"better than bit-perfect" audio players

[leeperry]

I was wondering if anyone tried xxHighEnd, cMP, Lilith Player and so?

they're supposedly so much better than foobar in ASIO/KS/WASAPI, and I got my own opinion....considering doom9 is a forum made of reasonable ppl, I believe it'd be interesting to discuss it here :devil:

the latest "magic" player is Amarra, it's a wrapper that goes between iTunes and your audio hardware...and using voodoo magic, it will make your music sound *SO* much better :cool:

http://www.weiss-highend.ch/amarra/Amarra.html

http://www.computeraudiophile.com/content/Amarra

[tetsuo55]

None of those players seem to do anything really special.

To sum it all up:
-Load the audio file completely into ram(if enough ram is available), this is basically over the top buffering
-Use Wasapi exclusive mode or kernel streaming, Every selfrespecting player has these modes, or has a plugin to enable them

On top of that could be added (lossy)
-Resampling
-Fancy DSP effects.

Please tell me if i missed anything.
What i did not see though is: any input file is treated with Internal 32bit floating point processing, output dithered to the highest supported value

Basically

ANY > 32bit floating point 192 KHZ > ANY
Then you would manually adjust to:
ANY > 32bit floating point (DAC LIMIT)khz > (DAC LIMIT)bit(DAC LIMIT decides floating point or integer) (DAC LIMIT)khz

[leeperry]

did you actually try them? I did try Lilith Player in ASIO and xxHighEnd in DS(couldn't find how to use anything better).

some ppl seem to believe that different HDD's sound different(!), and that HDD's add jitter....hence the need to copy on a ramdisk...

but jitter is not audible apparently : http://hddaudio.net/?p=460

Track 1 - 30ns
Track 2 - 0ns
Track 3 - 10ns
Track 4 - 100ns
Track 5 - 10ns

if you can hear the diff between #2 and #4, you're most likely half human/half machine :o

[tetsuo55]

i'm not saying those are not good ideas, but the relevance is limited and easy to implement in any player.

[leeperry]

and then Asus boasts about 1ppm clocking on their forthcoming improved soundcard :
http://www.asus.com/News.aspx?N_ID=vVME4fRYwvKILC6y

PC users can now experience 31% less audio jitter interference and enjoy ultimate sound quality that frees the audio signal from unsynchronized input streams and results in extensive music details and lively sound image for the ultimate listening experience. The improvement is most significant during general audio CD playback where the sample rate is 44.1kHz.

now, how far can you get in the bs exactly?

[leeperry]

i'm not saying those are not good ideas, but the relevance is limited and easy to implement in any player.
I don't think you actually tried them, right?

I didn't try cMP, but I tried these on my Asus STX(fit w/ LME49720 op-amps) :

-xxHighEnd, its coder says on computeraudiophile.com that his player uses a lot of EQ/DSP...and that's exactly how it sounds! DSP processing is not cool.
he says he's coded the player of his dreams, and some (fake?) fanboys say they would have been willing to play A LOT of cash for this player :rolleyes:

-Lilith Player, identical to foobar+KS on XP SP3, I tried REALLY hard....on both my KRK Rokit and DT770/600Ω, I couldn't hear any diff.

the idea of HDD's sounding different and adding jitter, that could only be fixed w/ ramdisk caching sounds like UTTER bs to me.

[tetsuo55]

No i did not try them, nor will i.

They don't add any scientifically proven things to the table.
I will accept scientificall and phycoacoustic improvements.
(Basically madVR but for audio)

[leeperry]

well, I kept reading about ppl in awe towards those players...wanted to check it :)

OTOH, this guy says that ASIO in foobar is broken : http://www.xbitlabs.com/articles/multimedia/display/asus-xonar_9.html

apparently going from source file on your HDD > your ears is not quite a trouble-free road as we seem to believe :o

when you manage to bypass KMixer, you're still at the mercy of a poor drivers design or implementation..

I still don't buy the jitter thingie, though! 100ns is WAY more than you would ever encounter IRL....and noone can identify these 5 samples in a DBT/ABX!

a nanosecond is a billionth of a second, what kind of human can hear that exactly? DAC designers such as AKM/BB/CS are not exactly complete retards....plus these chips carry built-in jitter correction :rolleyes:

[Inspector.Gadget]

Hydrogenaudio members routinely debunk "audiophile" claims. If you're into that sort of thing, and don't read it already, you should start :)

[leeperry]

Hydrogenaudio members routinely debunk "audiophile" claims. If you're into that sort of thing, and don't read it already, you should start :)
well these guys are too extreme....they'll tell you that noone can hear KMixer, and that it will not do ANYTHING to your audio.

a friend of mine(who's got a Lynx pro soundcard), was able to capture the same stream on XP SP3 in both KS/DS, and going through WaveSpectra he was able to measure the THD+N/IMD+N distortion rates....and also identify them in DBT.

I'm not willing to buy "sound" bs w/o proofs, but these HA are blindfolded IMO...and when you start discussing it, you either get banned(violating TOS# blabla) or you're being told that they won't play e-penis w/ you :rolleyes:

[tetsuo55]

The effect of kmixer on XP is pretty high, om very high end equipment you can clearly hear it.
vista and up have fixed the problem completely, effects on the audio should be in the 2% range

Still you can avoid that 2% but using exclusive mode

[leeperry]

well, these super smart HA ppl will tell you otherwise :
http://www.hydrogenaudio.org/forums/index.php?showtopic=5755&view=findpost&p=126964

Peter is the main foobar coder...I got banned there for saying that ASIO sounds better than DS+KMixer on XP(as this could mislead newbs :rolleyes:), these ppl are clueless and arrogant to the utmost!

ah well, xbitlabs.com says that ASIO in foobar is broken...and these "whiter than white" players are bs to begin w/...I guess I'll just stick to foobar in KS :o

foobar has a mixing matrix DSP(so you can use your own coeffs, like in ffdshow) and supports 32float VST plugins, great stuf!! :)

[tetsuo55]

Those guys can be a bit too sceptical.

Basically this is the ideal path:

Codec is decoded, using at least 32bits floating point precision (using scientific calculations, the limits of human hearing should be reached with audio encoded in +/- 34 bits floating point, if the nyquist limit applies to bits too then 64bit floating point would be ideal)

Stream is passed directly (no mixers or anything in between) through hdmi/spdiff to reciever (or if none is available straight to the dac of the soundcard)

Now the only thing the audio renderer has to take care of is making sure the audio samples arrive at the dac jitter free(Which it can achieve almost completely by using wasapi exclusive mode and media foundation)

[Chastity]

Just wanted to pop in and mention a few things about leeperry's quest for audio bliss:

1) In regards to the xbitlab's report on otochan's ASIO plugin and Foobar2000, otochan's plugin was replaced by Peter's own implementation because of otochan repeatedly not supporting certain functions for ASIO's compliance. eg. channel mapping. Current ASIO support uses whatever ASIO driver is supplied by the driver developer for their card, and Peter's foo_out_asio implementation.

2) "I got banned there for saying that ASIO sounds better than DS+KMixer on XP" XP's KMixer is notorious for adding distortion, since according to MS's whitesheet, it has a conversion quality of about -85dB, even in 24 or 32bit sources. This was why many software resampler plugins were developed, and better hardware resampling in the X-Fi series.

3) In regards to players sounding better than others: If a player can successfully setup a hardware stream to the soundcard, which is what KS / ASIO / WASAPI does, then output from these players should sound and be identical, barring any DSP applications. DSPs are also limited by whatever bitdepth they can apply their functions.

[leeperry]

well, xbitlabs tried the 0.9 ASIO component and the current is 1.24? I'll take your word on it, though...as there's no release notes available anyway. I plan on doing some RMAA one day or the other, I wanna see how good the LME49720 fare(as these things simply sound of this world...the distortion is simply inaudible :eek:)

foobar's DSP run in 64FP? I thought it was 32FP, and that's what Ozone4 shows me...but this sort of conversions are pretty lossless anyway :)

so anyway, all these jitter stories are utter bs then :confused:

what about Onkyo, who says that their new soundcard has a pretty jitter-free USB implementation?

http://translate.google.com/translate?hl=en&sl=ja&u=http://www2.jp.onkyo.com/product/products.nsf/wavio/C112D5E10EA1F3C9492574DA00020251&ei=Oi80SvzYENurjAf--8yWCg&sa=X&oi=translate&resnum=1&ct=result&prev=/search%3Fq%3DOnkyo%2BSE-U33GXV%26hl%3Den%26lr%3D%26safe%3Doff%26sa%3DG

http://www2.jp.onkyo.com/product/products.nsf/9179d78f072ad94d49256dd7000317dd/c112d5e10ea1f3c9492574da00020251/Feature/4.33B0?OpenElement&FieldElemFormat=jpg

just like Asus, it seems that they're boasting about billionth of seconds jitter....who the hell can hear that :o

[tetsuo55]

jitter in the digital audio world means that tiny parts of the audio signal are lost and have to be guessed back into the stream.

[leeperry]

I think it's clock regeneration that you're describing, every DAC will try to "fill" the gaps in real time w/ a PLL : http://www.google.com/search?hl=en&safe=off&q=clock+regeneration+jitter+dac&lr=&aq=f&oq=&aqi=

basically the idea is that the audio is not right on time, it's slightly "floating"(jiterring)...but we're talking about inaudible jitter anyway, try the link I posted above and see if you can "hear" jitter in the 5 lossless samples ;)

PS: I see my friend Canar is hanging around :D

[neoufo51]

I just use WASAPI output in Foobar and it's pretty much all I need.

[lych_necross]

You forgot to mention that these "high-end players" also use noise sharpening to improve the perceived quality of sound. As for the Hydrogenaudio guys, they are usually right, if not arrogant. The problem with bit-perfect audio playback vs non-bit-perfect playback is, can it be ABX'ed. If a listener cannot ABX two sound files with at least a 95% confidence level, then there is no perceivable difference between the files. As for xxHighEnd, cMP, Lilith Player, and Amarra, pure snake oil. Foobar is all I need in a music player.

[leeperry]

well you can use DS for a week, then move to KS...and enjoy the difference! sound is clearer, soundstage is wider....saying that KMixer is inaudible doesn't make sense to me.

I've read about cMP yesterday, basically the idea is that it will replace the regular XP shell w/ some very basic one....and supposedly you will "hear" it, the sound would be far more "accurate"...mostly because the files are copied on a ramdisk to avoid any jitter.

some ppl say that they can hear a difference between a file on a HDD, and the same on a USB flash key ?! this sounds highly dubious to me, but I'll try :D

the brain is very easy to fool, and far from being an accurate measuring tool.....it seems to me that many of these ppl fall for it :rolleyes:

the term "noise sharpening" sounds scary! I personally like to use Ozone4 on movies, but any DSP on music will only color it.

[tetsuo55]

I think it's clock regeneration that you're describing, every DAC will try to "fill" the gaps in real time w/ a PLL : http://www.google.com/search?hl=en&safe=off&q=clock+regeneration+jitter+dac&lr=&aq=f&oq=&aqi=

basically the idea is that the audio is not right on time, it's slightly "floating"(jiterring)...but we're talking about inaudible jitter anyway, try the link I posted above and see if you can "hear" jitter in the 5 lossless samples ;)

PS: I see my friend Canar is hanging around :D
Clock regereneration is a fix of sorts for jitter (clock drift)

[lych_necross]

Never underestimate the power of placebo. Everything can be considered a placebo unless proven otherwise ;)

[tetsuo55]

If the player buffers enough of the file into memory, so that it is never forced to read from disk realtime, there will be no jitter.

The first place jitter will occur is when the player tells the system to send the audio signal to the soundcard.

Obviously though, having the player constantly read bytes from the disk causes an ever so slight system overhead which could theoretically lead to jitter.
Loading the entire thing into memory before playback ensures the player does not encounter any I/O deadlocks and doesn't put any extra pressure on the system (however so slight)

I have actually seen jitter happen in real life with this tiny reads/writes. It happens in an emulator, the game will save to sram, and the emulator will write to file, but since the file is so tiny windows saves up the writes, and does one massive write every X minutes, during that write FPS drops to a crawl for a few seconds.

[leeperry]

Obviously though, having the player constantly read bytes from the disk causes an ever so slight system overhead which could theoretically lead to jitter.
indeed, w/ 450ms of latency in foobar everything plays fine for me in DirectKS on XP..even 5.1 24/96 losless files :)

I've followed a lot of the guidelines from this PDF : http://imageevent.com/cics/v03theartofbuildingcomputertrnsp?p=0&n=1&m=1&c=1&l=0&w=4&s=0&z=2

the guy is quite nuts, but w/o going SO extreme...he does give great advices! you want the windows shell in the back, and your A/V apps in the highest NT priority...so far works great, and makes the system a lot snappier! that's also why I hate Vista, so many CPU hogging pointless background processes :rolleyes:
Never underestimate the power of placebo. Everything can be considered a placebo unless proven otherwise ;)
well, that's the tricky part...not everything can be measured.

the LME49720HA sounds better than the LME49720NA, yet it can't be measured : http://www.diyaudio.com/forums/showthread.php?threadid=132471&perpage=25&pagenumber=3

everyone notices that the bass response improves and the sound is ever so slightly clearer, but this National engineer says that they measure 100% identical...it's prolly due to its metal case that kills a lot of EMI :confused:

and also audio cables make a difference, I've tried to recable my DT770/600Ω on several occasions...and all it did was saturating the bass and killing the trebles....even w/ uber-snake oil cryo'ed wire :confused:

anyway, it's been thorougly discussed w/ sound engineers on HCFR....foobar in ASIO/KS *will* output bit-perfect audio, period.

but then there's that jitter story, some nutcase on computeraudiophile.com says that he can ABX CD extracted files from EAC/iTunes, and he firmly believes that even though the CRC are identical the jitter is better w/ iTunes ?! he says the CRC/MD5 computing don't take jitter in account.....ah well, I'm gonna grab the Asus ST anyway as it'll be easier to install in my box than the STX(I got plenty free PCI slots)...but I don't expect any sonic improvement because of the "31% less audio jitter interference" :rolleyes:

[Gokumon]

some ppl say that they can hear a difference between a file on a HDD, and the same on a USB flash key ?! this sounds highly dubious to me, but I'll try :D

That's because they already expect to hear a difference so it ends up being a self-reinforcing feedback loop. It's no different than the people who claim they can tell the difference between a 10 dollar HDMI cable and a 100 dollar Monster HDMI cable. They already expect to see the difference so when they plug it in they will see what they want despite the fact that they are getting the exact same output (assuming that both cables are up to spec).

BTW if you want to keep up with the ridiculous audio/videophile snake oil you should get yourself a CD/DVD demagnitizer! (http://www.gcaudio.com/resources/howtos/demagnetization.html)

For a number of years audiophiles have reported improvements in performance after demagnetizing CDs and DVDs. Improved clarity and resolution of fine detail, cleaner top-end and a more developed soundstage are common enhancements. Videophiles offer similar reports of improvement related to picture fidelity, color saturation and detail.

ROFL. The people who make this stuff are absolute geniuses and have to be making a fortune off of these idiotic audio/videophiles.

[tetsuo55]

That's because they already expect to hear a difference so it ends up being a self-reinforcing feedback loop. It's no different than the people who claim they can tell the difference between a 10 dollar HDMI cable and a 100 dollar Monster HDMI cable. They already expect to see the difference so when they plug it in they will see what they want despite the fact that they are getting the exact same output (assuming that both cables are up to spec).bad example, hdmi signal's can degrade when run over 12feet, (visable as snow on the screen)

[Gokumon]

bad example, hdmi signal's can degrade when run over 12feet, (visable as snow on the screen)

I'm not sure what that has to do with what I'm talking about. I'm talking about 2 cables of equivalent length that are both spec-compliant. Both will give the exact same output but people, just like this audiophile nonsense, will claim to see all sorts of differences which just don't exist. Besides your speakers will be far more destructive on the audio waveform than any of this minute amounts of jitter or other issues that people claim to hear will ever be.

[tetsuo55]

I'm not sure what that has to do with what I'm talking about. I'm talking about 2 cables of equivalent length that are both spec-compliant. Both will give the exact same output but people, just like this audiophile nonsense, will claim to see all sorts of differences which just don't exist.i'm affraid HDMI cables behave differently even at same length, depending on the HDMI certificate revision they where made for.

there probably will be a difference between the cheap one and the expensive one.

The 100 dollar price tag is absurd though, the parts probably only cost 8-10$ (the 10$ one's parts costing 10-50ct)

There are differences between cables, but it all comes down to basic physics. The cable has to be thick enough, and short enough to carry the signal from sender to receiver without loss of signal, or introduction of new signal (in the form of emi or radio signals) some 100$ per feet cables don't even manage to get that right.

EDIT: here is an example of a cable company that applies all the technical facts to their cables, without the snakeoil:
http://www.hqproducts.com/eng/products/audiovideo/audioconnection/page_1/artl_HQSS3471/1.5 (Its not an hdmi cable)

The price is very reasonable, in the example given above, this would probably be an 18$ cable, same result as the 100$ one (Any improvements to the 100$ cable should fall into the 2% negligible category)

[Gokumon]

i'm affraid HDMI cables behave differently even at same length, depending on the HDMI certificate revision they where made for.

Fine I'll be even more pedantic. I was referring to two HDMI cables of the same revision that both are spec compliant. They will offer bit-exact output.

there probably will be a difference between the cheap one and the expensive one.

Not in the actual quality of the output. The difference is usually in the connectors that's about it.

The 100 dollar price tag is absurd though, the parts probably only cost 8-10$ (the 10$ one's parts costing 10-50ct)

Hence my entire point.

The fact of the matter is that whatever these people are quibbling over with jitter, etc will be nothing in the grand scheme of the mangling of the audio from your speakers themselves. Anyone who claims to hear any difference when we are talking about jitter on the order of nanoseconds is blowing smoke up their butt and only hears a difference because they expect to and they usually know which sample is which. You do any sort of double blind testing and you show these audiophiles for the frauds they are.

[leeperry]

I'm talking about 2 cables of equivalent length that are both spec-compliant. Both will give the exact same output but people, just like this audiophile nonsense, will claim to see all sorts of differences which just don't exist. Besides your speakers will be far more destructive on the audio waveform than any of this minute amounts of jitter or other issues that people claim to hear will ever be.
well, of course your speakers/headphones will add crazy distortion...but the cleaner the input signal, the cleaner the output IMO. At least that's what I found out when trying different op-amps, these days they are the true "sound makers"...a lot more than DAC's are :
http://www.avsforum.com/avs-vb/archive/index.php/t-792354-p-7.html
Everybody -in different locations, at different times, without knowing from each other- told the same story, that they found the differences between opamps more important than the differences in dac chips.
a LME49720HA literally turns your music into high detailed hi-fi stuff, compared to the usual NE5532/JRC2068 you find on soundcards :eek:
That's because they already expect to hear a difference so it ends up being a self-reinforcing feedback loop. It's no different than the people who claim they can tell the difference between a 10 dollar HDMI cable and a 100 dollar Monster HDMI cable. They already expect to see the difference so when they plug it in they will see what they want despite the fact that they are getting the exact same output (assuming that both cables are up to spec).
HDMI is trickly, because timing is what matters here. there's no EEC, only CRC...and the recipient will have ways to overcome transmission issues, but you can easily guess that a poor cable that doesn't conduct data in all wires at the SAME exact pace will create problems.
anyway, it's a non-issue to me, for the price of noname HDMI/DVI cables here in Europe....I can get real Monster on ebay.com
I'm not saying that they give a visible difference(they more than likely don't), but the plugs are very sturdy, the shielding is top-notch and well..you can get good stuff for the price of noname, why bother?
I paid $30 for this 8ft HDMI/DVI cable(in original sealed box)...that's a good investment to me as my projector runs on a temporary installation, so I plug/unplug the cable everyday.

BTW, you should look at these ppl trying to explain us why their $1500 player is so great...yet it doesn't do any EQ/DSP :D

http://www.head-fi.org/forums/f46/amarra-anyone-using-428550/index5.html

and an interesting review of the STX : http://www.xbitlabs.com/articles/multimedia/display/asus-xonar-essence-stx.html

this soundcard is really mind blowing...only the drivers suck, but I've annoyed the Xonar Product Manager to great extend and he promised to fix them :)

[saint-francis]

Really all of this just gives me a headache. Out of everything I have done to get a better sound (shielded cables, EQ's, Amp's, speakers and so on) rebuilding my room properly has made the largest improvement. A cheap system will beat the bag out of an expensive one if it's in a good room. Good bass traps made movie watching a completely different experience for me. Suddenly I could hear the mid range in a way I had never experienced before.

[leeperry]

indeed, but I mostly use headphones anyway :D

and jitter can be measured: http://www.stereophile.com/digitalprocessors/cambridge_audio_azur_dacmagic_da_converter/index3.html

[tetsuo55]

Because there are hardly any good soundcards for 7.1 audio i decided to use HDMI for everything, and let the expensive dac's in the receiver worry about jitter and stuff.

leeperry, i see your still using XP, Vista and up (much improved in 7) use a way better audio subsystem than XP. One of the things a lot of time was spent on is reducing jitter. (Audio path gets priority over everything else)

You should be able to get a measurable reduction in jitter when using wasapi exclusive mode

[leeperry]

ok, but your receiver prolly uses low quality op-amps(5532 or worse) and caps not meant for audio in the first place.

well XP works you know....I get bitperfect KS in foobar and Reclock, just great :)

HR doesn't work too well on Vista, and there's tons of crapware background processes....I really don't like that. I usually have <20 processes running on XP, on Vista SP1 it was more like 30/35 :rolleyes:

plus I really don't care for the fancy GUI...all I want is a fast and efficient system, that "works".

the master volume is always hardware accelerated(say the KX drivers engineers), so XP doesn't do jack when I send KS to the Asus STX...AFAIK :o

[pirlouy]

I'm not willing to buy "sound" bs w/o proofs, but these HA are blindfolded IMO...and when you start discussing it, you either get banned(violating TOS# blabla) or you're being told that they won't play e-penis w/ you :rolleyes:
Foobar2000 rocks (I speak about functionalities). (Some) Moderators sucks. Canar, with his beloved "TOS#", is indeed not appreciated and caused several people to leave forum. This guy is unable to have a discussion and ban everyone which disagree him. Great. :/

Glad there's not this spirit in Doom9...

[tetsuo55]

ok, but your receiver prolly uses low quality op-amps(5532 or worse) and caps not meant for audio in the first place.I have a onkyo 607 which is the udated version of the prizewinning 606, and is itself getting high scores in review too.

It's recommended for its awesome sound quality at the priceclass it's sold at, i'm using KEF 2005.3 speakers which are also best in priceclass.(Both receiver and speakers sound awesome toghether, a lot of people recommend and buy this combo)

well XP works you know....I get bitperfect KS in foobar and Reclock, just great :)But you still get a ton of jitter (still should be in the nanosecond range)


HR doesn't work too well on Vista, and there's tons of crapware background processes....I really don't like that. I usually have <20 processes running on XP, on Vista SP1 it was more like 30/35 :rolleyes:Vista is great in some areas and terrible in others, Windows 7 beats vista and XP (even in memory usage). Windows 7 runs way faster and more stable on my older 512MB machine. (Win7 was designed with low power low ram netbooks in mind)
Oh and i have 0 problems with HR but i'd rather use MadVR.


plus I really don't care for the fancy GUI...all I want is a fast and efficient system, that "works".Yeah i know what your talking about, its called windows 7.


the master volume is always hardware accelerated(say the KX drivers engineers), so XP doesn't do jack when I send KS to the Asus STX...AFAIK :othats a little too simple. Even kernel streaming is a feature designed and supported by windows. So you still have to pass by a ton of system process etcetera.

[leeperry]

hehe, so how many system processes do you have running at minimum on 7? 20 like XP? I don't see m$ removing bloat from Vista, all they can do is always adding more.

sending digital audio to your amp is indeed an option I guess, if you send uncompressed LPCM over HDMI(from Reclock for instance)...yet there's some problems sometimes w/ Reclock I think, at least that's what I read several times?! and most external amps carry op-amps like the 5532..

since when does XP give more jitter than 7? Vista has that exclusive mode, OK...but it's bit-perfect, just like ASIO/KS :o

HR constantly hiccups on Vista if Aero is enabled, and the jitter is very bad if you disable it anyway...I don't see any other VR doing HR's job this efficiently tbh, HR is a renderer of its own class(unmatched for smoothness IMO).

[tetsuo55]

hehe, so how many system processes do you have running at minimum on 7? 20 like XP? I don't see m$ removing bloat from Vista, all they can do is always adding more.Well they did, and i was as shocked as you are going to be. The number of processes depends on your hardware and settings, it doesn't really matter how many processes there are in windows7 though, speed and memory usage is awesome. You could try it on a 2nd harddisk.

sending digital audio to your amp is indeed an option I guess, if you send uncompressed LPCM over HDMI(from Reclock for instance)...yet there's some problems sometimes w/ Reclock I think, at least that's what I read several times?! and most external amps carry op-amps like the 5532..Reclock is not a good permanent solution, what we really need is a fully native wasapi/Media foundation audio renderer. ?I wouldn't know about the op-amps, but i thought i read the older model used burr browns


since when does XP give more jitter than 7? Vista has that exclusive mode, OK...but it's bit-perfect, just like ASIO/KS :oXP doesn't give more, 7 gives less, it has been designed from the ground up to reduce jitter (unlike xp which has been designed to make beeping sounds) bit-perfect in this case only means that windows does not actively resample the audio stream(but a lot of other things are still going on).


HR constantly hiccups on Vista if Aero is enabled, and the jitter is very bad if you disable it anyway...I don't see any other VR doing HR's job this efficiently tbh, HR is a renderer of its own class(unmatched for smoothness IMO).i wouldn't know about vista, but in windows 7 the entire desktop has been upgraded once again, and the drivers have been updates too (windows 7 aero runs on DX10.1)
Have you ever tried madVR? its better than HR in almost every way imaginable(scientifically proven).

[leeperry]

last time I tried mVR in MPC, it looked amazingly smooth for like 1H, then MPC would randomly drop frames after a while...MPC has always done that on XP(not just on my system), blame it on XP maybe.

KMP+HR is perfectly smooth, and doesn't drop any frame whatsoever...it does matter when you watch your movies on a 2m projection wide screeen :)

ok ok, maybe m$ actually removed all the bloat from 7 after all? I don't think you can lower the background process to 20ish, though...and I'm still doubtful, they know how to let the smoke out but not quite how to put it back in ;)

Reclock is the ONLY way to watch your NTSC/PAL movies at their 24fps genuine speed, together w/ perfectly smooth playback AFAIK....the regular DirectShow playback is meant to drop video frames by design.

Yes, I can easily imagine that anything will be better than XP for A/V rendering, using the HPET is a god bless for realtime stuff! ah, when KMP will support mVR(soon maybe?) and when mVR will have matured...maybe I'll do the big jump, but it'll have to be a lot less bloated than Vista :o

[ShadowVlican]

I just use WASAPI output in Foobar and it's pretty much all I need.
+1 and i've tested my setup with DTS wave files and it indeed has bitperfect playback

You forgot to mention that these "high-end players" also use noise sharpening to improve the perceived quality of sound. As for the Hydrogenaudio guys, they are usually right, if not arrogant. The problem with bit-perfect audio playback vs non-bit-perfect playback is, can it be ABX'ed. If a listener cannot ABX two sound files with at least a 95% confidence level, then there is no perceivable difference between the files. As for xxHighEnd, cMP, Lilith Player, and Amarra, pure snake oil. Foobar is all I need in a music player.
exactly... ABX or go home.

[reepa]

But you still get a ton of jitter (still should be in the nanosecond range)

Why would the operating system affect jitter? If the OS and/or CPU can't keep up then you hear repeated samples or blocks of samples. The DAC is a separate system on the sound card/chip. For example old Sound Blaster can be programmed to fetch a number of samples by itself via DMA.

[tetsuo55]

Why would the operating system affect jitter? If the OS and/or CPU can't keep up then you hear repeated samples or blocks of samples. The DAC is a separate system on the sound card/chip. For example old Sound Blaster can be programmed to fetch a number of samples by itself via DMA.There is jitter in the form of not all samples, or bits not reaching the dac at the same time, due to the system being busy with other stuff and thus temporary blocking bits from passing through.

Every soundcard uses a different approach and a different driver which is why in vista and up all basic logic for each soundcard has to be the same, so the bits always reach the dac at the right time even when the system is busy with other stuff.

It doesn't really matter which driver or player you use your always stuck with windows.
Windows decides in which order bits are processed, windows controls the memory and so forth..

[leeperry]

it takes a miracle to get fully synced A/V on XP mostly because the timers suck to begin w/....HPET is the motto here! this thing is VERY accurate, and is aimed at realtime...XP really isn't :rolleyes:

[leeperry]

about win7 audio improvements : http://blogs.msdn.com/e7/archive/2009/06/17/improving-audio-glitch-resilience-in-windows-7.aspx

actually I can't use Safari or Chrome on my XP SP3 box because of these ugly glitches w/ my Asus STX soundcard....IE6 and Opera9 don't "glitch" :)

[tetsuo55]

yeah that's one of the many improvements.

MS is really taking things seriously with Windows 7. (they are feeling the heat from linux and macos, and vista is not making them a lot of money)
Even with almost every improvement disabled in MPC-HC still get an almost flat jitter line (except for the source/display FPS difference, but with 24p/24 its a flat line)

[honai]

There is jitter in the form of not all samples, or bits not reaching the dac at the same time, due to the system being busy with other stuff and thus temporary blocking bits from passing through.

Every soundcard uses a different approach and a different driver which is why in vista and up all basic logic for each soundcard has to be the same, so the bits always reach the dac at the right time even when the system is busy with other stuff.

It doesn't really matter which driver or player you use your always stuck with windows.
Windows decides in which order bits are processed, windows controls the memory and so forth..

Not true. As a matter of fact audio is being played back at an amazingly low bitrate compared to the bus it's being transferred through, including various OS and driver pipelines, that you should never hear the effects of jitter on any system made in the last 5 years.

What you are trying to describe, I guess, is not jitter but the fact that under certain configurations the audio pipeline might run dry because another task/process/thread is consuming too much CPU time. That's not jitter but resource starvation, and it's not specific to audio but common to anything you do on the computer.

[tetsuo55]

What is the official meaning if 'jitter"

I think its: a difference in clock between source and target, the clocks drift apart.
What i describe is indeed microstarvation, which causes one of the clocks to completely stop (the drift could not be bigger or worse than that)

As you can read in the link leeperry posted, measurable starvation happens to +/- 20% of all systems

[reepa]

I think its: a difference in clock between source and target, the clocks drift apart.
What i describe is indeed microstarvation, which causes one of the clocks to completely stop (the drift could not be bigger or worse than that)

As you can read in the link leeperry posted, measurable starvation happens to +/- 20% of all systems

Why would the clock stop? If your sound hardware doesn't get the samples it needs, you'll hear repeated samples or blocks of samples. For example, if Windows crashes, sound often starts looping. That's not jitter. Jitter is caused by the clock generator supplying the DAC. If the clock is of low quality, it drifts around its programmed frequency. The operating system doesn't affect that, because the clock is a piece of hardware on the sound card/chip.

Also, if the sound hardware has an onboard buffer for samples (why wouldn't it?), the main bus being slightly busy won't affect the DAC.

The operating system isn't the cause of nanosecond level jitter.

[leeperry]

For example, if Windows crashes, sound often starts looping. That's not jitter. Jitter is caused by the clock generator supplying the DAC. If the clock is of low quality, it drifts around its programmed frequency.
good point! when I was getting crazy crashing on my STX soundcard(it doesn't like AC3Filter and CoreAVC CUDA :devil:), the picture froze in KMP/MPC, the sound would loop a bit and it'd simply restart....using ffdshow for A/V decoding fixed the problem altogether.

there's a good description of jitter here : http://www.google.com/search?hl=en&safe=off&q=stereophile+jitter&btnG=Search&lr=&aq=f&oq=&aqi=

and supposedly the 1ppm clock on the Asus ST would yield much lower jitter....but again, noone can hear 100ns jitter(check the samples on the link I gave on the first page!).

hehe tetsuo you're as sold on Win7 as I am on that STX soundcard, I'm listening to the 32bit remastered best of from Barry White in FLAC/KS w/ the LME49720HA op-amps....the SQ is simply out of this world! http://forum-images.hardware.fr/images/perso/alphat.gif

[tetsuo55]

Why would the clock stop? If your sound hardware doesn't get the samples it needs, you'll hear repeated samples or blocks of samples. For example, if Windows crashes, sound often starts looping. That's not jitter. Jitter is caused by the clock generator supplying the DAC. If the clock is of low quality, it drifts around its programmed frequency. The operating system doesn't affect that, because the clock is a piece of hardware on the sound card/chip.

Also, if the sound hardware has an onboard buffer for samples (why wouldn't it?), the main bus being slightly busy won't affect the DAC.

The operating system isn't the cause of nanosecond level jitter.Depends on the cause of the crash. When my hdmi-out gets starved sound stops completely.

Anyway, straight from the horse's mouth:
http://support.microsoft.com/kb/943253

[leeperry]

yeah, it's more or less explained on that link : https://www.microsoft.com/taiwan/whdc/system/sysinternals/mm-timer.mspx

the timers that are used in XP are not accurate enough for realtime stuff, and apparently they have dropped DirectMusic in Vista ?! so they have to rely on them in legacy mode, which are a far cry from HPET....still it's quite a shame that you can't play MIDI to external gear right on time in Vista ?! I guess there's a very good reason why ppl stick to XP to use Cubase :o

I'm personally using the PM clock in XP(which you can force through the boot.ini), it's much better anyway...but still much less accurate than HPET.