According to audiophiles...
[blockquote]We mustn't use an internal sound card, because there is noise, eg from the power supply, inside the case, and it will degrade the sound.
If we use USB interfaces/DACs, or even Coax, similarly, there is a chance that this demon noise will creep along the cable and enter our sound.[/blockquote]
My uneducated view...
[blockquote]A good soundcard is engineered to work in the PC environment. Good? I mean from a company like Lynx, RME, etc.
A good PC sound supply (let's face it, some of the nasty ones make audible noise) will not create a serious noise issue.[/blockquote]
But hey, in many ways, I'm just another audiophile, so, let me ask the experts...
[blockquote]Has the electrical/magnetic/etc noise inside a pc case and it's effect on audio processing ever been studied and quantified?
Can this noise be transmitted by USB/Other cables and cause a problem?
[/blockquote]
In a nutshell, is this a real problem?
Those advices, audiophile or not, are intended to reduce the likelihood of problems. (But audiophiles tend to be extremists)
Said that, what are those problems?
Well, I am unsure, but you might know about the "I move the mouse and there is noise coming from the soundcard", or the " when the harddrive is working, i hear noise out of the soundcard".
This used to happen on old computers and initially also in laptops (due to things being integrated and taking less than adequate care). Things have been changing since then, but still, the usual experience of an user with a soundcard is with an integrated one.
So now, what about discrete (internal, but not integrated), USB, and others?
Well, first of all, it only really matters if we talk about analog out. In digital, the most you could get is some sort of ground loop if using a coax cable (not likely).
So, in analog, what could happen is:
- Internal:
* Power comes from the PSU. The PSU could not have a steady enough power (having a powerful Graphics card, or a very cheap PSU are possible causes)
* Components inside the Box are not properly shielded. Interference could be captured by the soundcard. (This is what happened in the early days. Right now, either the motherboard has a bad design, or there shouldn't be much problems from other components)
* Problems with the data Bus (I don't think this is a problem nowadays, but with SBLive, there used to be cases of IRQ sharing that caused problems. Not really noise related in this case, but problems nonetheless).
- USB:
* If it uses the PC for power, it also can get problems with steady signal. Said that, it shouldn't be an usual problem.
* If it uses external power, depending on how it is done, it could get some ground loop. I don't know how usual this problem could be, if any.
* Obviously, interference from internal components wouldn't be a problem, because there would be a higher distance.
* There shouldn't be data bus problems, but generally USB audio has a higher latency.
In general, if an internal (average to good, not integrated) soundcard gets noise, it is the PC the one to blame. In the end, I believe the decision is more about how the soundcard is intended to be used than how quiet it will be.
Almost all notebook audio has some noise problems when using mains power but is often ok when on battery only. This is basically a notebook power supply design issue. There are several work arounds that can help with many notebooks. There is also a complete alternate type power supply replacement option, so the difficulty seems to be an economic choice by notebook manufacturers.
There have been many professional PCI audio cards that have no noise issues and some USB cards, ones that get their power from the USB line, that have pretty extreme noise problems. It seems obvious that it is all a matter of proper engineering. For non-notebook computers, the internal soundcard noise is largely a myth, except perhaps in the cheaper motherboard chip types.
Well that didn't answer the question ... but then I'm not going to either
If you care about good sound quality you won't be using your computer to generate it. I have no idea what your definition of "good soundcard" might be, or what you are listening through (powered speakers, headphones, etc). "Noise" shouldn't be an issue until you have an analogue signal. Degradation at the digital stage would primarily be due to poor quality components. So look at where your analogue audio signal starts and ends, and yes you probably will get some noise, but no it probably won't be the critical factor in your music quality. That said, there are plenty of computer speakers that hum, and worse.
P.S. Just occurred to me, but you might even be considering playing directly from an optical disc drive. That will give you acceptable sound, but you certainly shouldn't be worrying about the rest of the signal chain if you're doing that. Quality CD/DVD drives exist for computers (perhaps not quite Naim, but decent) but I'm fairly sure what you have will very rarely give you a jitter-free output.
If you don't hear any noise, don't worry about it!
Inside of a computer there is a ton of electromagnetic noise... It's not just the power supply. A soundcard manufacturer can do their best to make the card immune to noise, but every computer is different. I can hear some noise under come conditions (with my motherboard soundchip).
A computer power supply can be noisy... Once, I had the bright idea of mounting a power-jack in a blank PCI cover-panel and connecting it to the internal 12V supply, because I didn't have the wall wart for my speakers. The noise was horrible.
I've got a USB soundcard and it's "dead silent" with headphones when the sound is stopped/paused. (There is always some noise, but I can't hear it.) It's possible to get noise through USB power, but USB soundcards are isolated from all the other noise inside the computer. An external USB audio interface with it's own power supply will be completely immune to noise from the computer.
If you care about good sound quality you won't be using your computer to generate it.
I hope artists, DJ's, composers "et al" don't take that too literally
If you care about good sound quality you won't be using your computer to generate it.
I’d like to assume you mean something else and have just worded it badly – because taken verbatim, this is patent FUD (http://en.wikipedia.org/wiki/Fear,_uncertainty_and_doubt).
And I say this as someone who in many cases prefers dedicated audio hardware – but because those units offer a different user experience, stability due to being separate, and most importantly features I’ve not seen reproduced, at least cheaply, on a computer – not because computer-generated audio is somehow intrinsically inferior. It’s not. In fact, in many cases, hardware units have limitations to their processing and outputs that DAWs, VSTis, and so on
can beat.
So what did you really mean?
P.S. Just occurred to me, but you might even be considering playing directly from an optical disc drive. That will give you acceptable sound, but you certainly shouldn't be worrying about the rest of the signal chain if you're doing that. Quality CD/DVD drives exist for computers (perhaps not quite Naim, but decent) but I'm fairly sure what you have will very rarely give you a jitter-free output.
Oh: more nonsense. Are you aware that in any half-modern computer, CDDA is passed digitally to the playing application in a buffered form and is not subject to the mythical “jitter”? or any other supposed aspect of variation between drives (unless the variation is so great that the drive simply doesn’t work) I’ll also wager that, as well as not being unique to CDs, with any half-decent soundcard, the much-maligned “jitter” isn’t going to matter at all
Almost all notebook audio has some noise problems when using mains power but is often ok when on battery only. This is basically a notebook power supply design issue. There are several work arounds that can help with many notebooks. There is also a complete alternate type power supply replacement option, so the difficulty seems to be an economic choice by notebook manufacturers.
There have been many professional PCI audio cards that have no noise issues and some USB cards, ones that get their power from the USB line, that have pretty extreme noise problems. It seems obvious that it is all a matter of proper engineering. For non-notebook computers, the internal soundcard noise is largely a myth, except perhaps in the cheaper motherboard chip types.
Well, now, that certainly
is an answer, and a pretty good one. I was not aware of the laptop problem.
By the way, I don't have a problem. I haven't had a problem since I installed an RME card in a PC over ten years ago, and
really liked the result. I was very happy with both the analogue out (compared subjectively to other components) and digital toslink out played through my hifi. This, of course, was before
I was told I shouldn't be doing that I've had problems with PC audio from the dreaded DPC latency making a card/PC combination unusable to struggling with a firewire device and Linux/JACK, but I have never had sound quality issues with a
working PC system. I've used a PCI card, a Firewire Interface, and now a USB DAC.
But I'm interested in audio, hifi, etc (as I guess we all are) and mix, online and off, with others of various levels of audiophilia (I'm an audiophile in denial ). I hear these things, and I think, according to my experience, that is wrong --- but I have no evidence to offer, and if someone gave me an oscilloscope, I wouldn't know where to start.
Hmm... RMAA software ... aren't the (noise) result normally worse than what you would expect from the ADC & DAC (+opamp) by its specs (DAC or ADC itself probably isn't very noisy and normally the opamps aren't either)?
Both my Audiophile 2496s measure within about a dB of manufacturer specs with RMAA. This is the same as my M-Audio, wall wart powered, USB interface. Since I've noticed some significant differences depending upon where the computer is (I've used several computers with M-Audio and Echo PCI soundcards to record in different locations), I suspect most of the noise in some recordings (which is measurable when no signal is present, but not high enough to effect the recordings) comes from outside the computer.
Other equipment in the signal path, and hum/EMI from all the power cords, antennas, etc. that are too often too difficult to get far enough away from the audio signal path, are the culprits. Serious professionals would most likely be more concerned about sanitizing the recording environment, and using balanced input lines where appropriate, but I don't have to please anyone but myself. For playback, there is no issue.
Competently designed equipment will have its own voltage regulation and filtering, making the power supply largely irrelevant. Since the noise on an ATX power supply is fairly high frequency (well above audio frequencies), removing it is generally not difficult. Of course, often doing so is not a priority (e.g. many laptops).
Competently designed equipment will have its own voltage regulation and filtering, making the power supply largely irrelevant. Since the noise on an ATX power supply is fairly high frequency (well above audio frequencies), removing it is generally not difficult. Of course, often doing so is not a priority (e.g. many laptops).
And that lack of priority might extend to desktop motherboard built-in sound-card chips too? Probably just the same as laptop built-in, I suppose.
Thanks again for more
real-world responses on this.
I recall a Lynx Studio sales Q&A which went something like:
Audiophile buyers say that one can't use the analogue-out on a PC soundcard; answer, wrong, Lynx products are designed for high quality analogue out. I think it was an internal sales-dept document, and although it was, obviously, visible to the net, I have never been able to find it again.
In my experience, PSU noise is not an issue with onboard audio. On the other hand, depending on the motherboard implementation (especially cheaper microATX ones) there will be audible interference from nearby circuits (CPU voltage regulators, USB busses and external GPUs being the usual culprits).
For example, my Asrock motherboard produces interference that is audible during silent passages with high volume set on the external amplifier when moving the mouse or transferring data from devices connected to the rear USB ports. But the noise in only audible under those specific circumstances I mentioned, i.e. large gain on the analog side and low signal level from the source. The board is dead now though and I'll get a new one with digital audio output.
Thank you for an example in which it was an actual problem, rather than an assumed, maybe problem.
When I was initially struggling with my LInux/Jack/Firewire thing, I had, at one time, a situation where any screen update woudl cause drop-outs... but you wouldn't describe that as "noise," it was really "failure!"
It's true that my listening is only at low to moderate volumes.
In my experience, PSU noise is not an issue with onboard audio. On the other hand, depending on the motherboard implementation (especially cheaper microATX ones) there will be audible interference from nearby circuits (CPU voltage regulators, USB busses and external GPUs being the usual culprits).
For example, my Asrock motherboard produces interference that is audible during silent passages with high volume set on the external amplifier when moving the mouse or transferring data from devices connected to the rear USB ports. But the noise in only audible under those specific circumstances I mentioned, i.e. large gain on the analog side and low signal level from the source. The board is dead now though and I'll get a new one with digital audio output.
That is exactly the type of noise I would get from both the onboard sound (also an ASRock) and from my USB soundcard. It was low-level noise that was audible with the amplifier turned way up (loud, but still bearable levels) or with headphones. It would increase concurrently with large screen updates and also at a slightly lower level with mouse movement. The noise sort of had a regular beat to it, probably some kind of oscillation.
The noise was completely eliminated when I switched to an external DAC connect via toslink. My best guess is that the noise existed on the ground plane, which would explain why it could affect the USB soundcard. I live in an old apartment, so none of my outlets have a ground lead, so it's very likely that I had some kind of ground loop situation, although it was more rhythmic rather than an outright hum.
Pretty much all of the components in a PC (except maybe fans other very simple motors) are drawing current with a rate that varies in the many KHz to MHz. This means they are injecting ultrasonic to radiofrequency energy back into the power supply. This energy is nearly all outside of the range of human hearing, but if designers are careless, they can allow it to enter into the transistors and diodes that exist in ICs, and thus be downmixed into the audio frequency range. The fact that the noise didn't originate in the power supply itself isn't too relevant from the point of view of an audio device, its still a variation in input voltage with time that needs to be blocked.
Just moving a converter outside of a PC as in the USB example above obviously does nothing in this case, as the USB port is still electrically connected to the same power plane as a PCI part. If a designer is careless, the same signals can be coupled into the device in either case. Conversely, switching to toslink does prevent this problem, as toslink does not have an electrical connection and thus no energy can be coupled through it. Of course, breaking an electrical connection is not actually necessary, attenuating signals that are many octaves out of band 100+ dB below signal is not very difficult or expensive, and most quality PCI sound cards manage this just fine.
Well, I am unsure, but you might know about the "I move the mouse and there is noise coming from the soundcard"...
My computer is exactly like that. When I move the mouse, scroll down web page or move a window I hear some kind of noise coming from my headphone/speakers, it is easily audible.
My sound card is not an onboard/integrated, it is on PCI slot. Then I decided to make a test if I could hear that noise while something was playing through alsa using hardware device.
The result was: I couldn't hear anything but the music being played (couldn't notice a distortion)... I tested a silent track as well, couldn't hear it (the weird noise)...
So, I concluded that was not an issue... At least for my ears...
Now with my newer computer, I couldn't hear it... Actually, only once for a limited time (don't know why).
Anyway, I may be talking ****. If yes, please let me know.
Well, I am unsure, but you might know about the "I move the mouse and there is noise coming from the soundcard"...
My computer is exactly like that. When I move the mouse, scroll down web page or move a window I hear some kind of noise coming from my headphone/speakers, it is easily audible.
Been there, done that.
This problem is usually in the digital domain of software and computational domain of CPU chips, RAM chips and mass storage; and not in the analog domain of power supplies or EMI inside the case.
It is usually fixed by adjustments to software - device drivers, priorities and such. However, it can also be fixed with a different audio interface, mostly because a different audio interface comes with different device drivers.
My sound card is not an onboard/integrated, it is on PCI slot. Then I decided to make a test if I could hear that noise while something was playing through alsa using hardware device.
Oh, *nix. The world of questionable device drivers supported by tiny populations of users. Usually self-induced.
The result was: I couldn't hear anything but the music being played (couldn't notice a distortion)... I tested a silent track as well, couldn't hear it (the weird noise)...
So, I concluded that was not an issue... At least for my ears...
Now with my newer computer, I couldn't hear it... Actually, only once for a limited time (don't know why).
Anyway, I may be talking ****. If yes, please let me know.
I'd first try putting the PCI card into a different slot.
My computer is exactly like that. When I move the mouse, scroll down web page or move a window I hear some kind of noise coming from my headphone/speakers, it is easily audible.
I have had a computer like that several years ago. Asus motherboard using internal USB for mouse and internal soundcard, I believe.
-k
Oh, *nix.
Me too. although my opening-post question is not OS-dependent. Most of the time when people talk along lines of "you can't blah blah blah because of noise" they are talking about the OS that most of the people use.
Your comment about software-domain interference is very apt. Thanks.
Another specific, if I may.... There is a tweak in which the +ve power conductor in a USB cable is cut
to stop it carrying noise from the PC into a DAC. Of course, this relates to boxes which are not USB-powered. Is there any objective possibility that this is true? Could it depend on how the USB interface handles the un-needed connection inside the DAC?
Oh, *nix.
Me too. although my opening-post question is not OS-dependent. Most of the time when people talk along lines of "you can't blah blah blah because of noise" they are talking about the OS that most of the people use.
Your comment about software-domain interference is very apt. Thanks.
Another specific, if I may.... There is a tweak in which the +ve power conductor in a USB cable is cut to stop it carrying noise from the PC into a DAC. Of course, this relates to boxes which are not USB-powered. Is there any objective possibility that this is true? Could it depend on how the USB interface handles the un-needed connection inside the DAC?
Let me put it this way. Right after setting up a new computer with a FIIO E5 headphone amp (USB powered) I had this horrible noise. I had been powering the E5 off of one of the USB ports on the computer. I rummaged through my stuff, picked out a generic switchmode USB power supply, powered the E5 with it, and obtained utter silent bliss.
Later on I ran into a similar situation while trying to power an E5 and a Sansa Fuze from two different ports on a portable USB charger. The E5 sounded great if I unplugged it from the charger and ran it on its own battery. I think I'm figureing this out... ;-)
I use a FiiO E7 USB DAC/headphone amp at work. My old work desktop had a lot of noise through it's headphone socket and even my shiny new laptop (about 2 months old) has some very slight noise on it's headphone socket. I use IEMs (Westone UM3x) and there is no noise at all through the E7 so I'll continue to use it. It also depends on how sensitive your equipment is.
Let me put it this way. Right after setting up a new computer with a FIIO E5 headphone amp (USB powered) I had this horrible noise. I had been powering the E5 off of one of the USB ports on the computer. I rummaged through my stuff, picked out a generic switchmode USB power supply, powered the E5 with it, and obtained utter silent bliss.
Later on I ran into a similar situation while trying to power an E5 and a Sansa Fuze from two different ports on a portable USB charger. The E5 sounded great if I unplugged it from the charger and ran it on its own battery. I think I'm figureing this out... ;-)
Good. Noise
can be transmitted over USB power then.
Is there anything to suggest that it can be transmitted over the power lines of a USB connection to a device which is
not USB powered?
Let me put it this way. Right after setting up a new computer with a FIIO E5 headphone amp (USB powered) I had this horrible noise. I had been powering the E5 off of one of the USB ports on the computer. I rummaged through my stuff, picked out a generic switchmode USB power supply, powered the E5 with it, and obtained utter silent bliss.
Later on I ran into a similar situation while trying to power an E5 and a Sansa Fuze from two different ports on a portable USB charger. The E5 sounded great if I unplugged it from the charger and ran it on its own battery. I think I'm figureing this out... ;-)
Good. Noise can be transmitted over USB power then.
Is there anything to suggest that it can be transmitted over the power lines of a USB connection to a device which is not USB powered?
Edit: I guess I was unclear above. Yes, if you are electrically connected, then you can transmit a signal.
If a signal can be transmitted, noise can be transmitted. Electrically, they are both the same. That doesn't mean that the device design can't have extremely high noise rejection for noise that is not mixed with the signal. In other words, unless there is a defect, like a faulty cable, if the source audio is free of noise, noise is unlikely to be a problem on the audio output if either
a) the device is externally powered, i.e. not via USB
or
b) the USB power is properly filtered by the device
Of course, it is also possible to build a device with power supply that is so poorly done that noise does get through from its power supply. In the old tube circuit days, getting adequate filtering was significantly harder and more expensive than today. Electrolytic capacitors for high voltages were very large and somewhat expensive. Massive chokes were also used when expense be dammed was part of the design. I have various old LP, very well performed and recorded, where hum is definitely part of the listening experience, courtesy of the recording electronics of the day (i.e. audio that does not meet the criteria above: "if the source audio is free of noise ..."
Let me put it this way. Right after setting up a new computer with a FIIO E5 headphone amp (USB powered) I had this horrible noise. I had been powering the E5 off of one of the USB ports on the computer. I rummaged through my stuff, picked out a generic switchmode USB power supply, powered the E5 with it, and obtained utter silent bliss.
Later on I ran into a similar situation while trying to power an E5 and a Sansa Fuze from two different ports on a portable USB charger. The E5 sounded great if I unplugged it from the charger and ran it on its own battery. I think I'm figureing this out... ;-)
Good. Noise can be transmitted over USB power then.
Is there anything to suggest that it can be transmitted over the power lines of a USB connection to a device which is not USB powered?
Bottom line is that USB is based on a metallic conductive connection between the USB device (power or signal or both) and that makes it a potential participant in ground loops.
That suggests that noise can be transmitted via USB caqbles no matter how you connect them to the PC, or what kind of noise is or is not on the connection itself. The simple existence of the connection indicts it. BTW, the same is true of HDMI. But its not true of LAN connections or optical audio connections.
The crucial factor is the design of the USB device itself, whether or not it reject/filters noise adequately. The same is true of internal soundcards, whether built in or PCI/PCIe.
Again, thank you both.
I guess I am being spoon-fed basics here. I appreciate it that people take the time and trouble to do that. I guess I've said already that I'm trying to get real information rather than audiophile mythology: I hope many other seekers will find and benefit too.
Again, thank you both.
I guess I am being spoon-fed basics here. I appreciate it that people take the time and trouble to do that. I guess I've said already that I'm trying to get real information rather than audiophile mythology: I hope many other seekers will find and benefit too.
The important thing is that a top performing USB device has to hit on all cylinders of which there are at least two: Power
and signal.
One irony about my Fiio E5 story is that there is no digital processing within it. It is just a miniature analog power amp. Whatever noise that came from USB comes in via the power terminals. I don't think that shielding or the suual kinds of interference rejection had anything to do with it at all. The noise came in via a power supply ground wire. It was a grounding problem. This shows how USB power is a potential exposure all by itself.
I don't know if I made this clear but the E5 caused no problems with the first PC that it was installed on and ran cleanly for years as a booster for a number of different audio interfaces, both internal and external (USB). But it was problematical with a second PC as well as a portable Li Ion charger. This shows the importance of getting the power part of the equation right, which has nothing to do with shielding or the like.
It seems obvious to me that the E5 was not tested in all possible relevant configurations. Even though there is one spec for USB power, there are many possible implementations and they can make a difference.
The good news to me is that external USB power supplies are usually the fix and they are cheap and readily available. But, its not the cleanliness or absence of noise of the USB power supply that makes a difference. The difference is that the external power supply removes a potential ground loop that might be the fault of some other part of the system, in this case possibly the PC itself.