frezog
frezog's picture
Offline
Last seen: Never ago
Joined: Jul 2 2007 - 8:02am
digital is digital?
RGibran
RGibran's picture
Offline
Last seen: 2 years 4 months ago
Joined: Oct 11 2005 - 5:50pm

Perhaps if you told us what you are trying to accomplish we could narrow our recomendations. Are you building a desktop system for your office, feeding computer speakers, or are you wanting to feed your main audio system, which may be 20 feet away, ect. Gonna listen to iTunes downloads? Burn your entire music collection to lossy or lossless?

I believe there may be a few of us here who have been through this and have some soundcards we would just give you.

RG

Jim Tavegia
Jim Tavegia's picture
Offline
Last seen: Never ago
Joined: Sep 1 2005 - 4:27pm

Rgibran,

LOL What do you think of the Frezog question? The answer is slightly shorter than the Warren Commission Report! I know you can help him once the goal is set.

frezog
frezog's picture
Offline
Last seen: Never ago
Joined: Jul 2 2007 - 8:02am

I see you need more info. I am currently running spidf out from the onboard audio (RealTek HD audio) of my mb. In the past I used the spidf out from an Audigy1 soundcard. The digital signal is sent to my main audio/HT system which is about 25 feet away. The digital ouput is input to an Outlaw Audio 950 preamp/processor. The computer is integrated into my system in that I have a three monitor setup. 2 monitors on the desktop and the third monitor being my 52" dlp TV.

I dont use my cd or dvd player any more since all media is played through the computer. (DVDs are usually ripped to the hard drive and played from there)
My music collection consists of mostly .flac files but I also run .wav, .ape and mp3 files. As a music player/library manager I have been using MediaMonkey and have recently been playing around with foobar2000. With this setup I have no need for any type of a streaming/server solution.

I would like to be able to know that the weak point is my audio system, and that I am not degrading the audio/digital signal before it reaches the audio system.

Jim Tavegia
Jim Tavegia's picture
Offline
Last seen: Never ago
Joined: Sep 1 2005 - 4:27pm

I am not familiar with RealTek, but from their website and what info you have given us it sounds like you are feeding HT 5-6-7.1 to your home theater set up? You might find out from the MFG what the 2-channel jitter spec is for the digital out.

The AES/EBU pro digital format is preferred, I am not sure if next in quality is firewire or SPDIF, then optical and USB. BenchMark claims to have solved the USB issues with their new 2 channel DAC/Headphone amp. Check out the review on the Phile website.

Rgibran may know about your sound card with his extensive knowledge. The brand of HT receiver/processor would also help. You may be heading into Dr. Rubinson territory.

RealTek

Regards,

RGibran
RGibran's picture
Offline
Last seen: 2 years 4 months ago
Joined: Oct 11 2005 - 5:50pm

Frezog,

Thanks for the background. I see you have already achieved your objective and are in a questioning mode regarding your two channel audio for whatever reason. I will try to give you the short answers, because as Jim stated, the long answer is waaaaay too long.

You asked: I would like to be able to know that the weak point is my audio system, and that I am not degrading the audio/digital signal before it reaches the audio system.

NO! Can you improve it? Probably, at least on paper. Will you hear it, in your system? I doubt it.

Internal and external soundcards have their flaws. Media players have their flaws. All of these connection schemes have their flaws, which I suppose is why you are here posing these questions, apparently having caught wind of them in an article or two on the related subject. Jim has done a good job establishing their preferred order with the exception of omitting

Editor
Editor's picture
Offline
Last seen: 13 years 3 months ago
Joined: Sep 1 2005 - 8:56am


Quote:
I am not familiar with RealTek, but from their website and what info you have given us it sounds like you are feeding HT 5-6-7.1 to your home theater set up? You might find out from the MFG what the 2-channel jitter spec is for the digital out.

Which computer is the poster using? Earlier PCs with the AC97 codec, would only pass 20-bit audio from the onboard audio. You needed a separate soundcard using the ASIO protocol to get a 24-bit data output. The ability to handle 24-bit audio was built into the Mac's CoreAudio system from the start, however.

John Atkinson
Editor, Stereophile

frezog
frezog's picture
Offline
Last seen: Never ago
Joined: Jul 2 2007 - 8:02am

Thanks everyone for the responses. I guess most of my questions are theoretical because I do have a system now that I am happy with. I guess what I am trying to get info that can help with future systems. I would rate my present audio system maybe at the high end of "mid fi" but probably at the low end of "hi-fi". I'm kind of a cheap s.o.b. meaning that don't like spending money on something that I could of found cheaper, or for the same money got something better. I probably have about 4 thousand dollars invested in a system which has been put together over the past 20 years. I could see that all changing and putting 10 to 20 thousand dollars into a system over the next 10 years.

My main question is (again, assuming that all internal DACs are bypassed for an external DAC or processor). If a 24bit digital signal is coming from a computer can that be any better or worse than a signal supplied from a multi thousand dollar CD player. My Ideal system would have the computer (I usually build my own) as the only source component, if it can deliver the signal necessary for true audiophile quality.

I know I'm coming from a naive place here but it still seems to me that it shouldn't be a hard thing to get a [pure, raw,
unadulterated, series of 1s and 0s, (is there a term that can be used here)] digital signal out of the computer and to the audio system.

Elk
Elk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 26 2006 - 6:32am

A computer hard drive can send a high quality digital signal out as well as good CD transport.

The provisos: The CD has to be ripped accurately onto the hard drive in the first place. Also, the soundcard needs to be capable of sending the digital signal without truncating the word length and without adding any artifacts of its own.

There is oodles of info on this topic in more specialized forums. I'm sure you will find it with a few Google attempts. It's fun reading.

fred333
fred333's picture
Offline
Last seen: Never ago
Joined: Sep 17 2007 - 6:22am

I looked around the net and found some great stuff. I have enjoyed finding music online then putting them on my ipod. My son turned me onto this it is great. Such better quality.

59mga
59mga's picture
Offline
Last seen: 5 years 3 months ago
Joined: Jun 21 2006 - 6:52am

Fred,

If you would, please advise us on what you found and where you found it..inquiring minds want to know. Thanks.

Mike

ps: The wife-beater t-shirt look...that's how I relax and listen to my music.

struts
struts's picture
Offline
Last seen: 2 years 9 months ago
Joined: Feb 1 2007 - 12:02pm

frezog,

While I see you have already received some helpful advice here (and maybe you were really seeking practical advice more than theoretical enlightenment) I don't think anybody has really answered your questions. Let me have a go.

I would caution that engineering science is mixed with opinion pretty freely on these boards and it can sometimes be hard to distinguish between the two, especially for non-engineers. So I will offer the usual caveat: however persuasively anybody puts their case here always let your ears be the final judge, to the greatest degree possible before parting with your hard earned!


Quote:
Reading the other posts here makes be believe this is all a lot more complex than I had been aware of.


So true, so very true!


Quote:
1. digital is digital?
--- Are there different "versions" of "spdif out"?
--- I always asumed spdif out was the same no matter what the device outputting it. Are there factors that effect the "quality" of the digital signal?


Digital is digital until (and here's the rub) you come to convert it back to analogue. Suddenly you have to start worrying not just about the encoded audio samples "the bits" but the time-base or "clock" which the S/PDIF protocol actually embeds into the datastream using Biphase Mark Code. Whereas the datastream itself is quite robust and the probability of amplitude errors is not high, the timebase information is incredibly fragile and all too easily distorted by a whole host of real world factors that aren't apparent if you take an abstracted, purely digital view. These range from electrical effects caused by temperature changes or physical vibration, electromagnetic noise from power supplies, RF pollution from clocks and high frequency circuits and busses, ground-borne noise and even things like signal reflections caused by impedance mismatches in connectors. Distortions in the wordclock, called "jitter", degrade the sound quality since the right bit at the wrong time is the wrong bit. This paper contains a much more detailed explanation of interface-related jitter and the susceptibility of different DAC architectures to it. The bad news is that almost every one of these sources of distortion is present in abundance in a computer chassis. Professional and consumer audio products address them to varying degrees depending on factors like the sound quality ambition level, the skill of the designer and cost constraints.

Yes, there are different "versions" of S/PDIF. In short the two main variants are 75 ohm coax with RCA or BNC connectors and fiber with "TOSLINK" connectors. You will often see the opinion expressed that "Coax is superior to TOSLINK because it has less jitter". In my view this is a bit of an oversimplification. It may be true in theory at the boundary case, but the two versions have different theoretical strengths and weaknesses. My own opinion after much listening and reading is that, "statement" products aside, the implementation is far more important than the version in determining sound quality. To a first approximation a good implementation of either will likely outperform a poor implementation of the other.

The subject of jitter, "hearing it", understanding it, measuring it, modelling it and removing it is one of the areas where audio engineering research is currently focussed. See this thread for more detail and some recent comments from one of the pre-eminent researchers in the field.


Quote:
2. If there is a sound card that can output a "proper" digital signal, does the playback software have an effect on the digital signal?
--- Is foobar any better or worse than winamp j river etc.
--- What would be the most inexpensive sound card available if the only concern is the quality of spidf out.


No, if the codec is either uncompressed or compressed losslessly and the media player is purely decoding the file, i.e. we're assuming that it is not being used to intentionally alter the bitstream, for instance applying digital volume control or DSP, then the resulting bitstream should be exactly the same as the original and the choice of player should not affect the sound quality. There are of course many other differences between players such as breadth of codec support, output mode support (e.g. ASIO or Kernel Streaming - needed to bypass a part of the Windows O/S called kmixer which can upsample your data without you asking it to), metadata/tagging support, user friendliness etc. However since they work exclusively in the digital domain, and not at the point of conversion back to analogue, sound quality should NOT be one of them.

The sound card that is generating the S/PDIF bitstream definitely can affect the sound quality however. In theory it can do so quite considerably, although one would hope that in practice with reasonably well engineered products the actual variance would be much smaller.

Your question about which sound card to go for is a great one since most quality as opposed to feature comparisons focus on the analogue rather than the digital output. The Creative "Sound Blaster Audigy Advanced MB" card built into my Dell PC performs non-defeatable upsampling to 48kS/s so I bought an M-Audio "Audiophile 2496" card (US$129.95 list) which I chose mainly for its ability to handle high-resolution (24 bit 96kS/s) PCM datastreams. I should point out that "Audiophile" is the name of the card rather than any value judgement of mine on its performance! It works fine feeding my Grace m902 DAC via coax but there may well be better and cheaper cards out there that I'm not aware of.


Quote:
3. Assuming the computer's digital output is ideal, and internal DACs are bypassed on all devices, doesnt that mean the computer is as good as any other output device available?
-- Why spend thousands (tens of thousands) of dollars on a cd player if the computer can do the same thing? (can the computer do the same thing)


If the assumption were valid then the conclusion would be too. As I said, there are many reasons why it isn't which I tried to summarize above. You may also like to check out this seminal thread on the topic.

Your second question is truly the $64,000 one and opinion on this is deeply divided. I for one suspect that vested interests (among consumers as well as producers) are conspiring to cloud the issue. See this thread for my own attempt to kick-start this very debate.

However, I suspect that the newly announced Linn Klimax DS (computer audio for

ROLO46
ROLO46's picture
Offline
Last seen: Never ago
Joined: Sep 29 2006 - 10:16am

I concur with Mr Atkinson

Use a Mac

Music is better

I-Tunes is native

Garage Band is native for minor tweaks

Roll over and think of America.....

struts
struts's picture
Offline
Last seen: 2 years 9 months ago
Joined: Feb 1 2007 - 12:02pm


Quote:
If a 24bit digital signal is coming from a computer can that be any better or worse than a signal supplied from a multi thousand dollar CD player.

Missed that one! A 24-bit signal from a computer should ceteris paribus always be better than the signal supplied by any CD player, regardless of price, since the latter are confined to 16 bits at 44.1kS/s.

Unfortunately the bit-depth is only a small part of the equation and 24-bit encoding certainly doesn't guarantee SOTA sound quality, let alone the artistic merit of the performance. It does however remove one theoretical constraint on what is achievable.

Compare the 16/44.1 and 24/88.2 versions of the Dunedin Consort's outstanding reading of H

Elk
Elk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 26 2006 - 6:32am

Neither Mac's or PC's have an advantage for music reproduction. Both have their downfalls and need to be utilized correctly to get the best out of the them.

Specifically as to PC's, the AC'97 CODEC is an old (1997, hence the name) 20-bit architecture Intel designed for integrated audio in PC's. AC'97 supports 20-bit/96 audio. Of course all PC's, but for the most basic, have for years come with a separate soundboard - most with 24 bit capability.

In 2004, AC'97 was replaced with Intel's high Definition Audio. HD Audio supports up to 32-bit/192 stereo and up to eight channels of 32/96 audio.

If 32/192 isn't sufficient for your needs I really want to hear your system.

ROLO46
ROLO46's picture
Offline
Last seen: Never ago
Joined: Sep 29 2006 - 10:16am

Dual Core I-Macs run silent

Operate at 24bits

Have optical outs

I-tunes is free and good music management

Music runs better on a Mac

My 'Sound on Sound' mate who does home visits to fix project studios will not venture into any recording PCs

My Hifi mates at AVI also dislike PCs for music application.

Roger
IMHO

Elk
Elk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 26 2006 - 6:32am

Both Macs and PCs are great for music. Both can also be a disaster if not setup well. Neither is inherently better or worse.

The he-man big boy systems (such as Pyramix DAWs) are based on PCs, but this is probably the result of the PC's open architecture and configurability as much as anything.

Anyone in the business of working with project studios refusing to work with PCs on principle (rather than simply not knowing enough about them) is cutting himself off from a lot of business as there are many more PC based studios than Mac based.

I will never understand the silly Mac v. PC and the SS v. tubes wars. All are wonderful if designed and implemented properly.

Editor
Editor's picture
Offline
Last seen: 13 years 3 months ago
Joined: Sep 1 2005 - 8:56am


Quote:
Both Macs and PCs are great for music. Both can also be a disaster if not setup well. Neither is inherently better or worse.

According to John Siau at Benchmark, iTunes 7.5 is a disaster as it sample-rate converts on every frequency except 96kHz and introduces sonic junk as it does so. I have yet to confirm this. He recommends Foobar for the PC, BTW.

John Atkinson
Editor, Stereophile

Elk
Elk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 26 2006 - 6:32am

This is unfortunate. I knew there were issues with iTunes in that you needed to be sure to have all DSP off, etc., but this new problem with 7.5 appears to be unavoidable unless your file is 96kHz.

I wonder why these issues exist. Perhaps it is because I am not a computer programmer, but wouldn't it be easier just to leave the sampling rate and bit depth alone and just pass it on to either the onboard DAC or to the digital output?

Thanks for the info.

dwiggins
dwiggins's picture
Offline
Last seen: Never ago
Joined: Nov 20 2007 - 1:45pm


Quote:
This is unfortunate. I knew there were issues with iTunes in that you needed to be sure to have all DSP off, etc., but this new problem with 7.5 appears to be unavoidable unless your file is 96kHz.

I wonder why these issues exist. Perhaps it is because I am not a computer programmer, but wouldn't it be easier just to leave the sampling rate and bit depth alone and just pass it on to either the onboard DAC or to the digital output?

You would think it would be that simple wouldn't you? Here's what I have found in my travels down the computer audio path:

There can be a lot of software from different providers between the data file and the digital out. The player software (iTunes, Foobar etc.) takes care of reading the data file and sending it in the direction of the soundcard. It can use one of the Windows sets of sound functions to do this: either the older Multimedia Extensions (MME), Directsound (DS) or Kernel Streaming (KS), or, if the card provides an ASIO driver it can use the ASIO functions. These functions interface in a defined manner with the soundcard driver and mixer software which manages the hardware. Finally the hardware may have its own DSP's and firmware which might also interfere before the data gets to the digital out. ASIO drivers where available usually offer the simplest route to the soundcard as they are intended to provide very low latency and work directly between the player program and the soundcard with no Windows functions in between. However, this will depend on the quality of the ASIO driver and the player program has to specifically support ASIO.

I've just had a look at iTunes 7.5 and there are no options to select any particular sound functions. I tested playing a 16-bit 44.1kHz wave file (Rickie Lee Jones' Easy Money) through an M-Audio 2496 card and found that the data at the digital out was significantly changed from the data on the file.

These are the first few 16-bit samples from the digital out:

F7FF F6FF F7FF F6FF F7FF F6FF F8FF F5FF F7FF F6FF F8FF F6FF F8FF F5FF F7FF F6FF F7FF F6FF F7FF F6FF F8FF F6FF F7FF F6FF F8FF F7FF F8FF F5FF F8FF F5FF F9FF F5FF F9FF F5FF F8FF F6FF F8FF F6FF F8FF F6FF F8FF F5FF F8FF F5FF F7FF F6FF F8FF F6FF

and here are the same samples from the original file:

F3FF F1FF F3FF F1FF F3FF F1FF F4FF F0FF F3FF F1FF F4FF F1FF F4FF F0FF F3FF F1FF F2FF F1FF F2FF F1FF F4FF F1FF F3FF F1FF F4FF F2FF F4FF F0FF F4FF F0FF F6FF F0FF F5FF F0FF F4FF F1FF F4FF F1FF F4FF F1FF F4FF F0FF F4FF F0FF F3FF F1FF F4FF F1FF

(Each letter represents the hexadecimal value of 4 bits)

Its hard to fathom why iTunes changes the data. It is not difficult to offer an option to send the data out unchanged, it would be nice for iTunes Audiophile users if such an option could be added.

Subjectively I believe I noticed the difference in the sound quality as soon as I started playing the file with iTunes in that it sounded harsher than using my normal player. I am fairly confident in this observation as the difference is not slight and I always use this track to test for bit-perfect output, however it is a subjective judgement. The objective data is clear.

I did a similar test with Foobar and found that it produced bit-perfect output using ASIO or DS but the output with its KS driver was very similar to iTunes.

Elk
Elk's picture
Offline
Last seen: 3 years 6 months ago
Joined: Dec 26 2006 - 6:32am

Thanks, dwiggins!

Great info and very interesting.

And here I installed Foobar only because it plays FLAC. (I have a separate dedicated music server and use various computers for ripping and encoding.)

struts
struts's picture
Offline
Last seen: 2 years 9 months ago
Joined: Feb 1 2007 - 12:02pm

Dave,

Your primary research is a real breath of fresh air in a debate steeped in opinions masquerading as facts and the perpetuation of myths. Thank you so much for sharing these findings; I am sure I am not the only one here who sincerely appreciates it.

Welcome again to the forum, please keep it coming!

ROLO46
ROLO46's picture
Offline
Last seen: Never ago
Joined: Sep 29 2006 - 10:16am

According to John Siau at Benchmark, iTunes 7.5 is a disaster as it sample-rate converts on every frequency except 96kHz and introduces sonic junk as it does so. I have yet to confirm this. He recommends Foobar for the PC, BTW.

John Atkinson
Editor, Stereophile

Crikey
When will this be fixed?
I have noticed low level clicks recently
Roger

Log in or register to post comments
-->
  • X