Nonesense. All your gzip/lzip/xz/zip/7zip/winzip/peazip programs are lossless compressors.
1) read lossless format defintion;
2) scientifically establish that playing lossless format affects in measurable way DAC performance;
3) scientifically establish that playing raw format affects in measurable way DAC performance less than playing compressed lossless format.
...
For hard core audiophiles: I strongly suggest to check lighting, that is devices which produce light in their rooms. I am absolutely serious.
A couple of year ago I saw strange interference on my OpAmp based amplifier output. Luckily I was able to determine its fundamental was 50Hz, i.e. mains frequency in my case.
The interference shape was quite odd though.
After some thinking I looked at my table lamp - which uses a 13W fluorescent element. When I turned off the lamp the interference disappeared.
Now I'm using the lamp to check how well my circuitry is protected from EMI. I.e. I intentionally move circuitry closer to the lamp and watch whether the observed noise increases.
I got the impression that your erlier arguments that playing compressed lossless format made no differnce vis raw format
dave
Like someone said earlier, you have no idea how any particular file is stored in memory. The operating system may decide to compress it itself without you knowing. The ease of playing back uncompressed data may be outweighed by the higher volume of data that needs to be shuffled over data busses. A CD spins faster if the data's from the inside of the disc - think of all that extra load on the power supply! At any time, the operating system may be doing goodness knows what in parallel with playing your audio.
Basically, if there's contamination from your PSU, you're screwed whatever format you think you're playing from.
Let it go... it ain't worth it.
Basically, if there's contamination from your PSU, you're screwed whatever format you think you're playing from.
Let it go... it ain't worth it.
It is certainly worth a lot of talking and discussion. And I am being more and more convinced this is the principle motivation of an internet audiophile - just talking about things, not trying/testing/practicing them.
I still stick to it. I.e. in the jungle of events happening in the computer is a whole you won't be able to relate anything happening in DAC to specifically compressed <-> raw format difference.
maybe... but
Well, then his back-handed insult doesn't inspire much confidence... he has no idea what my HW is/was... not to mention lack of acknowledgement of vinyl's obvious failings (common amongst latter day vinylphiles trying to validate the megbucks spent on their collections)
John L.
Well, then his back-handed insult doesn't inspire much confidence... he has no idea what my HW is/was... not to mention lack of acknowledgement of vinyl's obvious failings (common amongst latter day vinylphiles trying to validate the megbucks spent on their collections)
John L.
Have you tried to compress music with zip? Work? Why not?
No, no, no. Wrong, wrong, wrong. What has to be scientifically proven is that can be constructed systems without loss, to accept the theory 'lossless'. Currently it seems to be well established by science is the Entropy: no perpetual motion.
If it can be shown, for example, would solve many energy problems, no transmission losses of electricity. No jokes, it would be a breakthrough and a instant Nobel Prize.
But for now, and unfortunately, can not be build systems without losses. Including electronic circuits.
I apologize, wasn't intended to be an insult. I should have said "maybe".
That's true.
Some of the "obvious failings" can be overcome/negated, but we're supposed to be talking about the failings of lossless formats.
I don't match that characterization.
jeff
Last edited:
I think you misunderstand the Second Law. Data compression/decompression IS lossless. But in accordance with thermodynamics, there's an energy input to achieve this.
Yes, if you zip and unzip music files, they're identical and sound that way. That's very easy to prove to yourself using the ABX plug-in in foobar or subtraction via Diffmaker..
Have you tried to compress music with zip? Work? Why not?
No, no, no. Wrong, wrong, wrong. What has to be scientifically proven is that can be constructed systems without loss, to accept the theory 'lossless'. Currently it seems to be well established by science is the Entropy: no perpetual motion.
If it can be shown, for example, would solve many energy problems, no transmission losses of electricity. No jokes, it would be a breakthrough and a instant Nobel Prize.
But for now, and unfortunately, can not be build systems without losses. Including electronic circuits.
You completely misunderstand what I'm saying.
I know it's senseless to compress music with 'gzip' and the such.
But apparently you lack fundamental knowledge about computers and about compression.
Any compression eliminates redundancy in case it is present. In that respect 'flac' and 'gzip' are the same - they just use different algorithms (and justly so) to find the redundancies to be eliminated.
Regarding loss of information - sure. And quite possible I know about this (much) more than you. Start from looking on the web for "synchronizer metastable state". My point is that with completely properly function HW there is no guarantee that when you press a key on your keyboard the keystroke will be detected by HW. You probably can't imagine how many synchronizers there is in modern systems. We all are at the mercy of probabilities. Remember my definition of G-d.
But all those attempts to relate DAC performance to lossless formats is still nonsense. And attempts to talk in this thread about information loss is nonsense too.
This is because bits in computer undergo gazillion transformations in gazillion of places due to gazillion of reasons. And information loss is possible in any of the transformations/places..
Have you heard that certain CPUs (e.g. P4) have ECC in their cache ? And that the cache can be enabled/disabled in BIOS ? Now I suggest to start another thread about influence of that cache being enabled/disabled on how DAC performs.
This thread depicts a grimly sad picture: people understanding almost nothing on the subject (i.e. how computer works and of what parts it consists) take an arbitrary factor (uncompressing lossless music format) and start building outlandish theories not substantiated by any statistical data.
CPU all the time does something. And it really doesn't matter whether it uncompresses a lossless music file or "sits" in the event loop of your GUI from time to time being "distracted" to re-render clock desktop clock applet.
I am saying "all or nothing". That is, either consider all the factors or none of them. There is absolutely no ground to view uncompression of lossless music as the dominant factor.
OK, this sounds interesting. So if I feed a completely asynchronous digital input into a clocked D type flip flop, I may end up with the flip flop in a metastable state. An error one way or another for one clock cycle around a transition might not be a disaster, but if the flip flop oscillated for two seconds it might be.
I see from a quick survey on the WWW that people think you can't design a metastable-free circuit:
As a believer in 'fuzzy' systems I find it hard to believe that a metastable-free system cannot be designed...
But if it's true then, presumably, this problem also applies to ADCs where there is a probability of more than a single LSB of error for more than one sample period..? Another thing for audiophiles to worry about?
The point of people who think so is that in order to avoid metastability one needs infinite gain, and people think there are no real devices/systems with infinite gain.
Is this one of those questions where, if you specify a limited bandwidth input signal (or slew rate or something like that), the infinite requirement goes away..?
I agree that metastability somewhere in a sampling circuit cannot be eliminated when sampling an asynchronous signal.
But I think I can come up with ways to, at least, eliminate the effects of that metastability as long as the incoming slew rate has a maximum limit. From outside the black box, the sampling circuit would be free of metastability (not just a low probability).
(I would like to think!)
You better think about this in terms of a mechanical analogy - I saw the picture in one of the articles I read in early nineties.
Suppose you have solid hemisphere, and on top of it a small solid ball. All this in the usual field of gravity.
Pay attention that the hemisphere is kind of variable inclination "plane" - on top of it inclination is 0 and near the "equator" inclination is 90 degrees.
So, if you push the ball on top of it just a little, it will stop rolling off, but slowly - remember, inclination at the very top is 0.
Now think about the following - the ball is not on top, but you pushed just enough for it to roll up to find exactly on top - then what ? Regarding this latter example - synchronizers are essentially cascaded latches (which flip-flops are too), so the "game" is between input signal change and the latch being made non-transparent but transition at is clock input. After that positive feedback in the latch is the only player, and it can find that metaphorical mechanical ball in any position with any speed.
...
In technical circles they simply assess probability of synchronizer going into metastable state. Rule of thumb is that 3 flip-flops are practically sufficient.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.