Dave, what is your latency setup? How often is your player awoken and asked to feed the uncompressed data to RAM for the soundcard to read? What other conversions does your chain (including the driver) do? What other processes/threads (both kernel and userspace) run on your system? What is the overhead of the filesystem reading and transferring about four times the data compared to the compressed version? Did you measure e.g. flac decoding CPU load? Did you analyze its impact on the power line in the light of all the other processes running on your multitasking OS?
If you did and you have really tested all of that, your statement is based on solid ground. Otherwise it is just a hypothesis, just like many others. One of the ways to circumvent the measurements/analysis is to prove that the effect makes a difference. How? By a double blind test which no subjective aspects can influence. You can either hear the difference (and your hypothesis is credibly confirmed), or the results suggest otherwise and the hypothesis has not been confirmed (but not rejected, of course).
Since i have memory play enabled, the entire song is loaded into RAM before it plays.
I have the OS running as skinny as i can (with more processes turned off as i find them). I don't have the kind of test gear needed to measure <10 µS glitches (few people do)... probably best done by reading the dynamic jitter at the output
It is just a plausible hypothesis -- a practical explanation as to why some people hear these things (and not likely the only thing) -- it is well understood that quality of power supplies have a serious role to play on sonics. We know that lots of people won't (conciously) hear them because they are not trained. I do not have the time or resources to do a valid double blind test (i've done lots of casual ones), and then it would still need to be independently verified to prove anything.
And, at least with ABX, even if a difference is not detected nothing is proven except that in that particular test no differences were heard.
Measurements are probably cheaper & easier, but until a correlation with what we hear is done, it to is also subjective -- ie i think these measurements explain the phenomenom.
dave
That is just a cache of the decoded samples. I am talking about the RAM buffer for DMA transfer to the soundcard. Some people claim lower playback latency sounds better (and never give a plausible explanation nor ABX test results) yet lower latency loads CPU just like flac decoding, often even significantly more. It is easy to measure, I have done so several times (incl. ABX). Or others claim jplay sounds better, yet its author claims he on purpose uses one whole CPU core for playback and wastes the remaining cycles by useless looping. See - some extra CPU load sounds better, some worse. Well, according to claims of some listeners, there are others who claim the exact opposite (larger latency, minimum CPU load).
Do you know what is interesting? Many of those who say they cannot hear the difference between different latency setup or between flac/wav, present their negative (i.e. inconclusive, not rejecting) ABX tests they found time and resources (miniscule BTW) to perform. On the other hand I have never ever read about a positive ABX test by anyone claiming to be able to hear the difference. Including you. The answer is always the same - "it takes too much time and resources to be properly conducted. I hear what I hear". Well, I know what looks more credible to me.
Jitter has nothing to do with music source. Jitter is a HW phenomenon.
SW at user level determines samples buffer size at best. Application buffer size may differ from audio driver buffer size.
Last edited:
One-sidedness on your side - you pay attention to the CPU (and maybe PSU load) because of the need of decompression, but you do not pay attention to the fact that compressed file requires less disk activity which is less load on the PSU.
If your audio "hears" power supply (my onboard soundcard, alas, hears power supply), your sound system is crap in the first place (mine is), and there is no sense of discussing audiophile nuances.
E.g. a page in web browser may get updated while you are playing music, or the system may check foe online updates.
Essentially, any computer activity can affect its energy consumption.
I think there is an indirect coupling between the music source and sound device clock - noise on the data/supply lines. Various communication technologies are vulnerable to various extent to this issue.
Unfortunately, even some developers of audio playback software (mostly commercial one for closed-source platforms) show complete lack of knowledge how the audio chain works by claiming the CPU communicates with the sound device directly (synchronously) and the timing of processing the samples has thus direct impact on the resultant jitter. I can understand regular users would believe so, but in case of professional developers it is always a sad reading.
There is nothing to talk about in the realm of HiFi audio if routine functioning of the device causes audible changes in noise.
I.e. in such a case everything is unpredictable and unmeasurable.
...
The whole sub-thread about lossless vs uncompressed music looks truly ridiculous to me. Exactly because factors of obviously malfunctioning HW has been brought into picture in the attempts to defend raw audio formats.
of cats and engineering types...
dave: try "Bengali" as a step between the "wild" cats and the domesticated ones. previous version:="Ocelot" (or similar),...last version "Voodoo" (my cat who still thinks he's a mix between a Tiger and a Dragon
), or last version:="Sylvester"
nowhere (and most engineering types, and maybe Dave too): Why not adopt a zero compression scheme? Easy on the ears, easy on precesses required. Bit for bit perfection without any (or minimal) perturbation. The only issue is memory and if dedicated to A/V use, a 500GB or 1TB (or even more) hard disk should be plenty to start with. Hard Disks are relatively inexpensive. And not everbody does "the cloud" thing, or have cd players elsewhere, so very easy to record to cd.
Ignorance is bliss.
All: I suspect a custom "MacMini clone" could be done for less than the cost of a new Mac mini. You select your hardware: mount separate good sound and video cards, a good I/O card (USB, Firewire, etc), and board in the main case, and an oversized 350W ps outside of the main case. Then record everything as a wma file (even in a Mac). The end result? At least a chance of getting a good audio signal out, a ps that "idles" along with no or little load and the ability to customize to your heart's content.
Simpler is better.
The issue regarding a quality USB (or cables in general for audio and video): In the past Bruce Brisson, Noel Lee, George Cardas, Bill Low, Ray Kimber and Roger Skoff had agreed that the difference between cables can be attributed to the dielectric, conductors, and geometries used in a particular cable. The biggest things that constitute a good cable (as stated by all of them) are as follows:
Remember that this is a hobby for most and the (as Dave states) it is that the reason for this hobby is enjoy music. So go enjoy some.
I may skip the last right to Mountain Lion* -- Apple is running out of cats, they are going to have to use a house cat sooner or later (ie OS X Pussy), they can say they have tamed the OS
dave
dave: try "Bengali" as a step between the "wild" cats and the domesticated ones. previous version:="Ocelot" (or similar),...last version "Voodoo" (my cat who still thinks he's a mix between a Tiger and a Dragon
), or last version:="Sylvester"nowhere (and most engineering types, and maybe Dave too): Why not adopt a zero compression scheme? Easy on the ears, easy on precesses required. Bit for bit perfection without any (or minimal) perturbation
. The only issue is memory and if dedicated to A/V use, a 500GB or 1TB (or even more) hard disk should be plenty to start with. Hard Disks are relatively inexpensive. And not everbody does "the cloud" thing, or have cd players elsewhere, so very easy to record to cd.Ignorance is bliss.
All: I suspect a custom "MacMini clone" could be done for less than the cost of a new Mac mini. You select your hardware: mount separate good sound and video cards, a good I/O card (USB, Firewire, etc), and board in the main case, and an oversized 350W ps outside of the main case. Then record everything as a wma file (even in a Mac). The end result? At least a chance of getting a good audio signal out, a ps that "idles" along with no or little load and the ability to customize to your heart's content.
Simpler is better.
The issue regarding a quality USB (or cables in general for audio and video): In the past Bruce Brisson, Noel Lee, George Cardas, Bill Low, Ray Kimber and Roger Skoff had agreed that the difference between cables can be attributed to the dielectric, conductors, and geometries used in a particular cable. The biggest things that constitute a good cable (as stated by all of them) are as follows:
- quality of the cable connector itself
- the workmanship in terminating the cable
- the quality of the conductors themselves
- the quality of the dielectrics used
- reasonably competent design that addresses a particular need
- inductance
- capacitance
- resistance (well perhaps more specifically impedance)
Remember that this is a hobby for most and the (as Dave states) it is that the reason for this hobby is enjoy music. So go enjoy some.
Because this is nonsense on multiple accounts.
Money saved on storage space can be spent on buying better audio HW.
And for those who religiously believe in decompression electrically affecting playback the recipe is simple - uncompress the whole album first - before listening to it. It takes several minutes: ... Actually, much less - the whole "Sticky Fingers" Rolling Stones album is converted from FLAC into WAV:
Code:
flac 1.2.1, Copyright (C) 2000,2001,2002,2003,2004,2005,2006,2007 Josh Coalson
flac comes with ABSOLUTELY NO WARRANTY. This is free software, and you are
welcome to redistribute it under certain conditions. Type `flac' for details.
track01.cdda.flac: done
track02.cdda.flac: done
track03.cdda.flac: done
track04.cdda.flac: done
track05.cdda.flac: done
track06.cdda.flac: done
track07.cdda.flac: done
track08.cdda.flac: done
track09.cdda.flac: done
track10.cdda.flac: done
real 0m22.704s
user 0m14.885s
sys 0m2.128s...
I've heard a lot of jokes about audiophiles, but couldn't imagine it is that bad. In the Microphone DIY list I learned that somebody refused to give positive comments on oxygen-free copper cables motivating it by the fact that his brain is not yet oxygen-free or something like that.
Exposing yet other nonsenses.
IIRC, the whole "Sticky Fingers" album is about 45 minutes long. It was first released on vinyl, and it was typical length.
Since uncompression takes roughly 23 seconds, we have relative CPU load/involvement:
23 / (45 * 60) = 0.0085.
I.e. less than 1%.
"Is there anybody out there ?" (c) going to seriously claim that those CPU "distractions" on uncompression taking place for less than 1% of music duration are going to significantly (in audio sense) affect power supply performance ? Remember, power supply and motherboard have pretty big electrolytic capacitors buffering those peaks of consumption.
IIRC, the whole "Sticky Fingers" album is about 45 minutes long. It was first released on vinyl, and it was typical length.
Since uncompression takes roughly 23 seconds, we have relative CPU load/involvement:
23 / (45 * 60) = 0.0085.
I.e. less than 1%.
"Is there anybody out there ?" (c) going to seriously claim that those CPU "distractions" on uncompression taking place for less than 1% of music duration are going to significantly (in audio sense) affect power supply performance ? Remember, power supply and motherboard have pretty big electrolytic capacitors buffering those peaks of consumption.
nonsense, really?
I guess good enough for the computer between my ears, good enough for me. There are some practical considerations, but non-compressed audio at any level will sound closest to the original. Period. No ABX testing required.
Try this: record your favourite album with a non-compressed format, bit to bit. Then make one compressed with whatever scheme you choose. Now take some (perhaps 20 or so) 50¢ discs and make a 20th generation recording using the same compression scheme. Compare to the non-compressed disc and the 20th generation compressed one. If you can't hear a difference (all else being equal), then for you, you have an answer.This is a typical engineered "solution" even though a solution that requires less engineering is usually the best one (as in simplest, so easiest to get correct)
Should this discussion now enter the quantum mechanical field theory? Of course engineers will argue that if something can't be measured, it (whatever artifact) doesn't exist. Now there's some nonsense. Just because we don't have the technology to measure certain artifacts at this time does not mean that those artifacts don't exist. An example: to test the unified theory, a particle accelerator (such as one found at CERN) would need to be constructed with an average radius equal to the average distance that Jupiter is from our sun. "Compressing" (as in scaling down so we can easily study these particles) does not provide for the complete solution.
'nuff about this. go listen to some music.
I guess good enough for the computer between my ears, good enough for me. There are some practical considerations, but non-compressed audio at any level will sound closest to the original. Period. No ABX testing required.
Try this: record your favourite album with a non-compressed format, bit to bit. Then make one compressed with whatever scheme you choose. Now take some (perhaps 20 or so) 50¢ discs and make a 20th generation recording using the same compression scheme. Compare to the non-compressed disc and the 20th generation compressed one. If you can't hear a difference (all else being equal), then for you, you have an answer.This is a typical engineered "solution" even though a solution that requires less engineering is usually the best one (as in simplest, so easiest to get correct)
er HW? Because of the cost of storage is now so low, not a huge savings. The question I hinted at (without begging it) was the quality of files intended for playback via a respectable level home system and some sort of reasonable compromise.
Should this discussion now enter the quantum mechanical field theory? Of course engineers will argue that if something can't be measured, it (whatever artifact) doesn't exist. Now there's some nonsense. Just because we don't have the technology to measure certain artifacts at this time does not mean that those artifacts don't exist. An example: to test the unified theory, a particle accelerator (such as one found at CERN) would need to be constructed with an average radius equal to the average distance that Jupiter is from our sun. "Compressing" (as in scaling down so we can easily study these particles) does not provide for the complete solution.
'nuff about this. go listen to some music.
Last edited:
I can assure you that my system, thou modest, is not crap as many who have heard it can attest.
If yours is crap then your arguements are completely theoretical since you do not have the resolution or DDR to hear differences even if the are there.
My arguement is also theoretical-- nowhere did i claim that i heard a difference between compressed lossless files & native formats, just that i used AIFF, a choice which was for practical reasons (despite you thinking it is nosense). I did provide a valid hypothesis for why some may hear a difference -- i've never bothered to compare them.
dave
Your ABX tests probably fall into the 95+ % of ABX tests that are confounded in some way. One of the more interesting parts of Kunchur's papers are his description of the measures he had to go to to not have the results confounded by the test.
dave
lossless???
Hi all you seekers!
About all our different music-format there are many thought and meanings(feelings)! I have tried many of the formats that is available in the digital domain. Flac, WMA, Wav and different mp3-formats. The last I have tried is the DSD(burned to DVD from the BlueCoast DSF files).
and my personal feeling(which is a far cry from the truth to everyone else is: The Hi-Rez Wav and DSF files from BlueCoast is somewhat better recorded and better sounding than many Hi-Rez 24/96 Flac files from many other supliers(M&W Musicclub among them).
But many 16/44.1 Wav files that are recorded in good studio with good equipment is equally good - listen to Bill Frisell - Gone-Just like a train. So tight and analog(honest) sounding record in only 16/44.1 !!!
And today I play them on my new Sony SCD-XE800 through my Pass B1 and my MonoBloc L\Amp Sit - and the Alpair 10.2.
So everyone to his own truth clings - but I feel it is mosly up to the engineer and the recording process - but compressed(even FLAC) will make the music a little more "naked" - and Wave and DSD will add some warmth.
Best to all - seekers or builder.
Olav
Hi all you seekers!
About all our different music-format there are many thought and meanings(feelings)! I have tried many of the formats that is available in the digital domain. Flac, WMA, Wav and different mp3-formats. The last I have tried is the DSD(burned to DVD from the BlueCoast DSF files).
and my personal feeling(which is a far cry from the truth to everyone else is: The Hi-Rez Wav and DSF files from BlueCoast is somewhat better recorded and better sounding than many Hi-Rez 24/96 Flac files from many other supliers(M&W Musicclub among them).
But many 16/44.1 Wav files that are recorded in good studio with good equipment is equally good - listen to Bill Frisell - Gone-Just like a train. So tight and analog(honest) sounding record in only 16/44.1 !!!
And today I play them on my new Sony SCD-XE800 through my Pass B1 and my MonoBloc L\Amp Sit - and the Alpair 10.2.
So everyone to his own truth clings - but I feel it is mosly up to the engineer and the recording process - but compressed(even FLAC) will make the music a little more "naked" - and Wave and DSD will add some warmth.
Best to all - seekers or builder.
Olav
This is a really weak argument...
(in the statistical & the real-world sense)
John L.
You are omniscient enuff to dictate how i best spend my time & resources? There are a zillion things more important in my queue.
dave
When sitting next to a PC, I find it is sometimes possible to actually hear actual mechanical noise emanating from the PSU (probably?) that is synchronised perfectly with, say, movement of the mouse. I have a very quiet desktop PC (with no speaker in case you're wondering) that does this quite reliably.
When assessing the audible differences between lossless formats (always an enjoyable pastime), do listeners take into account the mechanical noises that come from their sources? CD motors, PSU fans etc. Even in the next room I would expect these mechanical noises to swamp the genuine format differences by an almost infinite amount.
When assessing the audible differences between lossless formats (always an enjoyable pastime), do listeners take into account the mechanical noises that come from their sources? CD motors, PSU fans etc. Even in the next room I would expect these mechanical noises to swamp the genuine format differences by an almost infinite amount.
Statistics is not Chinese take-out
you can't pick and choose which statistical methodology & results you use to support what your opinions are. That's a fundamental flaw in many discussions... but we see this often... as is the case here
John L.
you can't pick and choose which statistical methodology & results you use to support what your opinions are. That's a fundamental flaw in many discussions... but we see this often... as is the case here
John L.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.