Greetings! Having just upgraded to i7 and having made redundant my Audigy 2 ZS (I now use the onboard motherboard soundchip), I'm going to experiment with another method of audio from my PC. I'll use this thread to post my findings and welcome others' input. There are many uPnP digital streaming devices (media renderers) on the market, some of which are open source and will stream audio or "internet radio" from just about any "broadcast" over ethernet. I'm fortunate in that I have access to many of these products as I work in the hi-fi field. My plan is to ditch all onboard sound processing, and send the audio from my PC to my digital streamer over ethernet. If successful I know the audio performance will destroy what is currently available, regardless of whether I'm shooting zombies or listening to music. To date I've cracked the music issue by ripping my music to FLAC, installing a media server application (Asset uPnp or TwonkyMedia), and playing back via a Linn DS (other media renderers are available - Logitec Transporter, Soundbridge, Naim HDX, Meridian Sooloos, etc...). I can also play Spotify and LastFM through the PC/Linn by broadcasting these as an "internet radio" station, and having the Linn receive, decode and playback that stream. Has anyone else used this method for their music and media, and how would you recommend getting audio from games etc... available over ethernet hence accessible to the Linn or other media renderer? Many thanks, Cleggy.
Excuse my ignorance but how is this different from using a SPDIF connection with regards to gaming in particular?
When using SPDIF from the PC, you're using either the onboard soundchip or soundcard to handle the initial audio processing. D/A conversion and analogue processing is then managed in your receiver/offboard DAC. This method has already "tainted" the quality of the output to a certain extent. Taking the audio at an earlier stage would allow me to bypass the poor quality elements of my PC, and output in a raw data form over ethernet to be managed solely by my digital streamer and hifi. When I listen to music currently, my PC sends data in the form of FLAC or whatever into my network, to my digital streamer, which then processes that stream for the best possible performance. Listening to ripped CD's in this manner sounds far better than playing said CD's on a CD player - far less data loss at the transport/reading stage, and far less data loss through the DA/upsampling/filtering stage due to the absence of machanical and electronic noise. Maybe what I'm looking for isn't possible but I wanted to explore it anyway...
erm excuse my ignorance as well, but wouldn't the digital stream be unmolested before being sent over SPDIF ? I thought the audio stream is sent in digital form to the SPDIF out to whatever equipment to have to decode and convert it
Yeah, almost, though the more I think about this the less I reckon my theory is possible... As it stands, data output by the SPDIF (as a digital bitstream) still comes from somewhere, in the case of games it probably comes from a bunch of samples held within the game's files in WAV/FLAC or MP3. I want to intercept the data in that form, before it's turned into a digital bitstream by the computer, and hand all processing over to a dedicated digital streamer... However I don't think that's possible at the moment... More likely I'll end up buying the best quality soundcard I can afford (recommendations please - I hear Auzentec and M something are good) and taking the SPDIF into a receiver. Yes I am anal, I know...
but sound files on the pc are already digital? I dont think you should buy an expensive soundcard to just do digital out, the expensive part is about the DAC units not the digital side.
yes, because an audio file stored on a computer is already a digital bitstream isn't it? Its just a long list of 1s and 0s effectively.
Believe it or not there's a huge difference in quality from SPDIF outputs, even in high-end hifi hardware. With regards the data, you're right that the audio data is held digitally on the PC - it's held in a "file" format (WAV/FLAC/MP3/whatever), before being decoded into a bitstream (a string of binary 0's and 1's). It's the decoding process and the components the data traverses that massively influence the quality of the bitstream outcome, hence ultimately the quality of the sound that reaches your lugholes once it's been through the DAC and analogue stage. Anyways, I reckon I'll buy a decent card and test it against my Audigy and onboard sound. Will post my findings when done. Cheers for the input folks, Cleggy.
I'm a computer science graduate not an audio engineer so I may be way off the mark but I don't see how the bitstream is 'degraded' in anyway by the on-board components when using SPDIF. Unpacking/decoding a file format into its relative bitstream is granted - a fairly involved process, but ensuring that the '1s and 0s' end up in the right order during transit is as simple as avoiding basic data corruption. This is something that computers have been capable of dealing with years (cyclic redundancy checks, for example). IIRC the most likely stage of causing any damage to a digital stream is the 'jitter' that a poor transport mechanism (read: bad cd player) may introduce. I use foobar2000 with the WASAPI plugin (which basically ensures a bit-perfect output via SPDIF and grants foobar2000 exclusive access to the audio card). A Beresford DAC then handles things from there. From what I can gather my digital stream is arriving at the DAC 'untainted'. As I say though, its not my area of expertise and I'd love to understnad this better
exactly how I was picturing it, digital is digital it will either work or it wont, unlike analogue where you might hear distortions. Its the same story with a £5 HDMI cable and a Monster HDMI cable that costs £50, your paying for nothing.
Every conversion from digital to analog and back again in an audio system results in a loss of quality, or at least some different colouration of the audio. It's arguable as to whether the conversion of digital to analogue in a DAC has much of an effect on the audio (I mean, it has to happen somewhere along the line, so it hardly matters), but when it goes the other direction there's an inevitable change. For every situation where digital audio data is converted to analogue by a DAC, sent to another digital device and converted back to digital by an ADC, there's going to be a loss of quality, depending on the quality of the DACs/ADCs. This is due to the ADC producing an approximation of an analogue signal by reading values at sample-points along an incoming audio wave and outputting those values digitally. - WhiskeyAlpha, the loss of quality doesn't occur anywhere in the digital values, but in the accuracy of those digital values in representing the analogue sample they were taken from. A simple and somewhat-accurate analogy would be if you imagine plotting graph-points along a sine wave, then redrawing that sine wave by joining the dots with straight lines (or stair-steps to be truthful). Now imagine that same sine wave, but between two of your points on the wave there's a little dip in amplitude or some other fluctuation that your points don't take into account, thus that information is lost when the wave is redrawn later from the points you noted. The same holds true in converting analogue audio to digital values and goes to explain why a higher bitrate makes for higher quality (you're taking more points along the wave, thus creating a more accurate image of the wave when you redraw it), but also explains why a conversion from digital to analogue and vice versa results in a loss of quality. Due to limitations of processing power and memory, you can't have an infinite number of digital points along the curve of that wave. Here's an image to illustrate the plotting of those points: Another analogy is if you zoom in too far in a photo and begin to see pixels. From a distance it looks like a smooth, cohesive analog image, but up-close you see the quantization of the digital format in pixellation. [Cue a CSI meme image involving zooming & enhancing]
I'm not too sure I can agree wholeheartedly with you there, as SPDIF connections aren't present in older stereo recievers, which ironically(and interestingly) do provide much better sound quality with just analog RCA than a modern integrated run-of-the-mill $800 receiver with SPDIF. Although that might have to do with expensive speakers and the old reciever being a vaccum-tube.
@Zurechial: Thanks for that, without sounding ungrateful I'm quite familiar with the processes you describe. Though its nice to have an in depth description in the thread The point I was really trying to make is most closely related to this: The OP was discussing the diminishing effect on audio quality that occurs when data is moved in its digital format (from decoding to SPDIF output). My argument was that this data in its raw form has already been converted into its digital form (from its analogue source) and that computers have had methods to ensure basic data integrity for many years that are proven and work extremely well. I don't see 1s suddenly becoming 0s or bits dissapearing, this simply doesn't happen in computing, unless intentional or some corruption/fault is occuring. The protocols are presumably in place to avoid such issues? It was this point alone that I was disputing/discussing. I just don't see how digital signals can be 'coloured' in the same way analogue sine waves are simply by moving through a series of chips/circuits (putting aside additional post-processing of course, I'm talking pure untampered digital bit-streaming). To summarise my opinion: when bit-streaming and not adding additional processing via the sound card or the OS/application, every SPDIF should output the same pattern of 1s and 0s from a given source.
Not entirely true. Getting your digital source from your HDMI output to your display's HDMI input will still require a cable. 1s and 0s are delivered onto said cable using high and low voltages accordingly. A bad, unshielded cable over a distance can still therefore degrade the bitstream and being a 'realtime' process there is no error correction provided (unlike a TCP ethernet connection for example). When the signal arrives degraded, you will either get breakup or dropout, or the display will employ some technical wizardry to try and correct/guess it (interpolation). The breakup/dropout or interpolation may be invisible to many people (if of course it's not severe). The very same theory applies to DVI and SPDIF connections BTW. Many people assume that if the image arrives okay then the £5 cable must be as good as a £80 QED job. I certainly think that a good quality digital cable is important but I certainly wouldn't go excessive at all. I think many of the problems related to digital cable interference are grossly over-exaggerated and I've seen no evidence to suggest that it is as important as analogue cables were to an A/V system. The thing to remember with digital signals is that you simply want to ensure that your bit-stream arrives with all the 1s and 0s as they were sent. With no error correction in place, you can only use the best cable you are prepared to shell out for (be it £5 or £500) and hope for the best. With analogue signals, the impact that every element has on the sine wave carried will 'colour' the picture/sound in some way. Clearly, the cable is crucial in this regard. Good article here
If your entire signal path is digital then this should indeed be the case. For all intents and purposes a totally-digital signal path will maintain the full quality of the original digital source file, but if there's a conversion from digital to analogue (and back again) anywhere in that signal path other than at the speakers, then there will be a loss of quality. Presumably an SPDIF output should provide an unmolested digital signal, as has been said, but it seems to me like the OP is doubting the integrity of the digital signal from SPDIF and wants to tap the audio stream earlier in the DSP path. Aside from some strange and unlikely situation where the soundcard does a DAC/ADC conversion itself somewhere in the signal path, the signal should already be a 1:1 copy of the audio output at the software level and that very same digital signal should be what comes out of the SPDIF output.
WhiskyAlpha and Zurechial i applaud you for expressing what i was trying to in such an in depth manner. rep to both of you. I too believe that the conversion was already done when the disc or audio file was created so a bit stream should not be affected by passing through SPDIF.
That's the point I'm trying to make The OP states that even hi-end Hi-Fi components have very different sounding output via SPDIF. The only sensible reason for this must be additional processing carried out by the device, or the inherent mistakes of a poor transport mechanism. Take a CD player for example. A CD player can lift the raw audio bit-stream directly from the CD and output it via its SPDIF (How well it achieves this is really down to the transport mechanism, as no error correction is in place). For arguments sake lets assume the transport mechanism is perfect and reads every disc exactly as it was pressed. If another CD player does the same tasks (again, with a perfect mechanism) and the resulting bit-stream is different to the first, then it can only be down to some internal processing. It's one of the main reasons I actually prefer to listen to FLAC over a CD. I can recreate the very same bit-stream without the concern of a transport mechanism introducing flaws. In fact, I stream my audio from a fileserver to all the machines in my house, and the TCP protocol ensures that nothing goes astray along the way. EDIT: Sorry Baboon, wasn't disregarding your post, just that I got distracted halfway through replying
Cheers folks, some very good discussion here. I've used a number of digital streamers for listening to music stored on my PC, with interesting differences in performance albeit from the same sound files on the hard disc. I currently use a Linn Akurate DS for music. This takes an ethernet feed from my PC (via the network - the Linn is just a network device), and turns the data (FLAC, WAV, AIFF, ALAC, MP3 or whatever) into music (file into bitstream, bitstream into analogue) before passing into my preamp. It sounds wonderful, much better than the same CD played on a reference CD player, due to a couple of facets: Firstly, when a CD is played through a CD player, that CD player has to read, decode and output in real time. Regardless of the quality of transport and electronics, some data will be lost in the process since it's a "read once" scenario. If I rip the disc to FLAC using good software on my PC, I know that the file that ends up on my hard drive is an accurate rendition of what was burnt onto the CD at the pressing plant, due to the ripping software reading the disc multiple times until it's happy. Secondly, when this FLAC file is then played via a streamer (the Akurate DS in my case), the playback "mechanism" has a very easy life. It is not subject to the mechanical, electronic and thermal interferance present inside a CD player since there are no moving parts - my music is protected by TCPIP as it moves along the network, and is then processed by authentic grade components rather than the stock components present in a PC or soundcard. The outcome is great - musically without peer. However, I want to use this "system" to play games through my DS too - imagine Bioshock, L4D and Call of Duty played through such a system. Uberdrool! Thoughts welcome... Cleggy. *EDIT* WhiskeyAlpha you are spot on, though regrettably no such perfect transport currently exists and never can. What is interesting that digital audio reproduction is now surpassing analogue, albeit with studio-master resolution (24-bit, 192khz). Hearing music at this resolution is life-changing!
Your best bet sounds like a card with Dolby Live or DTS Connect. It's not technically bitstream (it's transcoding to DD/DTS after all), but there's no ADC-DAC process between the source files and the receiver. I don't think there are any comparable solutions outputting LPCM, but it's a moot point: First, lets assume your receiver can decode the unusual audio filetypes used in games (i.e. generally not mp3, occasionally ogg). If you sent the raw files to your receiver for decoding, your games would sound pretty crap. This is because of all the processing that goes on for sound source positioning, gunshot echoes, spatialisation when underwater, and the like. Without this processing, games would sound very weird. Essentially there is no bitstream for you to bitstream, the actual audio 'output' the card would otherwise run through a DAC to your amp has already undergone processing. As long as you avoid that internal DAC (by using SPDIF, for example) you're getting the closest to the perfect output as is technically possible.
