BlackCatSound said:
Because I believe it is not that easy to improve the sound as it might look on the first glance.
Still - I think they are quite some people around using the
Lynx AES16 as reference digital PCI card for SPDIF output
feeding external DACs. They are all more than happy with it.
I am wondering if the gain would be worth the effort though.
Look at the cost e.g. for a USB-DDDAC. What is it? A tenth of above?
That's all you need to get a far more than decent sound, which most probably comes very close to above solutions!
I am really happy with it.
I'll first fix the easy wins on the PC, before starting up
the hardware battle, if still necessary.
Cheers.
But if you look at it:
CPU -> PCI -> USB -> I2C -> DAC
or
CPU -> PCI -> I2C -> DAC
Less stages in the way.
You've also got less software stages in the way as you don't need to go via a USB driver. The soundcard driver talks direct to the PCI bus.
It just seems to me the audiophile way is always to add MORE to the equation rather than take parts out.
CPU -> PCI -> USB -> I2C -> DAC
or
CPU -> PCI -> I2C -> DAC
Less stages in the way.
You've also got less software stages in the way as you don't need to go via a USB driver. The soundcard driver talks direct to the PCI bus.
It just seems to me the audiophile way is always to add MORE to the equation rather than take parts out.
soundcheck said:
I am pleased that you are happy. However, the quality of audio using a Lynx AES16 with top quality dac and fanless computer rivals high end hifi and seems to have that fluidity not easilly achievable with transports.
You USB thingie is limited to low resolution ausio whereas the Lynx will do 24/192. And yes, it is worthwhile going hirez - no more digititis!
BlackCatSound said:
Don't forget to consider several stages on the soundcard.
Usually you have receivers, transmitters, filters, buffers, DSP, SRC on the sound card, all potentially adding some kind of rubbish the stream.
If that in the end sounds better - Who knows. Until know I havn't seen any comparison. I think it is not possible, with the limited facts available, to come to a conclusion here.
fmak said:
I think we shouldn't go into discussion around upsampling, hirez,
oversampling, non-oversampling. There are at least 100 threads
in the forum discussing that.
If you look for NONOS DACs threads here around, you'll find that the little USB thingies are very much able to rival so called high-end hirez DACs costing ten times more.
That a PC as source under certain conditions is able to rival any high-end setup, I am well aware of. And this is why I started this thread, to find out more about best setups and conditions for using a PC a source.
soundcheck said:
And you don't think you have that with your USB DAC?
Have a look at M-audio PCI cards, they are just PCI -> buffer -> DAC.
No transmitters or receivers as you've got with your USB interface.
About the lowest latency you will get. They do DMA transfers direct to the DAC buffer so lower interrupt overheads and less software involvement compared to USB. ie all the things you seem to be worried about.
HW setup and config
As it has been discussed earlier , some devices you
are running causing noise and potentially do cause other problems.
E.g. all devices that are not needed will consume power and generate heat. If they are not needed just switch them off.
I just would like to share some HW-setup tips:
First of all I did set-up a dedicated audio hardware-profile.
While booting you can choose if you boot your stripped down audio profile or you boot your standard setup.
This you can do under system control/hardware/hardware profiles
What I am doing is switching off all devices not needed during
playback:
These could be:
Internal sound card
Network card
Modem
Pcmcia
CDROM
WLAN
Infrared
Even other USB devices could be switched off.
While playing back I let automatically switch off the screen after one minute.
One more config tip:
Set your sw-player process priority to "lowest".
Taskmanager-Processes- right mouse click on the corresponding player process. You can also try the realtime setting!
Feedback appreciated!
As it has been discussed earlier , some devices you
are running causing noise and potentially do cause other problems.
E.g. all devices that are not needed will consume power and generate heat. If they are not needed just switch them off.
I just would like to share some HW-setup tips:
First of all I did set-up a dedicated audio hardware-profile.
While booting you can choose if you boot your stripped down audio profile or you boot your standard setup.
This you can do under system control/hardware/hardware profiles
What I am doing is switching off all devices not needed during
playback:
These could be:
Internal sound card
Network card
Modem
Pcmcia
CDROM
WLAN
Infrared
Even other USB devices could be switched off.
While playing back I let automatically switch off the screen after one minute.
One more config tip:
Set your sw-player process priority to "lowest".
Taskmanager-Processes- right mouse click on the corresponding player process. You can also try the realtime setting!
Feedback appreciated!
fmak said:
Dont' take me wrong. I just wanted to say, that quite some people wouldn't regard upsampling and bit depth enhancement as an advantage, especially when looking at ringing filters etc.
I just tried to avoid this misleading discussion.
Ask Lavry about 192kHz? He'll come up with 96 is the way to go.
Ask Digital Audio Denmark they are telling you 384khz is the only way to get lossless recordings? I guess we'd end up nowhere.
Anyhow your point is taken.
To me, and I think I made that clear in the beginning, the adressable areas, from my perspective, related to my setup,
of improvements starting earlier than the soundcard.
And if this is the case because my DAC could potentially be better designed. OK. Myself and a lot more USB-DAC owners have to live with it.
By the way: DDDACs are running at 44,1kHz/16bit. There is no conversion to 48kHz ongoing.
Hi all
I was just thinking these days about a complete (and perfect ) PC source.I am intersting in DVD playback and some HD video content so the things are getting a little complicated.
One of my concerns is how to make a noise free powersuply. I am not sure that a pc powersuply is up to the task (just think how carefuly designed supllies you'll find in hi-end preamps, transports, amplifiers...etc.) So how about building a transformer based powersuply and not a switching one?.
Another problem would be the need of a pci audio card vs. usb extraction of the digital signal. I tend to favour the usb because I am planing a DAC with ad1896 for synchronous sample rate conversion and ad1955 or wm8740.
I think the lynx AES16 would be great for a further use of it for digital crossovers, but until then i find the price to much without using all the 16 channels available. If you think that a pci card would be better for 24/196 than using a sample rate convertor, then what cards would you sugest?
Oh, and by the way.. I find the tip with special bios setup for audio great!
Thanks
I was just thinking these days about a complete (and perfect
) PC source.I am intersting in DVD playback and some HD video content so the things are getting a little complicated.One of my concerns is how to make a noise free powersuply. I am not sure that a pc powersuply is up to the task (just think how carefuly designed supllies you'll find in hi-end preamps, transports, amplifiers...etc.) So how about building a transformer based powersuply and not a switching one?.
Another problem would be the need of a pci audio card vs. usb extraction of the digital signal. I tend to favour the usb because I am planing a DAC with ad1896 for synchronous sample rate conversion and ad1955 or wm8740.
I think the lynx AES16 would be great for a further use of it for digital crossovers, but until then i find the price to much without using all the 16 channels available. If you think that a pci card would be better for 24/196 than using a sample rate convertor, then what cards would you sugest?
Oh, and by the way.. I find the tip with special bios setup for audio great!
Thanks
The PSU is a relatively small source of noise in a PC. All the stuff connected to it will make more noise.
Personally I'd go for PCI rather than USB. The USB sound implementation is such a nasty hack IMO.
I have an M-audio audiophile192 PSI card and the output is SILENT. You can't tell where the volume on my external amp is set if there is no source playing. There is no hiss, no video noise, no HDD noise. A far cry from creatives offerings.
Personally I'd go for PCI rather than USB. The USB sound implementation is such a nasty hack IMO.
I have an M-audio audiophile192 PSI card and the output is SILENT. You can't tell where the volume on my external amp is set if there is no source playing. There is no hiss, no video noise, no HDD noise. A far cry from creatives offerings.
I am very intersted in using a last generation cpu (intel core 2, vista ready). So the psu will have a fan for sure.
Maybe I am blinded by the exotic designs in the audio industry, but I realy think the psu is a very important component regarding the reproduction of sound. That is another reason I'd like to try a transformer psu with diy regulators. I wonder if it is possible, because I am not sure how to design the control circuit wich conects to the motherboard (unlike a simple AT psu, the ATX is controled by the motherboard).
Do you use the onboard dac of the m-audio, or an external dac feeded by the digital output of the card? I know m-audio uses AKM dac's but I'd like the flexability of a diy dac. How about using an ASCR?
How the signal is afected using the usb?
Maybe I am blinded by the exotic designs in the audio industry, but I realy think the psu is a very important component regarding the reproduction of sound. That is another reason I'd like to try a transformer psu with diy regulators. I wonder if it is possible, because I am not sure how to design the control circuit wich conects to the motherboard (unlike a simple AT psu, the ATX is controled by the motherboard).
Do you use the onboard dac of the m-audio, or an external dac feeded by the digital output of the card? I know m-audio uses AKM dac's but I'd like the flexability of a diy dac. How about using an ASCR?
How the signal is afected using the usb?
The PSU is just turned on and off by the mobo. You won't get enough amps using a transformer and linear regs.
Either way its pointless, it will still pick up noise from all the digital signals.
If you really wanted to do it you should just run the analogue stage of the sound card from a regulated supply.
I use the DAC on the sound card.
The method used for clock sync with these USB sound devices is just so badly wrong. USB was never intended to do such things.
Either way its pointless, it will still pick up noise from all the digital signals.
If you really wanted to do it you should just run the analogue stage of the sound card from a regulated supply.
I use the DAC on the sound card.
The method used for clock sync with these USB sound devices is just so badly wrong. USB was never intended to do such things.
Soundcheck -- thanks for starting this thread.
Here's my setup:
Shuttle XPC (modded for quietness, that was a huge pain)
2 external hard drives (400GB + 160GB = 1,800 CDs as APE or FLAC)
Ripping using Exact Audio Copy .
Playback using J River Media Center v12 beta, Asio4all output.
(though soundwise Foobar with Secret Rabbit Code upsampling to 48khz is slightly better).
Volume controlled in player (not preamp). Have a Microsoft Media Center remote (with replacement driver) controlling JRMC + using HIP (free alternative to Girder).
USB cable - nothing fancy, but well shielded.
DDDac MkII, USB + 2 towers, upgraded caps (Elna Silmic II's for the 100uf, Vishay 1837 for the 100nf).
.... amps + speakers.
In terms of sound it easily beats the best CD player I've ever owned (a Marantz CD17). The DDDac is also far less fatiguing than the Benchmark DAC1 and matches it on detail. But that could be because I was feeding the DAC1 with SPDIF from a modded transit, rather than I2S via USB, which may have been the weak link in my previous setup.
The biggest challenge, I'd say, is getting a silent PC -- my system's noise floor is so low now that I can hear hard drive seek and some whirring in quiet music passages...
Not to mention that setting up a system takes time -- ripping and tagging takes a long time. Cost wise, you also need to factor in hard drive back up space.
On my "to do" list are:
1. get a really silent PC, or a cable extender system so I can put the PC in another room.
2. get an optical usb extender.
3. try out the usb-audio driver you recommend.
4. get a Tent XO clock for the DDDac USB board (ran out of stock when I ordered).
5. upgrade the DDDac Rload resistors to Rikens/Tantalums...
and so on......
Here's my setup:
Shuttle XPC (modded for quietness, that was a huge pain)
2 external hard drives (400GB + 160GB = 1,800 CDs as APE or FLAC)
Ripping using Exact Audio Copy .
Playback using J River Media Center v12 beta, Asio4all output.
(though soundwise Foobar with Secret Rabbit Code upsampling to 48khz is slightly better).
Volume controlled in player (not preamp). Have a Microsoft Media Center remote (with replacement driver) controlling JRMC + using HIP (free alternative to Girder).
USB cable - nothing fancy, but well shielded.
DDDac MkII, USB + 2 towers, upgraded caps (Elna Silmic II's for the 100uf, Vishay 1837 for the 100nf).
.... amps + speakers.
In terms of sound it easily beats the best CD player I've ever owned (a Marantz CD17). The DDDac is also far less fatiguing than the Benchmark DAC1 and matches it on detail. But that could be because I was feeding the DAC1 with SPDIF from a modded transit, rather than I2S via USB, which may have been the weak link in my previous setup.
The biggest challenge, I'd say, is getting a silent PC -- my system's noise floor is so low now that I can hear hard drive seek and some whirring in quiet music passages...
Not to mention that setting up a system takes time -- ripping and tagging takes a long time. Cost wise, you also need to factor in hard drive back up space.
On my "to do" list are:
1. get a really silent PC, or a cable extender system so I can put the PC in another room.
2. get an optical usb extender.
3. try out the usb-audio driver you recommend.
4. get a Tent XO clock for the DDDac USB board (ran out of stock when I ordered).
5. upgrade the DDDac Rload resistors to Rikens/Tantalums...
and so on......
soundcheck said:
In JRMC I don't like the upsampling so much. But in Foobar I find the Secret Rabbit Code gives a bit more 'air'... odd that, as I started my quest for PC Audio with the obsession for 'bit perfect'....
I would strongly suggest ditching .wav and moving to FLAC or Monkey's Audio. Sound is the same, but you gain more hard drive space and tagging.
soundcheck said:
I'm personally going to stick with XP for the time being because...
- I'm just finally getting comfortable with it and setup so it's rock-solid stable.
- will wait until Vista SP1 comes out before I even think of switching
soundcheck said:
Can;t help there, I dabbled a few years ago then gave up. Though I just downloaded Kubuntu and will give it another go...
Curious about Ramdisk -- never tried it -- but sounds terribly inconvenient having to drag files to another folder before playing them....
I have JRMC setup so all I see is album thumbnails, very easy.
Hi folks.
I was surprised that the thread is coming back to live again.
I think that's great.
To comment your posts above:
Some words about PC Audio hardware:
1. To avoid noise and power instabilities due to bad onboard power regulators you better choose highest quality main boards.
2. Your power-supply needs to be over dimensioned. 500W should do well.
It is pretty much the same as it is with audio-equipment. It is always better to
have some headroom..
Otherwise stick to battery powered notebooks, what I do for the time being.
3. Passive cooling throughout the system is of course the best choice.
4. When it comes to harddrives -- better go for professional grade over
dimensioned HD. There are some 500GB WD which are highly recommened
And don't fill them up too much!
5. Onboard graphic cards are prefered over PCI graphic cards.
Some words about audio interfaces.
I think it has been clear and proved by now, that USB is not performing worse than SPDIF. There are quite some threads in the WEB discussing and proving
that.
Keyissues, IMO, to dissuss are:
1. Clocking
2. Samplerates and SRC
3. Volume control
1. a. Clocking at its best is a synchrounous Master/Slave clocking scenario.
Your DAC runs a precision master-clock, which is clocking the PC and the
application as slave. You need special interface cards such as Lynx AES 16
or an RME PCI card to be able to do that.
I havn't seen any DIY projects yet addressing this.
b. My prefered (best price/performance ratio) solution is the USB interface, based
on the PCM2707. It is a 16bit/48kHz solution though. Which is really
a bottle-neck. I don't know why Burr Brown is not comming up with a
24/32bit USB/I2S interface chip.
I stick to my solution for the time being because I think the overall chain
matches perfectly.
c. Clocking within the PC is working better with 48kHz than 44,1
( most of the soundcards do internal upsampling to 48kHz -
delivering unacceptable performance though.)
What I did. I got myself the Voxengo R5Brain pro tool and upsampled my files
to 48kHz.
The sound improvement was clearly audible. Much more space, better pinpointing, more realistic stage.
2. SRC
I am not happy with realtime SRC. I checked the JRMC upsampling against
my offline upsampling. JRMC is not at all able to beat the offline conversion.
HW-SRC higher than 48kHz will heavily depend on the DAC and the related
SRC you use. I don't feel the drive right now to run after higher sample rates.
3. Volume control
I recently removed my S&B TX102 and my DAC coupling caps and connected
the DAC directly to the AMP input.
Unfortunately I do have a 16 bit DAC. Still, if running at high volumes, the
lost resolution is limited to a minimum.
The tansients and details increased so much that I can easily accept the small
disadvantages, loosing 1 or 2 bits resolution.
Beside the JRMC-DSP is calculating volume conversion with 32bit/float. That
is giving you already better results than e.g. AFAIK foobar volume control.
And beside that you don't have to use the windows-mixer for volume control.
My newest PC tweaks:
1.
I shutdown most of the services under system control/administration/services.
You can configure this for your specifc Audio HW profile.
My task manager opens up with only half the processes listed earlier.
2. Turn off your monitor, while playing back
Comment to "Playback from RAM":
This is a must. As mentioned earlier - Foobar offers full file buffering. There
is no need for RAMdisk.
If you use Samplitude for play back you can also configure full-file playback.
Comment to SW-player and operating systems:
OS Windows:
Samplitude 8.11 demo is my reference single-track player for the time being.
Multitrack playback award goes still to JRMC.
OS Unix.
I managed to setup Unix for my PCM2707 and it works great. I can confirm the earlier postings about UNIX, you might have read.
Once you have it running, it outperforms any Windows setup I listened to.
You can have a realtime system, delivereing latencies below 1ms
(remember Windows can do 12ms max.)
Just listening to the simple XMMS player over ALSA in realtime mode is better than
anything else when it comes to pure PCM playback.
Unix gives you a lot of options and it is all free! And it runs well on older and low profile PCs.
For now my first choice.
Highly recommended distributions:
Dyne:Bolic 2.3 www.dynebolic.org (or pure:dyne)
64studio www.64studio.com
As a start Ubuntu 6.10 works also fine. For people with less Unix background
this is most probably the best entry. Even ubuntu with XMMS and ALSA will
be better than any MS setup, when it comes to PCM playback.
That'll be it for now.
Cheers
\Klaus
I was surprised that the thread is coming back to live again.
I think that's great.
To comment your posts above:
Some words about PC Audio hardware:
1. To avoid noise and power instabilities due to bad onboard power regulators you better choose highest quality main boards.
2. Your power-supply needs to be over dimensioned. 500W should do well.
It is pretty much the same as it is with audio-equipment. It is always better to
have some headroom..
Otherwise stick to battery powered notebooks, what I do for the time being.
3. Passive cooling throughout the system is of course the best choice.
4. When it comes to harddrives -- better go for professional grade over
dimensioned HD. There are some 500GB WD which are highly recommened
And don't fill them up too much!
5. Onboard graphic cards are prefered over PCI graphic cards.
Some words about audio interfaces.
I think it has been clear and proved by now, that USB is not performing worse than SPDIF. There are quite some threads in the WEB discussing and proving
that.
Keyissues, IMO, to dissuss are:
1. Clocking
2. Samplerates and SRC
3. Volume control
1. a. Clocking at its best is a synchrounous Master/Slave clocking scenario.
Your DAC runs a precision master-clock, which is clocking the PC and the
application as slave. You need special interface cards such as Lynx AES 16
or an RME PCI card to be able to do that.
I havn't seen any DIY projects yet addressing this.
b. My prefered (best price/performance ratio) solution is the USB interface, based
on the PCM2707. It is a 16bit/48kHz solution though. Which is really
a bottle-neck. I don't know why Burr Brown is not comming up with a
24/32bit USB/I2S interface chip.
I stick to my solution for the time being because I think the overall chain
matches perfectly.
c. Clocking within the PC is working better with 48kHz than 44,1
( most of the soundcards do internal upsampling to 48kHz -
delivering unacceptable performance though.)
What I did. I got myself the Voxengo R5Brain pro tool and upsampled my files
to 48kHz.
The sound improvement was clearly audible. Much more space, better pinpointing, more realistic stage.
2. SRC
I am not happy with realtime SRC. I checked the JRMC upsampling against
my offline upsampling. JRMC is not at all able to beat the offline conversion.
HW-SRC higher than 48kHz will heavily depend on the DAC and the related
SRC you use. I don't feel the drive right now to run after higher sample rates.
3. Volume control
I recently removed my S&B TX102 and my DAC coupling caps and connected
the DAC directly to the AMP input.
Unfortunately I do have a 16 bit DAC. Still, if running at high volumes, the
lost resolution is limited to a minimum.
The tansients and details increased so much that I can easily accept the small
disadvantages, loosing 1 or 2 bits resolution.
Beside the JRMC-DSP is calculating volume conversion with 32bit/float. That
is giving you already better results than e.g. AFAIK foobar volume control.
And beside that you don't have to use the windows-mixer for volume control.
My newest PC tweaks:
1.
I shutdown most of the services under system control/administration/services.
You can configure this for your specifc Audio HW profile.
My task manager opens up with only half the processes listed earlier.
2. Turn off your monitor, while playing back
Comment to "Playback from RAM":
This is a must. As mentioned earlier - Foobar offers full file buffering. There
is no need for RAMdisk.
If you use Samplitude for play back you can also configure full-file playback.
Comment to SW-player and operating systems:
OS Windows:
Samplitude 8.11 demo is my reference single-track player for the time being.
Multitrack playback award goes still to JRMC.
OS Unix.
I managed to setup Unix for my PCM2707 and it works great. I can confirm the earlier postings about UNIX, you might have read.
Once you have it running, it outperforms any Windows setup I listened to.
You can have a realtime system, delivereing latencies below 1ms
(remember Windows can do 12ms max.)
Just listening to the simple XMMS player over ALSA in realtime mode is better than
anything else when it comes to pure PCM playback.
Unix gives you a lot of options and it is all free! And it runs well on older and low profile PCs.
For now my first choice.
Highly recommended distributions:
Dyne:Bolic 2.3 www.dynebolic.org (or pure:dyne)
64studio www.64studio.com
As a start Ubuntu 6.10 works also fine. For people with less Unix background
this is most probably the best entry. Even ubuntu with XMMS and ALSA will
be better than any MS setup, when it comes to PCM playback.
That'll be it for now.
Cheers
\Klaus
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.