Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Playing audio, PC to speakers—best way? most sensitive parts of chain? (Read 11682 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Hi all,

I have a simple question, but have never been clear on the right answer!

Basically I want to play digital audio stored on a PC via loudspeakers. What is the best way to do this and where do the sensitivities lie in this decision?

Simple right?
But slightly vague, so let’s tighten the criteria.

  • I want to transmit from the PC digitally.
  • I could send the audio via HDMI or S/PDIF coax
  • I want the solution to handle video from the PC as well as audio.

Current equipment (although I do not want to limit this discussion to current equipment)
AVR (handling spif or hdmi or phono) to loudspeakers

OK,

Some assumptions:
  • I can use any lossless format
    As compressed or otherwise all lossless files create the same PCM stream. This is an example of an option which I beleive has no sensitivity to the outcome

  • I can ensure that the lossless file is converted into a bit-perfect PCM stream within the PC

  • If I park design standards for a moment, then digital data that is going to be converted into analogue can be asynchronous but there is limited capability for error correction in the transmission.
    So signals need to arrive at a DAC complete and in the right order, at or ahead of time, but not necessarily in sync between transmitter and receiver.

  • The data-rate can vary to support different sampling rates (eg 44.1, 48, 96, etc)


Some sub-questions:
  • I assume that PCM on S/PDIF is actually a regular signal (eg 44.1k) however the clock in the PC’s soundcard (built-in, PCI or USB) will not be synced to the DAC in the receiver (AVR or otherwise), so is there handshaking or a buffer?

  • If there is a buffer, which end is it and generally what order of magnitude is its size?

  • Given that most recording are made at above the final storage’s sampling rate (eg a 44.1k CD could be mastered at 192k) there should be no noise in the digital signal due to filtering or other effects, what really matters in the DAC?

  • Why would you not use a DAC in a mid-range AVR (of which many use TI/Burr-Brown DACs)


Hope that all makes sense!
icstm

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #1 –
If your goal is as you say to "I want to play digital audio stored on a PC via loudspeakers" then plug it into your AVR via HDMI or S/P-DIF. I'm not really sure why you need answers to all those questions? For example, I don't understand why you're asking the last one? No-one on here that I've noticed has ever said not to use the DAC in an AVR?

Obviously S/P-DIF doesn't send video so you'd need to make sure your video card can send over HDMI if that's what you want?

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #2 –
>I can use any lossless format
You can if your media player supports it.
All lossless formats will yield a bit identical result on decoding.

> I can ensure that the lossless file is converted into a bit-perfect PCM stream within the PC
You need a media player supporting a bit transparent driver like ASIO or WASAPI


>digital data that is going to be converted into analogue can be asynchronous but there is limited capability for error correction in the transmission

This is protocol depended.
SPDIF and HDMI are not asynchronous.
Adaptive mode USB is not asynchronous.
All these protocols slave the DAC to the sender.
Asynchronous USB is the only protocol that can be used to feed a DAC that is truly asynchronous.

All these protocols maintain a realtime stream.
This makes it impossible to do a retry in case of an error so no error correction


>The data-rate can vary to support different sampling rates
Yes but it is hardware depended. Both sender and receiver must be able to support these sample rates. Likewise 16/24 bit

>I assume that PCM on S/PDIF is actually a regular signal (eg 44.1k) however the clock in the PC’s soundcard (built-in, PCI or USB) will not be synced to the DAC in the receiver (AVR or otherwise), so is there handshaking or a buffer?
No, the DAC will be slaved to the clock in the PC driving the SPDIF
http://www.thewelltemperedcomputer.com/KB/SPDIF.html

>Given that most recording are made at above the final storage’s sampling rate
If you down sample to 44.1 there is by design no signal above 22.
Depending on the quality of the software used down sampling can create artifacts.
TheWellTemperedComputer.com

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #3 –
RoseVal Thanks for your answers.


Yes my players do support most lossless formats and yes I do use ASIO and WASAPI, but as I said I was assuming that once you choose any of these they do not impact the result.

OK, so I did not realise that the receiver slaves itself to the transmitter. If that is the case, what are the critical issues that can impact sound quality in such a system.

We seem to be saying that:
The timing to the DAC is determined by the PC
So the PC can impact sound quality even with a lossless bitperfect signal
I assume, as I did before that latency is not an issue, so I assume that the PC can buffer the PCM signal to ensure it can be sent with good timing without problem

What determines a good DAC?
I assume the main issues are post the conversion to analogue and that these issues are the usual ones that impact an amp (pre or power), such as clean power supply, suitable analogue filters, etc.
However are there any differences in the digital part of the amp (such as up/over sampling circuits, or the conversion itself)?

Thanks again,
icstm

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #4 –
As PCs runs a multi-tasking OS by design it is not guaranteed that data will be delivered at a specific moment. So all I/O is buffered.  In practice it works fine most of the time and is able to fill the buffer of an audio device in time.

As long as you use a synchronous protocol the quality of the timing is determined by the clock of the PC. You might compare SPDIF and HMDI as HMDI often has a very poor jitter performance.
http://thewelltemperedcomputer.com/HW/Conn...DMI_connect.htm

DACs often use PLLs/ ASRC to reduce the impact of input jitter.

In case of a DAC of course everything counts, the quality of the power, the filtering, the analog stage, etc, etc.
In practice you often find the same chip (by Wolfson, Burr-Brown, ESS) in all kind of products.
Modern well designed consumer grade products often do have a very good DA conversion.
TheWellTemperedComputer.com

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #5 –
>> Playing audio, PC to speakers—best way? most sensitive parts of chain?

This hobby named 'hi-fi music appreciation', it's a lifelong pursuit...

The most sensitive part of the chain is your own ears.
The most important part of the chain is owning fantastic speakers.
The signal quality coming out of the PC is a trivial issue if you have not spent many years developing those first two things to a high degree.

Obviously, the signal from the PC needs to go through an amplifier to power your speakers, and this too plays a big role in the sound quality coming out of the speakers.
For a long time, I just used a recent model Yamaha 5.1 Reciever and thought the sound was great ...until I met a really nice guy selling vintage Hi-Fi audio; he would allow me to take home any amplifier in his shop (for free, no strings attached, price of amp no problem!) and hook it into my system and hear for myself in my time, playing my music, in my environmental space.
After many months of 'borrowing' amps from his shop like a kid in a candy store, my awareness of "quality sound" coming out of the speakers was highly trained.
That Yamaha Receiver I had was total crap, but I didn't have good enough "aural" experience to make judgements. That Hi-Fi dealer and his altruism were an incredible blessing to me.

In summary, if your amp and speakers and ears are not cultivated to a high degree over a long period of time, the signal coming out of the PC is insignificant

Basic tweaking of the PC parameters is good enough, such as using foobar2000 with the 'Kernel Streaming' plugin and piping the signal out of the built-in line out to your amp.

Buy nice thick speaker wire. Buy nice cables with connectors made by Neutrik AG (as used by audio pro's). Ultra dear audiophile equivalents are a waste of money.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #6 –
Quote
I want to transmit from the PC digitally.


This is probably the best product on the market atm.

RME Fireface UCX

There are many less expensive alternatives.

However since you also think you want

Quote
I could send the audio via HDMI or S/PDIF coax
I want the solution to handle video from the PC as well as audio.


You could consider this instead.

MoTU HDX-SDI

MoTU have a good reputation for software on Apple platforms. Not so hot with Windows.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #7 –
Quote from: RonaldDumsfeld on
This is probably the best product on the market atm.

RME Fireface UCX

can you elaborate why this is a good solution? i simply cannot imagine why OP would need a 36-channel audio interface that has been optimized for ultra-low latency operation and includes things such a high-quality mic preamps, a DSP for mixing and EQing in hardware, etc.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #8 –
I did say he could get something cheaper.

If he didn't require HDMI or video handling an EM-U 0204 or similar @ ~£100  would be fine.

The OP seemed to be interested in what was the optimum quality for the job in hand. RME offers genuine pro quality, buller proof build quality, rugged construction, unrivaled ongoing product support (software support in particular) and instant residual resale value. It's what an audio professional would have for his personal use. The OP can can look at the specs and make a comparison with units offered by the hi fi distribution chain and frankly at ~<£1,000 the price compares pretty well with what he'd be offered for something with a nice box and gold plated sockets etc.

I had a broadly similar requirement to the OP and bought an M-Audio Fast Track Pro (~£150). It was fine but I almost immediately realised how useful the multichannel IO and DSP would be. So I got a 10 in  / 14 out unit and now wouldn't be without it. It replaced my old int. receiver. I plug all my old stuff in, record decks, headset, microphone, tape, cassette, CD/DVD telly, laptop, desktop, guitar, tuner etc and it feeds floorstanders, desktop monitors, headset, the spare set in the kitchen, all recorders etc. Instantly record anything , multiple programs simultaneously. If I knew what I knew now I would have saved a while longer and got an RME. probably the less expensive Fireface 400.

I don't know anything cheaper that deals would deal with vodeo but there may be.

Anyway it's a good way to do comparative shopping. Look at the best and see how it compares with what you want to spend.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #9 –
Quote from: RonaldDumsfeld on
I did say he could get something cheaper.

sure. i just wanted to provoke an explanation. tbh, i still don't understand what OP would get out of a Fireface UCX what he couldn't get out of, say, the Babyface. Not trying to argue the Babyface is OP's best bet, just noting that there are interfaces one-third the price of the UCX which provide the exact same functionality OP needs.

Quote
If he didn't require HDMI or video handling an EM-U 0204 or similar @ ~£100  would be fine.

yeah, the interfaces you mention don't provide "HDMI or video handling." they mereley provide coaxial SPDIF, something you can find in interfaces costing a small fraction of the price of the RME or Motu you suggested. please cmiiw.

Quote
The OP seemed to be interested in what was the optimum quality for the job in hand. RME offers genuine pro quality, buller proof build quality, rugged construction, unrivaled ongoing product support (software support in particular) and instant residual resale value. It's what an audio professional would have for his personal use.

the same could be said about the Babyface..

Quote
I had a broadly similar requirement to the OP and bought an M-Audio Fast Track Pro (~£150). It was fine but I almost immediately realised how useful the multichannel IO and DSP would be. So I got a 10 in  / 14 out unit and now wouldn't be without it. It replaced my old int. receiver. I plug all my old stuff in, record decks, headset, microphone, tape, cassette, CD/DVD telly, laptop, desktop, guitar, tuner etc

idk, OP mentioned he has an AVR which can probably handle most of that stuff.

EDIT: i guess the babyface merely provides optical SPDIF, not coax. but i don't think that affects the core of my argument.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #10 –
Quote
the interfaces you mention don't provide "HDMI or video handling."


Not to start a row or anything, simply as a point of information please have a look at the MoTU limk again.

From the first page., third paragraph

Quote
For the best-possible image quality, capture uncompressed video via HD-SDI or HDMI with 4:2:2 10-bit color depth and full-raster 1920x1080 HD resolution.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #11 –
I recommend using TCP/IP over ethernet or 802.11 to send a digital audio and video signal to a media box with upnp/DLNA rendering features, and then from the media box to my TV and AVR.  I find this offers higher quality in terms of signal transmission and can be cheaper and more flexible to implement than HDMI, USB or other methods.  Many use something like the Western Digiital TV Live box.
 

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #12 –
Quote from: icstm on
RoseVal Thanks for your answers.


Yes my players do support most lossless formats and yes I do use ASIO and WASAPI, but as I said I was assuming that once you choose any of these they do not impact the result.

OK, so I did not realise that the receiver slaves itself to the transmitter. If that is the case, what are the critical issues that can impact sound quality in such a system.

We seem to be saying that:
The timing to the DAC is determined by the PC
So the PC can impact sound quality even with a lossless bitperfect signal
I assume, as I did before that latency is not an issue, so I assume that the PC can buffer the PCM signal to ensure it can be sent with good timing without problem

What determines a good DAC?
I assume the main issues are post the conversion to analogue and that these issues are the usual ones that impact an amp (pre or power), such as clean power supply, suitable analogue filters, etc.
However are there any differences in the digital part of the amp (such as up/over sampling circuits, or the conversion itself)?

Thanks again,
icstm


You're asking a pretty broad question (lots of variables) but I'll try to help. 

You start with a digital file and you want a clean output signal; make sure your components don't degrade the pristine signal you started with.

PC    - as long as it doesn't lag it shouldn't degrade your signal
Software - If you're getting bit-exact output you haven't degraded the signal path.  Look for workflow that works for you.
ASIO - use it if you get some errors but don't bother with it unless you have to
DAC  - Some are cleaner, clearer, better than others--go out, research and listen (You really have to.  There are no shortcuts. Even if you wanted to throw a ton of money at a sound-system you still may end up with crud unless you listen hard and demo it.)
EQ/Soundboard - see DAC above
Cables              - see DAC above
Speakers          - see DAC above

I'm not trying to be an a&&.  You have to go out and listen.  If you can, take full advantage of the "30 day return policy", demo stuff and play with signal chain and then switch it out and do it again.  I've been building my sound-system for going on 12 years and I'm still tweaking bits and pieces daily/weekly/monthly to get just that little bit closer to perfection.  If you give up, send me $20,000 and I'll build you a system which will knock out half the windows in your neighborhood and sound so good the neighbors might not even complain.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #13 –
Quote from: godrick on
I recommend using TCP/IP over ethernet or 802.11 to send a digital audio and video signal to a media box with upnp/DLNA rendering features, and then from the media box to my TV and AVR.  I find this offers higher quality in terms of signal transmission and can be cheaper and more flexible to implement than HDMI, USB or other methods.  Many use something like the Western Digiital TV Live box.


It's a neat idea.  I've seen audio pushed over Ethernet (http://www.axiaaudio.com/mobility/) so I know it could work.  I'm more of a traditionalist though, I'd go USB/Firewire because I know it'll work and I hate headaches.  You wouldn't gain/lose sound quality either way (unless you're dropping packets).  If you have a long cable run, I'd experiment with Godrick's Ethernet suggestion. 

Godrick, have you seen audio dropouts under heavy network traffic or are you running signal over dedicated Ethernet?  I love the idea of 802.11. Has it run solid for you?

PS - FWIW I think MOTU/Fireface is overkill here.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #14 –
I'm a traditionalist, too, which is why I use ethernet - it's been around far longer than USB or firewire.  Even RTP and RTCP aspects (to better accomodate realtime audio and video over IP networks) were developed before USB and firewire.  I see nothing novel in using ethernet for media, and I can avoid much of the cost, time and complexity in using methods like USB to overcome limitations.  Dropped packets do not cause quality issues unless you mess with the automatic error correction and retransmit aspects, which I don't. 

I have no problems with dropouts that I notice.  I don't run a dedicated network.  No special configuration.  Combo of wired and wireless throughout the house.  It's almost embarrassing how simple and cheap it is to use ethernet instead of USB or other methods and I think is superior for its error correction, simplicity, cost, length of cable runs supported (if wired), and lack of need for cables (if you go wireless).

Once 802.11n came along the only issue in using ethernet for audio and video was waiting for consumer electronics companies to offer products that used it for media, and with the media streamers, not to mention other products, that time has come.  I don't make use of the applications that come with a media streamer for playing my content - I really just use it as a cheap and high quality network transport device and very good DAC (when I use the analog outputs of a WD TV Live box).

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #15 –
Quote from: TangoTiburones on
DAC  - Some are cleaner, clearer, better than others--go out, research and listen (You really have to.  There are no shortcuts. Even if you wanted to throw a ton of money at a sound-system you still may end up with crud unless you listen hard and demo it.)
EQ/Soundboard - see DAC above
Cables              - see DAC above
Speakers          - see DAC above


The typically tiny measured differences between DACs and cables can be heard and ranked by 'listening hard'?  What does that mean?

Quote
I'm not trying to be an a&&.  You have to go out and listen.


A method that's got some large inherent problems.  Hence TOS#8 here.

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #16 –
"most sensitive parts of chain?"

A. the speakers
B. their positioning
C. your room

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #17 –
Quote from: Roseval on
As PCs runs a multi-tasking OS by design it is not guaranteed that data will be delivered at a specific moment. So all I/O is buffered.  In practice it works fine most of the time and is able to fill the buffer of an audio device in time.

As long as you use a synchronous protocol the quality of the timing is determined by the clock of the PC. You might compare SPDIF and HMDI as HMDI often has a very poor jitter performance.
http://thewelltemperedcomputer.com/HW/Conn...DMI_connect.htm

DACs often use PLLs/ ASRC to reduce the impact of input jitter.

In case of a DAC of course everything counts, the quality of the power, the filtering, the analog stage, etc, etc.
In practice you often find the same chip (by Wolfson, Burr-Brown, ESS) in all kind of products.
Modern well designed consumer grade products often do have a very good DA conversion.
Ah yes, a PLL, cannot beleive I had forgotten about those, sorry.
That HDMI vs SPDIF article was interesting. What is driving that, has that changed over time by using different chips and controllers? I thought theoretically they should be the same, though reading the article, it appears that in SPDIF the Tx dictates the data rate, whereas in HDMI the Rx does - how weird!


Quote from: godrick on
I recommend using TCP/IP over ethernet or 802.11 to send a digital audio and video signal to a media box with upnp/DLNA rendering features, and then from the media box to my TV and AVR.  I find this offers higher quality in terms of signal transmission and can be cheaper and more flexible to implement than HDMI, USB or other methods.  Many use something like the Western Digiital TV Live box.
That is quite interesting. When you say a MediaBox, do you mean a WDTV or ACRyan, PopCorn, etc type (ie the sort that run on a Sigma or RealTek micro controller)? So H/W decoding of the codecs?
The issue there I would have thought is:
  • Limited codec playback (relise on firmware updates
  • Limited customisation of skins (less important) and layout (more important) - though I know BRad or someone use to offer custom ROMs for the old boxes
  • More problems with some encoded files (eg too many P-Frames or other problems)
However I think the fact it is a relativly cheap solution and is fanless and idiot proof for others.

Quote from: mzil on
"most sensitive parts of chain?"

A. the speakers
B. their positioning
C. your room

Completely agree and there is a D, where you sit in the room
I take the best seat in the house and place it in the best spot 

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #18 –
Quote from: derty2 on
>> Playing audio, PC to speakers—best way? most sensitive parts of chain?

This hobby named 'hi-fi music appreciation', it's a lifelong pursuit...
Agreed!

Quote from: derty2 on
The most sensitive part of the chain is your own ears.
The most important part of the chain is owning fantastic speakers.
The signal quality coming out of the PC is a trivial issue if you have not spent many years developing those first two things to a high degree.

...until I met a really nice guy selling vintage Hi-Fi audio; he would allow me to take home any amplifier in his shop (for free, no strings attached, price of amp no problem!) and hook it into my system and hear for myself in my time, playing my music, in my environmental space.
After many months of 'borrowing' amps from his shop like a kid in a candy store, my awareness of "quality sound" coming out of the speakers was highly trained.
That Yamaha Receiver I had was total crap, but I didn't have good enough "aural" experience to make judgements. That Hi-Fi dealer and his altruism were an incredible blessing to me.
Indeed you were very lucky, I hope you bought from him in the end.
You kind of wish there were more people like that

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #19 –
Quote from: godrick on
I recommend using TCP/IP over ethernet or 802.11 to send a digital audio and video signal to a media box with upnp/DLNA rendering features, and then from the media box to my TV and AVR.  I find this offers higher quality in terms of signal transmission and can be cheaper and more flexible to implement than HDMI, USB or other methods.  Many use something like the Western Digiital TV Live box.


This product doesn't even do gapless playback

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #20 –
1) I never claimed perfection in what I use
2) Lack of gapless (or difficulty in getting it so) is the biggest issue, and really the only issue, I have with using upnp/DLNA renderers
3) I am by no means an expert or even up-to-speed on efforts to overcome the gapless issue, but this appears to be something that may be achievable at some point via the control point per
http://yabb.jriver.com/interact/index.php?topic=68934.0
http://www.hydrogenaudio.org/forums/index....showtopic=84665

Playing audio, PC to speakers—best way? most sensitive parts of chain?

Reply #21 –
Although I haven't tried it yet, it seems there is a solution for gapless using foo_upnp per http://www.hydrogenaudio.org/forums/index....showtopic=85443

Since WD TV Live now seems to support up to 96k and 24 bits, this should work for me, once I try the Playback Stream Capture setting in foo_upnp.