. I've read that it's what they did with the K-mixer and stack mumbo-jumbo being completely different between the two. I do not know of A/B/X data which confirms or refutes that there is a difference, but I have two PCs, side by side, with which I staggered my upgrading to be certain I was okay with Vista. They both have identical MB's and I was able to quickly move my coax from one to the other, playing the same recordings on two instances of WMA. (I wasn't into foobar yet) There is no difference outside of video cards and processors (one is for games, one is for general stuff) To me, the difference was detectable between operating systems.
2.) Similarly, I still have my tianyang(?) ZERO Dac, which I still use for games. Now, both systems have Vista. If I simply move the cables around between the Zero, and the PS Audio the difference is there, subtle of course, but there certainly. I'm not talking about volume changes etc., I hear sticks hitting brass cymbals, instead of hiss-scratch. I hear greater 'space' between instruments - it isn't all moshed together anymore... I can go on but you get the idea, I think my DAC plainly sounds better. My point and question is how do I then go about posting those unscientific observations to be challenged confirmed or refuted here? iis there a database or posting of consolidated test results somewhere?

Another question using a different analogy:
It is generally known (but I suppose not 'proven') that Blu-Ray has an immediately discernible difference vs. DVD. I believe I can easily see it on my big assed plasma display. What if though, we all agreed we'd do an A/B/X between the two using a 4 inch black and white battery operated CRT like Dad used to take to football games. Would the A/B/X still show a difference? I'm not so sure. My point is that as the rest of the audio chain gets better in resolution, differences become more apparent. Just because a bunch of people can't see something, doesn't mean it does not exist.
I simply cannot tell a reliable difference (well, I can, but I'd do poorly on an A/B/X) between any of my four aforementioned DAC's on my home stereo, which isn't a terrible system by any means. (Lord knows I've tried, I'd love to say my $1.5k DAC smokes my $200 one) What I think are differences are not enough for me to rule out all the psycho-acoustic stuff. That doesn't mean that there is no big difference. I can plainly hear differences on my headphone rig.

As an aside, on the other boards I routinely call out anyone to show me anywhere any proof whatsoever that anyone can tell a difference between a cheap power cable and those super expensive ones that are around. I'm not a big believer in cable differences. Not that they don't exist, but because no one has prove-ably heard them.

Again, not criticizing, just learning... Thanks.

1. The only difference between Kmixer and Vista's output is a volume difference of 0,003% (Vista's is identical to the input). This can be verified by looping back Kmixer output via S/PDIF and comparing it to the original. 0,003% volume difference is far from every scientific finding about the thresholds of human hearing. That means there remain three possibilities to explain your findings:

• The heard differences are pure imagination.
• The setup was not as identical as you claim, e. g. the software configuration.
• Your mainboard's audio codec can only process a fixed bitrate, most of the time 48kHz, and the heard differences are not due to Kmixer vs. Vista's mixer but due to different sample rate converter implementations.

2. This comparison, to be valid, requires that you volume match the outputs first with a voltmeter. Else it is very probable, that the perceived differences are just slight volume differences. 1db difference, for example, can't always be identified as what it is by humans, but is commonly described with all kinds of attributes as "better defined transients", "better soundstage", "better bass", etc.

Simple to answer. Tests related to the extremely outlandish claims you made can easily be replicated.

For example you made some incredibly sweeping statements that not even well-known authorities in the usually extreme sport of high end audio will agree with such as:

"all soundcards and internals sound like garbage."

How about an opinion from a well-known "subjectivist"?

http://www.stereophile.com/computeraudio/280/

John Atkinson, editor of Stereophile wrote:

"What I can say is that, at $595, the CardDeluxe is the most cost-effective way of making a PC an integral part of a high-end audio system."

Yes, the Card Deluxe is an internal sound card and no less than John Atkinson thinks its suitable for inclusion in a high end audio system. BTW, this review was done no less than 9 years ago. The art of making internal sound cards has improved by approximately one order of magnitude since then.

As seems to be your habit, your supposition is completely wrong.

It is easy to prove that with the right source material, display and viewer location, there is an immediately obvious reliable discernable difference between the Blu Ray format and the DVD format.

Of course with certain other source material, display and/or viewer location, it can be nearly impossible or impossible to discern the difference.

This is one of the common mistakes that poorly informed people seem to be prone to make, They think that there is some sort of equivalence between reproduction equipment related to seeing and hearing. In fact there is no such thing. Video gear in general does not perform at levels of resolution and dynamic range that would stretch the capabilities of the human eye. In contrast, audio gear, even some very humble-appearing and modestly-priced audio gear has resolution and dynamic range that taxes and evem readily exceds the capabilities of the human ear (and brain).

Speaking of 'outlandish'... Posted right next no less to an ad for one of these ridiculous life changing power cord ad's of which I just spoke. I tend not to believe everything I read in magazines.Really? Could you point me to where I can do some reading about how to quantify a nine-year-approximate order of one magnitude? Did you A/B/X it too, or would that also be outlandish?Poorly informed according to who, you? I think it goes without saying to all but the most myopic, that this was an extremely contrasted hypothetical example made to illustrate a point. I Apologise if you didn't get that.
I see where this is going, and I'm not here to get dog-piled, nor do I want to start trading veiled insults. I am completely and utterly wrong, you are completely and utterly right. It's your sandbox, you win.

Folks, I got it. I won't make sweeping generalizations, nor unsubstantiated opinions. I'm here to learn and interact, and from now on, I'll simply lurch and not post. unless I have a direct and simple question to which I'd like to hear someone call out semantic irrelevancies in rebuttal rather than talk about the meat of the issue.

Ban me if you must, but I don't think I've done anything wrong after being set straight. I'm done in this thread though.

I think you're overreacting and hope by now you have calmed down as I would really like to address one of your earlier posts:


At this point you really need to build a case to back this claim. Let's see what you got...


This doesn't exactly make me feel particularly confident that your research has given you much insight into the subject, though links to what you read could go a long way, especially if authored by reputable people.


As an aside, individual test results typically confirm rather than refute. At this point, however, your assertion that there is a difference on this forum needs to come to a complete stop, everything you've said afterward is irrelevant...


The answer is quite simple, you don't; not unless you can demonstrate that you didn't imagine the difference through a double-blind test.

rpp3po already mentioned that you can record the output of both systems in order to conduct a test after verifying that your recordings were carried out in a very precise fashion and that they are perfectly matched in level and have been edited so that the are synchronized in time. To acknowledge your exaggerated point earlier, when playing back these recordings, of course you don't want to send them through a 5W transistor radio with a 3.5" speaker. Feel free to play them with your Vista machine through your headphones or whatever makes you happy, just make sure you follow the proper ABX protocol. Feel free to repeat the double-blind test while playing the samples back though some other piece of hardware. If you can show that you can tell the difference, feel free to post your results along with the samples that you've used, just make sure they are no longer than 30 seconds if the material is protected by copyright.


Yes, there's a dedicated forum here for listening tests. Will you be contributing any?

But since you're only testing the difference between bluray and DVD capabilities, and not whether certain people at certain distances can tell the difference, you can just run resolution tests and have people sit as close as possible. There is a difference with the audio world, in that unless the image is microscopic, you can always watch closer if you are resolution-limited. I don't think there's an equivalent scenario for audio.

Sure there is - turn the volume up until you can hear the dither, in order to prove that 16-bits isn't enough.

Doesn't work during the actual content (you'll be deafened!) but works during inter-track silence.

(Not so good for proving the "superiority" of vinyl though! )


I think the differences between various OSs and various ways of routing audio within them are well understood. It's sometimes a little trickier to know a-priori exactly what will happen on a specific system, with a specific player, OS, and sound card - but enlightenment comes by re-capturing the digital output and comparing it with the lossless original - using test signals if necessary.

This will tell you a lot more about the issues than listening (mainly because all kinds of subtle changes are almost certainly inaudible, but most people would seek to avoid them anyway when they're unnecessary!). We're not talking about psychoacoustic codecs where listening is vital - we're talking about a (potentially) lossless mathematical system where we want to check that it's working correctly.

I think this is a subtle point to the newcomer, not explained in the HA TOS (probably because it was felt to be obvious / irrelevant) - lossy systems are checked by double-blind listening tests, but (supposedly) lossless systems are checked by a simple binary check, with further mathematical investigation to quantify any issues.

Doing listening tests on (supposedly) lossless systems is a little silly. Why go to the pain of an ABX test when you can do a simple bit comparison?

Cheers,
David.

I was talking about resolution actually. To me what you say would be like zooming into the dark parts of the image to see if an image coming from a 12-bit RAW has more or less noise than one coming from a 14-bit RAW.

No, compared to many people who post here.


You said what you saiid, no smileys, no caveats, no hedge words. If we can't trust you to speak what you believe is true, then we can all simply skip over your posts.


That would be called by many passive agressive behavior. It is a means for among other things, delivering veiled insults. Now what's going on here Scott - first you say that you don't want to start trading veiled insults and next sentence you are!


It is very much in style here to avoid sweeping generalizations and unsubstantiated opinions. ;-)


That would be lurk, I think.

However Scott, you again contradicted yourself in the next sentence. Lurking is not interacting. It is sitting there passively. Nothing wrong with that, but I think it is helpful to not contradict yourself in the next sentence so much of the time. ;-)

From a pure data perspective resolution in digital video is analogous to the number of channels in audio. In Redbook audio you capture 44100 times a second 2 samples of 16 bit depth. In 1080/24p digital video you capture 24 times a second 2073600 samples of 32 (3x8) or more bit depth (it's different for analog video).

The mere numbers demonstrate the vast differences between auditory and visual processing in the brain. Visual information can be captured at a rate of only 24 Hz and the eye is already satisfied, at the same time your retina is a massive sensor matrix, able to capture huge amounts of data from each frame in parallel. Audio information needs much higher temporal resolution but 2 channels are basically already sufficient (considering an individual HRTF is applied).

As playing stereo signals through just one speaker disregards much of the contained information, playing back a 2073600 "channel" signal through a 640x480 matrix can only preserve the quality that fits into a 640x480 matrix.

24fps is very non-transparent wrt reality

the eye only has high spatial resolution at the centre

two delivery channels with HRTFs applied are only sufficient if the listener's head is clamped - head movement destroys the illusion unless head tracking is involved


but the speakers (channels) to pixels "analogy" is really good one, and makes your point very well.

With conventional audio we punch two holes in a wall to hear through; with conventional SD video we punch 704x576=405,504 holes in the wall to see through!

Cheers,
David.

You're right, and I have, ... for the moment, LOLMumbo-Jumbo implied that I'm not a computer expert. It looks to me that Microsoft did some major re-engineering of Window's audio facilities between XP and Vista. What I got out of it was that there were 'sample conversion' issues with XP, and that this was causing problems with latency in recording, and data rate changes that could be audible, depending on what software, drivers, etc. were being used. It's mumbo-jumbo to me, But what got me interested in it were latency issues while I was using a program called Sonor.
Earlier I posted a personally unsubstantiated claim that I believed I heard a difference which I will not do again, as asked. I'd LOVE to A/B/X on it for a couple days though and see if what I believe I heard is replicable. Unfortunately I have no PC with XP available. I can no longer replicate, so I must therefore withdraw the claim.
Here are two Microsoft Articles I had bookmarked from then though, if it helps... One of them claims audibility, I believe, and they are from Microsoft.
http://support.microsoft.com/kb/925901
http://blogs.msdn.com/larryosterman/archiv.../20/471872.aspxOkay, no problem.Cool, I see how it works, thanks. Sounds fun! I'll do some reading about the particulars and try it.
Two more questions though.
Can we assume that just because someone else may A/B/X two samples through a system and not hear a difference, that it does not necessarily mean there isn't one? e.g., If I hear it with my system, others would need to replicate my system to accurately try their hand at it, right?
Is there accounting for whether small inaudible flaws on their own may (or may not) combine or piggy back in all sorts of ways with other inaudible flaws within other equipment, drivers, codec's software, etc. to create an audible difference? The sum being greater than the parts so to speak?
The short of it is, Am I allowed to still make my original assertion about all audio reproduction being a chain, and that chain only being as strong as it's weakest link or links?[

Freddie, my sincerest hope is that you do just that.It's so clear to me now, thank you Dr. Freud! I understand Freud liked contemplating cigars too.Check spelling on your last several posts before pointing at mine, Mr. Kettle.Now you're just acting like a pedantic ass, Freddie.

Yes, exactly. An ABX test either provides or fails to provide evidence that a difference exists.

Proving a negative is ... really hard.

I think that was my original concern with XP. I figured if I was going to spend all this time with foobar, and lossless files, as well as a good rig to listen with, I wanted to be sure the DAC was getting the same numbers in order as are on the drive. Doesn't foobar have a snap-in to do a binary test somewhere? Thought I saw one at one point.

It's possible for there to be an audible difference that cannot be detected by all people under all circumstances, absolutely.


Perhaps, but I don't think this necessarily has to be the case, especially if the person claiming to hear the difference describes the difference as night and day (not that I'm suggesting that you have).


A well-designed test should eliminate everything but the specific item being tested so that this does not happen.


I don't see that as a problem. Others may challenge your assertion and if it hinges on claiming X sounds better than Y then you might have a problem on your hands. If you're talking about reproduction of high frequency and X is a woofer and Y is a tweeter then you're safe. If you're talking about speaker wire then there will likely be an issue.

As said, the only difference, when you play two signals at full volume through Vista and XP is a 0,003% volume difference. This should not be audible in contrast to your claim. Vista's improved sample rate converter only kicks in - well if sample rates must be converted. When your hardware is capable of playing back 44.1kHz files, this is not the case and all you get is a 0,003% volume difference.

Before I switched to Mac I had Foobar on XP also set up to bypass the Kmixer. But the reasons were solely of idealistic nature. I have tried to ABX it on way above average gear and found no reason to claim an audible difference.

This also being why multiple people's results are aggregated within the other forum... The more results the more accurate a given statement likely is. AhhSo...Now wait just a minute. I've got some triaxial braided phase-aligned cobalt infused cryo-balanced race wire you need to hear then. It's like watching color TV vs. black and white for the first time.
Sorry, couldn't resist.

Damn, I wish I had made some recordings. I went back and forth for days, I didn't really think of A/B/X because of what I took for granted was a clearly audible difference. It could just as easily have been a slight volume difference caused by the tiniest voltage irregularity or something weird like that between mainboards, kinda like you posted abut earlier.... This is actually great news - one less thing to worry about.

Okay then, back on topic, 'member I mentioned about op-amps versus a mo'betta output stage? (not just with DAC's but anything with gain) That I can A/B/X, even now, but no one called me out on that. Am I generally considered to be correct in that regard?