Nixie said:
ASIO provides low latency and control of i/o synch. It's benefit isn't so much what it does, but what it doesn't do -- pass the sound through a bad software mixer designed for games rather than DAW use. ASIO is a way to bypass Microsoft kmixer, etc.
On RME cards, I suspect that the WDM emulated sound driver is bypassing any/all Windows mixing.
Cards from RME (HDSP9632?) or Lynx are probably every bit as good as cheap external DAC's. Is anyone using a sound card direct into their amp and not having problems with ground loop type issues?
Member
Joined 2003
I never had problems with my M-Audio Audiophile 2496. Though I did cut the ground connection between the card and the chassis, which may have lowered the noise floor, but I wouldn't know as the noise floor was already unheard at all listening volumes. This card was plugged into my Kookaburra preamp, then from the preamp into my Mauro LM3886 amplifier. I now use an external DAC to improve sound quality, nothing to do with noise.Daveis said:
Member
Joined 2003
The M-Audio card has a transformer at the digital I/O dongle, but I can't tell visually if it is for the digital input or the output. Maybe I should get out my multimeter and check.
Just to contribute something to the topic, in Windows I strictly use Foobar, it's great in so many ways. In Linux (my primary OS), I haven't heard any audible differences between players so I just use XMMS, and mplayer for quiet tracks. Mplayer has a volume normalizer plugin that works great for quiet tracks, while XMMS's volume normalizer is horrible; I could never get it to work right without blaring real loud music in my ears every once in a while. I would like to try out Aqualung on Linux, as it is geared towards audio fidelity.
Just to contribute something to the topic, in Windows I strictly use Foobar, it's great in so many ways. In Linux (my primary OS), I haven't heard any audible differences between players so I just use XMMS, and mplayer for quiet tracks. Mplayer has a volume normalizer plugin that works great for quiet tracks, while XMMS's volume normalizer is horrible; I could never get it to work right without blaring real loud music in my ears every once in a while. I would like to try out Aqualung on Linux, as it is geared towards audio fidelity.
I'm also using kxdrivers; mainly for tri-amping xover functions. I use optical outs to avoid all the nasties from the PC. I had jumped on the optical band wagon before even thinking about digital coax and transformers; though I doubt it would even make a difference to me.
This is all on a HTPC running Media Center, so of course WMP is used as my player. I have tried a few other players before and could not convince myself I heard anything different, let alone "better".
This is all on a HTPC running Media Center, so of course WMP is used as my player. I have tried a few other players before and could not convince myself I heard anything different, let alone "better".
aqualung
DecibeL
I tried the windows version of aqualung and it seems that you always have to have a sample rate converter loaded in the dsp part of the chain. The best quality converter is rated at 97db snr which is lower than a number of soundcards. There may be a way of turning off src, but I havnt been able to identify any way of doing so.
The linux version may of course be different, and it has the advantage of being able to load ladspa plugins into the dsp (might be a nice crossover plugin somewhere). Please report back if you try this player.
DecibeL
I tried the windows version of aqualung and it seems that you always have to have a sample rate converter loaded in the dsp part of the chain. The best quality converter is rated at 97db snr which is lower than a number of soundcards. There may be a way of turning off src, but I havnt been able to identify any way of doing so.
The linux version may of course be different, and it has the advantage of being able to load ladspa plugins into the dsp (might be a nice crossover plugin somewhere). Please report back if you try this player.
Daveis: " ... Is anyone using a sound card direct into their amp and not having problems with ground loop type issues? ... "
DcibeL: " ... I never had problems with my M-Audio Audiophile 2496. Though I did cut the ground connection between the card and the chassis, which may have lowered the noise floor, but I wouldn't know as the noise floor was already unheard at all listening volumes. This card was plugged into my Kookaburra preamp, then from the preamp into my Mauro LM3886 amplifier. ..."
The M-Audio 24bit/96k card was designed specifically to reduce or eliminate ground loops ... and its 24bit DAC has a very good noise floor to begin with.
There are several serious issues with PC plug in cards from other makes that do have difficult noise issues and more than a few do have ground loop potential. The modern M-Audio plug in cards and modern Apple Mac LossLess optical output being noteworthy exceptions, as a general rule one should avoid the PCI / PCMCIA and other plug-in sound cards ... external USB and FireWire devices almost always have provisions like linear voltage regulators to isolate the dirty, noisy PC power supplies from the audio stream.
As far as the playback software goes, they all work equally as well (as per Nixie, et al), exceptions being failure to interface with the better drivers (ASIO, etc.) and lack of plug-ins.
PC based audio quality or lack thereof is all primarily hardware related, software not really being an issue. 24bit DACs rule, dudes ... live with it.
DcibeL: " ... I never had problems with my M-Audio Audiophile 2496. Though I did cut the ground connection between the card and the chassis, which may have lowered the noise floor, but I wouldn't know as the noise floor was already unheard at all listening volumes. This card was plugged into my Kookaburra preamp, then from the preamp into my Mauro LM3886 amplifier. ..."
The M-Audio 24bit/96k card was designed specifically to reduce or eliminate ground loops ... and its 24bit DAC has a very good noise floor to begin with.
There are several serious issues with PC plug in cards from other makes that do have difficult noise issues and more than a few do have ground loop potential. The modern M-Audio plug in cards and modern Apple Mac LossLess optical output being noteworthy exceptions, as a general rule one should avoid the PCI / PCMCIA and other plug-in sound cards ... external USB and FireWire devices almost always have provisions like linear voltage regulators to isolate the dirty, noisy PC power supplies from the audio stream.
As far as the playback software goes, they all work equally as well (as per Nixie, et al), exceptions being failure to interface with the better drivers (ASIO, etc.) and lack of plug-ins.
PC based audio quality or lack thereof is all primarily hardware related, software not really being an issue. 24bit DACs rule, dudes ... live with it.
PCI
This thread seems to have morphed into something very similar to the other pc source thread.
Would a well implemented pci card not generally best a usb or firewire dac? Specifically, since current solutions for both of these introduce just as much jitter as spdif (albeit different jitter) Also, in a lot (even all?) of cases usb/firwire ports on a pc use the pci bus internally - use of a pci card which buffers audio directly into the onboard dac cuts out the added stage of further buses and protocols which should make for a simplified audio path. Improved shielding, EMI reduction and improved supplies for soundcards may be a useful approach to DIY in this area. Blackcatsound put this well in the other thread.
Of course if the dac chip preferred is not available as a soundcard you must build something externally and use spdif, i2S, usb or firewire and deal with the implied problems as best you can (or live with them).
Apart from buffer sizes and audio interfaces (asio, ALSA etc.) I also can't see why one player would differ from another which has the same features. I guess code wise no player is likely to implement the same feature in exactly the same way as another. But if there is an effect upon sound, this must surely be close to inaudible unless the code is put together very badly or algorithims are chosen to reduce cpu overhead rather than perform optimally?
I can imagine that if carrying out upsampling etc. then there would be significant benefits in doing this offline using a dedicated studio grade algorithm rather than in realtime.
This thread seems to have morphed into something very similar to the other pc source thread.
Would a well implemented pci card not generally best a usb or firewire dac? Specifically, since current solutions for both of these introduce just as much jitter as spdif (albeit different jitter) Also, in a lot (even all?) of cases usb/firwire ports on a pc use the pci bus internally - use of a pci card which buffers audio directly into the onboard dac cuts out the added stage of further buses and protocols which should make for a simplified audio path. Improved shielding, EMI reduction and improved supplies for soundcards may be a useful approach to DIY in this area. Blackcatsound put this well in the other thread.
Of course if the dac chip preferred is not available as a soundcard you must build something externally and use spdif, i2S, usb or firewire and deal with the implied problems as best you can (or live with them).
Apart from buffer sizes and audio interfaces (asio, ALSA etc.) I also can't see why one player would differ from another which has the same features. I guess code wise no player is likely to implement the same feature in exactly the same way as another. But if there is an effect upon sound, this must surely be close to inaudible unless the code is put together very badly or algorithims are chosen to reduce cpu overhead rather than perform optimally?
I can imagine that if carrying out upsampling etc. then there would be significant benefits in doing this offline using a dedicated studio grade algorithm rather than in realtime.
I'm sorry if this was already covered, I didn't read all the posts. I thought I would throw in my few cents. I find that some of the biggest problems with computers as sources is fan noise and other self noise, and poor quality power supplies. Computers use poorly regulated switching power supplies that are very noisy. There are companies that make much better quality power supplies, even some places that make custom hand built power supplies, however they are all switching supplies. I feel that linear supplies sound better on analogue cuircits, so one idea I have had is to use an external DAC so that its analogue cuircits can be operated by a seperate linear supply. Also, it would allow better regulation to the dac. Also, I want to modify the power supply in my PC source, to reduce noise and improve regulation. Looking into this idea, I started just redesigning the power supply, and after a while, thought it might be best to simply make a totally new power supply from scratch. Reducing fan noise is a simple task of using low noise fans or no fans, and using something like liquid cooling or large heat sinks and plenty of ventilation.
Essentially, I agree that the limitations of PC audio is in the PC itself, not the software. By all rights a computer has the ability to play back music better than any cd or dvd player can. Because it uses buffering and ram, its capable of much less error, and the potential for lower jitter as well. Much larger and more varied files can be supported, so the computer can change with the times, taking on new formats. You can use a far less compressed files, for better sound that CD or DVD. You can also process the files using software DSP, similar to the now popular software DSP being used in turnkey players, but you can customize it yourself, or update it with new software more easily.
Essentially, I agree that the limitations of PC audio is in the PC itself, not the software. By all rights a computer has the ability to play back music better than any cd or dvd player can. Because it uses buffering and ram, its capable of much less error, and the potential for lower jitter as well. Much larger and more varied files can be supported, so the computer can change with the times, taking on new formats. You can use a far less compressed files, for better sound that CD or DVD. You can also process the files using software DSP, similar to the now popular software DSP being used in turnkey players, but you can customize it yourself, or update it with new software more easily.
In My Opinion, backed by a very significant number of professional studio engineers, is that the FireWire connected DACs are the best = least amount of jitter, more channels of 24bit/96K or 24bit/192K, faster & better than USB or EtherNet connected devices or any of the plug-in cards.
Commercial/professional examples:
http://www.m-audio.com/products/en_us/FirewireAudiophile-main.html ... minimum acceptable, home or studio.
http://www.m-audio.com/products/en_us/FireWire410-main.html ... studio device
http://www.m-audio.com/products/en_us/ProFireLightbridge-main.html ... serious studio device
( http://www.m-audio.com/products/en_us/Transit-main.html ... USB playback only = as good as Apple's LossLess via Apple's I2S optical ports = optically isolates PC power supplies from your audio system)
http://rolandus.com/products/productdetails.aspx?ObjectId=731&ParentId=114 ... studio device
http://rolandus.com/products/productdetails.aspx?ObjectId=702&ParentId=114 ... studio device
http://tascam.com/Products/fw1804.html ... studio device
http://www.digitalaudio.dk/ax24.htm ... very serious mix master recording studio device (this is how the best European DVD-A and SACD recordings are made).
There are some very nice, high quality USB DACs, but they are playback only * often fail to make the translation to 24bit/96K ...
(A Mercinary Announcement: Get FireWire = this works famously ... http://industrialcomponent.com/firewirestuff/fws46603.html = for the price of the cable you get two plugin FW cards, one for Windows or Mac, one for Linux = no drivers required, compatible with ASIO ... from my company's web site. We are also M-Audio distributors.)
Commercial/professional examples:
http://www.m-audio.com/products/en_us/FirewireAudiophile-main.html ... minimum acceptable, home or studio.
http://www.m-audio.com/products/en_us/FireWire410-main.html ... studio device
http://www.m-audio.com/products/en_us/ProFireLightbridge-main.html ... serious studio device
( http://www.m-audio.com/products/en_us/Transit-main.html ... USB playback only = as good as Apple's LossLess via Apple's I2S optical ports = optically isolates PC power supplies from your audio system)
http://rolandus.com/products/productdetails.aspx?ObjectId=731&ParentId=114 ... studio device
http://rolandus.com/products/productdetails.aspx?ObjectId=702&ParentId=114 ... studio device
http://tascam.com/Products/fw1804.html ... studio device
http://www.digitalaudio.dk/ax24.htm ... very serious mix master recording studio device (this is how the best European DVD-A and SACD recordings are made).
There are some very nice, high quality USB DACs, but they are playback only * often fail to make the translation to 24bit/96K ...
(A Mercinary Announcement: Get FireWire = this works famously ... http://industrialcomponent.com/firewirestuff/fws46603.html = for the price of the cable you get two plugin FW cards, one for Windows or Mac, one for Linux = no drivers required, compatible with ASIO ... from my company's web site. We are also M-Audio distributors.)
pjpoes: " ... I agree that the limitations of PC audio is in the PC itself, not the software. By all rights a computer has the ability to play back music better than any cd or dvd player can. Because it uses buffering and ram, its capable of much less error, and the potential for lower jitter as well. Much larger and more varied files can be supported, so the computer can change with the times, taking on new formats. ..."
You still have to get the digital information out of the computer and far away from that noisy, dirty PC power supply ... or optically isolate it with a decent external interface (few plug-in sound cards have that optical port ) ...
The pro devices all have linear power supplies (above links = as you note is almost an absolute requirement). There are ways to optically isolate your USB DAC. but they can cost as much as the USB DAC itself ( http://industrialcomponent.com/opticis/opm210003.html ).
My advise = don't settle for anything less than multi channel 24bit/96K ... and if you plan on doing a decent job of ripping music from various sources, your interface should be bi-directional = a FireWire connected device 'cause USB, et al, can't do it.
You still have to get the digital information out of the computer and far away from that noisy, dirty PC power supply ... or optically isolate it with a decent external interface (few plug-in sound cards have that optical port
) ...The pro devices all have linear power supplies (above links = as you note is almost an absolute requirement). There are ways to optically isolate your USB DAC. but they can cost as much as the USB DAC itself ( http://industrialcomponent.com/opticis/opm210003.html ).
My advise = don't settle for anything less than multi channel 24bit/96K ... and if you plan on doing a decent job of ripping music from various sources, your interface should be bi-directional = a FireWire connected device 'cause USB, et al, can't do it.
I don't do windoz (and don't care much for it), but while discussing the use of iTunes on the PeeCee for playback, i was told by the programmer guy that if a player interfaces with the built-in Windoz audio routines (as iTunes for PC does) the signal gets screwed up (and i understand Vista is going to be WAY worse) so it is necessary to have a player that uses the better drivers (i'll assume ASIO is one of them).
On the Mac CoreAudio is well implemented & iTunes is a no-brainer for playback.
I've used foobar on a PC and it has a UI from hell (but i guess that describes many windoz based pieces of sw)
dave
FastEddy said:You still have to get the digital information out of the computer and far away from that noisy, dirty PC power supply ... or optically isolate it with a decent external interface (few plug-in sound cards have that optical port
The pro devices all have linear power supplies (above links = as you note is almost an absolute requirement). There are ways to optically isolate your USB DAC. but they can cost as much as the USB DAC itself ( http://industrialcomponent.com/opticis/opm210003.html ).
But why on earth would you want to use an interface that was never designed for constant rate data like audio and has been fudged to try and sync the external device to the PCs internal clock?
The way they have implemented USB DACs is IMO the most horrible way possible. The best way to isolate it is to disconnect it and throw it away.
Have you ever done any tests to see how noisy a PC PSU actually is?
" ... But why on earth would you want to use an interface that was never designed for constant rate data like audio and has been fudged to try and sync the external device to the PCs internal clock? ... The way they have implemented USB DACs is IMO the most horrible way possible. The best way to isolate it is to disconnect it and throw it away. ..."
Yes, "fudged" syncronization = inconsistant data transfer rates, that is a good point as well. USB sucks for bulk file transfers = too much handshaking 'tween data digestors & data translators (= jitter), really just single duplex, etc. ... direct bus connect plug-in cards should have an advantage here, but 'cause of PC PS noise, etc., they don't get the job done very well. Exceptions to this are slim and none.
The big advantage for FireWire is simply that it has ... a) four wires instead of two (double duplex = totally bi-directional data & control) ... and b) is peer to peer = all FW devices have equal "weight"/value/master&slave relationship = always in sync. = no fudging allowed.
Yes, "fudged" syncronization = inconsistant data transfer rates, that is a good point as well. USB sucks for bulk file transfers = too much handshaking 'tween data digestors & data translators (= jitter), really just single duplex, etc. ... direct bus connect plug-in cards should have an advantage here, but 'cause of PC PS noise, etc., they don't get the job done very well. Exceptions to this are slim and none.
The big advantage for FireWire is simply that it has ... a) four wires instead of two (double duplex = totally bi-directional data & control) ... and b) is peer to peer = all FW devices have equal "weight"/value/master&slave relationship = always in sync. = no fudging allowed.
FastEddy said:all FW devices have equal "weight"/value/master&slave relationship = always in sync. = no fudging allowed.
Thats got nothing to do with DAC clock sync though. Firewire was designed for digital video which does not have any susceptibility to jitter.
But the big question is are PS PSUs actually as noisy as you think? I'd have to say no they are not.
" ... Have you ever done any tests to see how noisy a PC PSU actually is? ..."
Nah, never bothered, not when you can actually hear it. A fully loaded PC often has very obvious noise coming right out of the sound card and out through the other ports ... the fix is to throw the sound cards away and get something external for audio. Those switching supplies used to give network integrators fits = talk about handshaking & data transfer retries & packet collisions ... quite often remeddied by upgrading the SM Power Supplies to something with a lot more power ... that's the real raeson those clunky old network servers had extra large, 500 watt suppies ... more power = more headroom above the switching noise floor. (Novell Certified #92 talkin' here ... That's also why God gave us silver wire in the better EtherNet cables = better signal to noise ratio & noise floor + better conductivity & impedence matching = less PS noise down stream.)
Nah, never bothered, not when you can actually hear it. A fully loaded PC often has very obvious noise coming right out of the sound card and out through the other ports ... the fix is to throw the sound cards away and get something external for audio. Those switching supplies used to give network integrators fits = talk about handshaking & data transfer retries & packet collisions ... quite often remeddied by upgrading the SM Power Supplies to something with a lot more power ... that's the real raeson those clunky old network servers had extra large, 500 watt suppies ... more power = more headroom above the switching noise floor. (Novell Certified #92 talkin' here ... That's also why God gave us silver wire in the better EtherNet cables = better signal to noise ratio & noise floor + better conductivity & impedence matching = less PS noise down stream.)
One of the worst sound devices for PC crossover I've used was an MAudio Firewire Solo. I heard audible clicks/dropouts when using firewire that I dont have with my PCI RME card.
Now, the fault doesnt necessarily lie with MAudio. On Thuneau's site many have had problems with Windows and firewire audio. Many are attributing the problem to poor generic Windows firewire drivers.
Also, I've had some audible problems with some motherboards, AC97 sound cards, and RAID controllers under windows.
Now, the fault doesnt necessarily lie with MAudio. On Thuneau's site many have had problems with Windows and firewire audio. Many are attributing the problem to poor generic Windows firewire drivers.
Also, I've had some audible problems with some motherboards, AC97 sound cards, and RAID controllers under windows.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.