Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[JimH]

I added some Audiophile Info to our Wiki this week. 

And last weekend, in the course of rebuilding my home PC, I installed Vista and gave WASAPI a try with an optical cable connecting to my receiver.  Wow!  The sound was much more clear. 

I think I've underestimated the importance of what a difference better equipment can make.

If you haven't already done so, I recommend giving WASAPI and a digital connection a try.  (On XP, you'd have to do ASIO in place of WASAPI).  Links are in the Wiki article.

[rjm]

Does anyone know the maximum cable length for WASAPI?

Google and my favorite computer store do not provide any clues.

I would like to replace the 60' standard audio cable I have running from my computer to my receiver. Will WASAPI work over 60'?

[JimH]

Can you use an optical cable?  That should work.

[jmone]

WASAPI is the API in Vista for the rendering of the sound that is then passed to the "transmitter" (eg the HDMI / S/PDIF / Analog hardware that physicaly makes the connection and sends the info from your PC to your Receiver) it should make no difference from how you do things now.  You should be able to just select WASAPI instead of Direct Sound or Wave Out and away you go.  That said, 60' sounds like a lot for Analog, and pending you wire gauge I'd image you are suffering a fair bit of signal strength loss.  If you are looking at other connection choices then 60' should be OK with S/PDIF (Coax or Optical) but HDMI may require a repeater at that length (I can't find it at the moment, but there was a good test of cheap HDMI cables that found there was no difference between HDMI cables of any price EXCEPT as the the distances got larger).
Thanks
Nathan

[jmone]

I added some Audiophile Info to our Wiki this week. 
...I installed Vista and gave WASAPI a try with an optical cable connecting to my receiver.  Wow!  The sound was much more clear. 

Welcome aboard!  Now all we need is a DirectShow WASAPI filter.....(like the reclock guys are thinking of doing....)

[rjm]

WASAPI is the API in Vista for the rendering of the sound that is then passed to the "transmitter" (eg the HDMI / S/PDIF / Analog hardware that physicaly makes the connection and sends the info from your PC to your Receiver) it should make no difference from how you do things now.  You should be able to just select WASAPI instead of Direct Sound or Wave Out and away you go.  That said, 60' sounds like a lot for Analog, and pending you wire gauge I'd image you are suffering a fair bit of signal strength loss.  If you are looking at other connection choices then 60' should be OK with S/PDIF (Coax or Optical) but HDMI may require a repeater at that length (I can't find it at the moment, but there was a good test of cheap HDMI cables that found there was no difference between HDMI cables of any price EXCEPT as the the distances got larger).
Thanks
Nathan

I misunderstood that their was physical interface on top of WASAPI. Jim talked about an optical cable so I assumed that was standard for WASAPI. Thanks for setting me straight.

I was warned about the 60' cable being risky, but I had it custom made with good quality shielded cable and it seems to work fine.

[mabes]

Right now my sound card only has the mono output. It connects to a cable with left and right RCA jacks that goes to my amp. Will ASIO help any, or should I wait and get a better sound card.

[glynor]

I would like to replace the 60' standard audio cable I have running from my computer to my receiver. Will WASAPI work over 60'?

Can you use an optical cable?  That should work.

Generally optical cable is the least robust way to achieve long cable runs.  Optical audio cables are generally built from POF (Plastic Optical Fiber) which is pretty lossy stuff (but is much more flexible, less fragile, and doesn't need to be used in bundles like glass fibers).  Anything much over 50' is not recommended, even when using extremely high quality POF cables.

I'd strongly recommend you go with high-quality, certified copper coax instead.  BlueJeansCable sells very high quality cables for cheap, and they thoroughly test their cables at very long lengths: http://bluejeanscable.com/store/digital-audio/index.htm

[rjm]

perfect Glynor, thanks

[mark_h]

Replay Gain (as mentioned in the Wiki entry) has no place in an audiophile system...

Mark

[benn600]

When I try either ASIO or WASAPI I get playback errors.  I'm using HDMI from the computer.  Do I have to use optical for this?  It's nice using HDMI because it's one high-end cable as opposed to two.

[JimH]

WASAPI is only on Vista.

ASIO requires ASIO4ALL.  It's in the link above.

They are digital, so you need an optical cable or digital coax cable.

[jmone]

WASAPI works fine over my HDMI connection.  You may want to post your details in teh WASAPI thread http://yabb.jriver.com/interact/index.php?topic=49536.0 where I'm sure you will get some pointers.

Jim - I'm not sure the physical connection method matters as WASAPI as the Audio Renderers are normally ignorant of the transmition method (but then again it could be me that is ignorant on the matter).

[benn600]

I installed it and enabled it.  Surprisingly, it appears to be working because I get the popup when I start audio.  I am using HDMI audio.  Now I'll just have to see if I can hear a difference.

[newsposter]

??

Using optical/laser interconnects for high-def signals is A/V 101 kind of stuff guys.

[benn600]

My audio dealer recommends HDMI or coaxial digital audio because there are fewer conversions.  And many TRUE audiophiles prefer red/white phono plugs.  At some point, it gets turned into analog, right?  So if you have a high end player that outputs phono with excellent cables, you're golden.

[benn600]

Just how component can be better than HDMI in many cases...google it and you'll find a lot of articles and reasons.

[)p(]

HDMI in most cases has much more jitter then spdif but imho if your dac has good jitter rejection and most do only "audiophiles" will hear a difference

peter

[Peter_T]

Past a certain point, I don't believe a word audiophiles say. 

I do, however, have about $10k sunk into an amp, speakers, and cables.  It sounds great.  But hearing the difference in digital cables even with a great DAC... that kind of stuff seems unbelievable to me.  I think superman had really good ears, didn't he?

[craigmcg]

I know this is slightly off topic but I had understood ASIO4All to be designed to permit ASIO to be used by sound cards/ onboard chips whose drivers don't natively support ASIO. I'm currently running MC on 4 machines (using E-mu 1212m PCI, Echo Indigo DJ, M-Audio Audiophile 192, and Auzentech X-Meridian) at home using ASIO without having installed ASIO4All (or is it part of the MC install and I didn't notice it). That said, I do find that the sound is better using ASIO. I just installed Windows 7 beta on one of my PCs so I will try WASAPI soon.

Jim, I'm wondering if you were previously using analog cables to your receiver because if so, the improvement could also be related to a better DAC in the receiver than on your sound card. Forgive me if I have misunderstood.

Enjoy the weekend all!

[Alex B]

... Jim, I'm wondering if you were previously using analog cables to your receiver because if so, the improvement could also be related to a better DAC in the receiver than on your sound card. ...

I think that's exactly what Jim was trying to say. He just didn't want to put too many details in the wiki article.

Probably he used analog connection and either an onboard sound device or an old sound card on his old PC. If he was using SPDIF, then probably something was not correctly set or Windows Kernel was resampling the content in the background. (This can happen when the sound device not support the used sample rate. The lower end sound devices typically support only 48 kHz SPDIF pass-through.)

WASAPI (in exclusive mode) and ASIO can both deliver unaltered audio signal to a HT-receiver or an external stand-alone DAC.

[JimH]

I switched from an 1/8 inch mini stereo and RCA plugs cable to an optical cable.

But I also switched motherboards (three year old Sony VAIO desktop) to an ASUS P6T motherboard and Intel i7 CPU (Matt made).  It's booting from a solid state disk (Patriot 128GB Warp).    Programs are on the solid state disk.  Data is on a 1TB Western Digital Caviar Black.

The new motherboard has sound on board. 

And I switched from XP to Vista at the same time.

[JimH]

... hearing the difference in digital cables even with a great DAC... that kind of stuff seems unbelievable to me ...

I didn't mean to say it was the cables alone.  I just wanted to start a thread.  I knew some folks who actually knew something would jump in.

[)p(]

1. By using wasapi you now pass bitperfect to the receiver.

2. Either you're now using the receiver dac instead of the onboard dac or
more likely you also bypass one additional a/d conversion because the receiver will digitize any analog signal it gets. Some even in bypass direct mode

I see no reason to ever send a receiver an analog signal if you have the option of going digital. Also because a pc is noisy some prefer to use the optical out because it electrically shields the pc from the external dac. Personally I never noticed a difference.

peter

[MerlinWerks]

DirectSound in 32 bit XP can be bit-perfect provided your hardware/drivers/software don't resample or apply DSP etc. Unfortunately many do. More info HERE...

[JONCAT]

Replay Gain (as mentioned in the Wiki entry) has no place in an audiophile system...

Mark

sorry, the amount of bits tossed out with a minor reduction per Replay Gain with 32bit internal volume control will not and can not be audible (average -9db reduction here).....and I would bet my entire system that ABX testing could reveal RG was activated.

Furthermore you could make the argument from a qualitative perspective, that the discriminating audiophile has no place in their system for annoyingly varied amplitudes.

I do, however, have about $10k sunk into an amp, speakers, and cables.  It sounds great.

QFT.

Yes! When I listen to my system, and it sounds amazing, I'm not thinking about RG or how much more money I can dump into my stereo.

dc

[craigmcg]

Since you were using onboard audio, I'm relatively certain that much of the improvement is due to the better DAC in your receiver. When I do get a good home theater receiver, I will look forward to having one cable instead of 6 (so will my wife).

[JONCAT]

Also, optical is not susceptible to RFI or EMI as is coax or "analog" wire, although that doesn't mean that's always an issues affecting fidelity.

DC

[Peter_T]

I switched from an 1/8 inch mini stereo and RCA plugs cable to an optical cable.

OK, I believe you could hear that upgrade...

[jmone]

I've got a reasonable system but nothing exteme (Axiom Audio Speakers being Driven by a Yami V2700 Receiver) and given I'm well out of my teens my best hearing is long behind me - the difference between WASAPI and DirectSound/WaveOut (over the same HDMI connection) is dramatic. I don't know (or really care) if it is the "bit perfect" playback or the proper switching to 2ch but the result is great!
Nathan

[gappie]

WASAPI is only on Vista.

ASIO requires ASIO4ALL.  It's in the link above.

They are digital, so you need an optical cable or digital coax cable.

well. asio requires asio drivers. and when your sound card have none you could try asio 4 all. but i doubt it would make any difference then. but when your card has its own, i would recommend to use those.

 
gab

[newsposter]

it is also worth while taking a look at all of the 'conversions' done in the path from data file to speakers.

Not just the d/a conversions, but any potential d/d conversions using multiple DACs, including DACs that may be running at different bit resolutions.  Be sure to look at both hardware and software manipulations.  Going through multiple 12, 14, 16, 18, and 24 bit conversions and then undergoing the final digital to analog conversion for the end user can play hell with digital media.

In a predominantly digital system (like an HTPC), you want the data path to remain the same (bit rate, resolution, etc) from the stored media file to the ultimate output/display device.  Interconnects need to be digital, preferably optical.  But if using optical cables adds mis-matching d/d or d/a conversions, look at, in order, digital coax or doing an early conversion to a pure analog processing path.

In a higher-end mixed analog/digital system, you want to make the d/a conversion once, and as close to the playback device as possible.  Again, the use of quality, matching analog interconnects (bandwidth, capacitance, resistance, EMI resistance, etc) is just as important as working to maintain the integrity of a digital signal path.

Don't ignore the internal interconnects of your components.  If a mid-range receiver/amplifier uses internal components (even something as innocent looking as too-narrow copper PCB traces to move signals from pre-amp to amp sections) that muddy up a digital or analog signal, get rid of it.  If a media player or digital distribution system shoves bits through multiple conversions for no good reason, look closely at cleaning up that mess.

So draw out a block diagram of your media system from the storage/playback device all the way to the speakers and displays.  Leave enough room to document the software, hardware, d/a & d/d conversions, and type/bandwidth of interconnects.  Starting with this document, you will be able to identify the sources of distortion and/or interference and then work to correct them OR rework your system to fix things up.

DO NOT start out with trendy fixes like power filters, 'pure' oxygen-free copper cables, name-brand audio-feely black boxes and crap like that.  Get your system DOCUMENTED and then work out an improvement plan.  Once that is in hand you can start to apply effort and $$ (or $$$$$) to work things out.

[Frobozz]

JRiver should have a budget audiophile computer setup for the office.  That way y'all can play with the toys, enjoy good audio, and verify that Media Center plays well with the big toys.  Let Matt pick the equipment (he's got his eyes on the Sennheiser HD-800). 

Talk up the audiophile credibility of Media Center.  It's a great selling point.  There are many many media player programs but very very very few that have audiophile performance (true gapless, bit-perfect, high quality math, etc.). 

[Peter_T]

Talk up the audiophile credibility of Media Center.  It's a great selling point.  There are many many media player programs but very very very few that have audiophile performance (true gapless, bit-perfect, high quality math, etc.). 

...agreed.

[newsposter]

JR is going to have to work up direct control/hookup protocols for network connectable high-end components from Denon, Onkyo, etc.  These would be on top of the 'simple' optical/digital coax hookups that HTPC hardware is capable of.

One thing that high-end digital gear is generally capable of is live microphone monitoring and in-line correction/setup of 5.1 and 7.1 setups.

This is certainly a partnership opportunity to join up with a few manufacturers.  The software that comes with high-end components often sucks rocks.  The high-end guys are probably feeling sales competetion from everything from iPods to Roku boxes to digital feeds on sat radio and cable TV.

[hit_ny]

DirectSound in 32 bit XP can be bit-perfect provided your hardware/drivers/software don't resample or apply DSP etc. Unfortunately many do. More info HERE...

An interesting writeup, the author has written his own driver for c-media soundcards and says its better than one provided by the vendor.

When i pased it on to a friend, here is what i was told

I don't use a receiver for music, only analog through pretty souped-up pro soundcards. Not consumer cards. The setup we're talking about was Dynaudio BM6 passive monitors with a homemade LM3886 amp and fed off a stock E-mu 0404.

I invite you over to my house for a demonstration with the same equipment (the soundcard will be different and even more resolving) - you would be deaf to not spot it. On my music rig (Usher BE-718, Buffalo DAC, 1212m and Cardas cables) the difference is immediately apparent though it's not as if one is better, they're just different.

And you can only measure bit-perfect with a DTS capable signal and DTS-capable receiver - Creative got around this smartly by ensuring they were bit-perfect at 48KHz but not anything else. kmixer also operates at 48KHz. There's really no way to tell that the hardware is NOT resampling at any other input sampling rate (kmixer does for sure, the program is coded to operate at 48KHz. At maximum volume it is still active, even though no effects (volume is an effect) are being rendered. The bit-mangling is even more apparent when volume is lowered, which is why professional cards always use their own (hardware) mixers.

[MerlinWerks]

An interesting writeup, the author has written his own driver for c-media soundcards and says its better than one provided by the vendor.

And believe a lot of people would agree... 

And you can only measure bit-perfect with a DTS capable signal and DTS-capable receiver

Yes, this is widely accepted as the "acid test". You can get DTS encoded wave files for testing HERE...
You may be able to forgo the HT receiver if your soundcard has built in DTS/AC3 decoding, although I've never had the opportunity to try this.

kmixer also operates at 48KHz. There's really no way to tell that the hardware is NOT resampling at any other input sampling rate (kmixer does for sure, the program is coded to operate at 48KHz.

KMixer does not necessarily run at 48khz or do any resampling as explained HERE...

At maximum volume it is still active, even though no effects (volume is an effect) are being rendered.

This is one point that puzzles me as I've read it other places as well, I believe it's because of the math used by XP's KMixer (fixed vs. floating?) but,if this is true how can the setup detailed by DOgbert work? Board denizen thomaspf is quite knowledgeable about this, perhaps he can shed some light on it.

Some more interesting reading on "high-end" PC audio can be found HERE...

[Peter_T]

Some interesting reading, thanks.

One thing I don't understand:  how can a USB cable leading to a DAC possibly change the sound that the DAC produces?  Isn't that the whole point of digital: if the bit get there then the bits get there?  I can understand upgrading analog cables, but does the $1149 USB cable deliver the bits better?

[Alex B]

KMixer does not necessarily run at 48khz or do any resampling as explained...

You are correct. Kernal Mixer in XP does not resample if the sound device supports the sample rate of the incoming signal and there is only one incoming signal or all incoming signals have the same sample rate.

However, its mixer capabilities are always active and it does never output unaltered signal. Its volume control is actually of quite high quality and tries to do things correctly. To achieve this it always dithers the output and adds a very small amount of dithering noise. This noise is measurable, but way below of what is generally considered audible. Without dithering the output after any volume adjustment would contain aliasing artifacts that would show up as increased distortion in measurements and more possibly be audible than the negligible amount of added dithering noise.

On XP ASIO can bypass Kernel Mixer and provide a different path to the sound device. On Vista WASAPI in the exclusive mode can do the same.

[Alex B]

how can a USB cable leading to a DAC possibly change the sound that the DAC produces?  Isn't that the whole point of digital: if the bit get there then the bits get there?  I can understand upgrading analog cables, but does the $1149 USB cable deliver the bits better?

1) It can't. (edit: -- unless it is broken and data is not transmitted correctly. If the cable is not working reliably the output of the external audio device will be seriously corrupted in way or another. A "bad" USB cable can't just slightly alter the fine details of the outputted audio signal. That would be an impossible coincidence.)

2) USB aka Universal Serial Bus is not an "audio signal transmitter". It is a serial bus standard to interface devices to a host computer.

When it works correctly it is not different than for instance PCI bus. It may be slower or more complex because of its technical limitations and because it needs to be controlled by additional devices and drivers, but that's a different matter.

3) I don't think it does.

[MerlinWerks]

Some interesting reading, thanks.

One thing I don't understand:  how can a USB cable leading to a DAC possibly change the sound that the DAC produces?  Isn't that the whole point of digital: if the bit get there then the bits get there?  I can understand upgrading analog cables, but does the $1149 USB cable deliver the bits better?

First off I would in no way try justify those kind of cables digital or analog, especially since I DIY my own analog cables for around $15

Second, it's often overlooked, but digital audio is not just 1's and 0's, it's 1's, 0's and timing and I believe it's in dealing with this last element where the cable-razzi claim that the highend cables are better. Deciding whether a signal is a 1 or 0 is relatively trivial, recovering an accurate clock signal is more difficult and may be more susceptible to electrical interference, etc. so a higher quality cable may help here, but you'll never see me buying one 

Some good info on USB audio modes HERE...

and coaxial SPDIF cables HERE...

Although this guy is selling cables, his explanation is technically accurate. You can always make your own high quality spdif cable 

How much audible difference ANY of this makes will continue to be debated ad nauseum... 

[hit_ny]

The take away, for me anyways was...

you would be deaf to not spot it....the difference is immediately apparent though it's not as if one is better, they're just different.

meaning he can differentiate between WAV playback & FLAC or other lossless encoders.

He qualifies it by saying its not necessarily better, just different.

But if the chain is bitperfect then this is not possible.

Therefore bitperfect is being affected some way along the chain and thats how he can hear the difference, because...

There's really no way to tell that the hardware is NOT resampling at any other input sampling rate

So maybe the ones that insist on WAV only for playback have a point ?

To which i wonder why would any other sampling rate be used, at all  ?

[Alex B]

The data that is transmitted through the USB bus is not yet any kind of "synced" audio signal. It is raw memory data that is transmitted from a computer device to another. As well you could say that an external USB hard drive should be connected by an audiophile cable. I am sure that some cable dealers say that.

SPDIF is a different matter. SPDIF transmits clock synced PCM data.

EDIT

I read through some USB specifications. Apparently it is also possible to transmit more or less synced "audio" through USB. If the external device is not a fully featured sound device that is connected by USB (= similar to a sound card) I suppose syncing issues can be possible. (I need to read more about that...)

EDIT 2

Here is a discussion at HA: http://www.hydrogenaudio.org/forums/index.php?showtopic=53640
It supports my thought that USB audio is not anyhow synced before it is processed by the chip in the external sound device .

However, I am still trying read and understand this three-part article:
http://www.planetanalog.com/showArticle.jhtml?articleID=12801995
(it may take some time)

[MusicHawk]

I would like to replace the 60' standard audio cable I have running from my computer to my receiver.

60' (feet, right?) is way too long for common high-impedance audio cables (typically loaded to 10K ohms, though there's no real standard). Even if you use lower-gauge wire (thicker) for less resistence, you're going to have capacitance loss of high frequencies, and be exposed to inductance pickup of hum and other junk. Even shielding can be a problem, because while it can block outside interference it increases the capacitance roll-off.

A professional studio routing line-level audio would use balanced twisted pair within a grounded shield, loaded to 600 ohms. I can't even count the miles of Belden 8450 and 8451 cable I've soldered into radio/TV stations over the years. It's a "puny" 22 gauge, but used correctly it works as desired. However, consumer gear doesn't have balanced inputs/outputs.

Similarly, sending speaker-level audio any major distance shouldn't be done at 8 ohms. Sound distribution systems in big venues (theaters, concerts, etc) use a variety of methods but expensive speaker wire isn't one of them (though sufficient gauge is very important). For instance, speaker-level audio is often up-converted to 70 volts at the amplifier to send through long cables, then at the speakers back down to the few volts needed to drive them. The higher voltage allows much lower current which reduces the loss due to the cable's resistance. And to the driving line, each speaker appears to be much higher impedance than 8 ohms, which makes it much easier to connect many speakers in parallel because they do not load the line as much and they are well-isolated from each other. (It's amazing how many consumer audio systems are degraded by connecting multiple 8 ohm speakers in parallel, which not only misloads the amplifier but allows the speakers to interact and send the overall impedance and frequency response on a roller-coaster ride.)

None of the professional solutions involve buying the silly "audiophile" cables sold to novices, just an understanding of what comprises and affects the audio chain and appropriate gear and methods.

If you really need to run analog audio a long distance, there are devices available in the professional audio/PA world that will provide the right impedance, re-amplication and/or frequency tilting to do it.

[JONCAT]

It would be unique, and attract attraction if an ABX test system was an MC Advanced Option/Tool.

DC

[BillT]

60' (feet, right?) is way too long for common high-impedance audio cables (typically loaded to 10K ohms, though there's no real standard). Even if you use lower-gauge wire (thicker) for less resistence, you're going to have capacitance loss of high frequencies, and be exposed to inductance pickup of hum and other junk. Even shielding can be a problem, because while it can block outside interference it increases the capacitance roll-off.

It's not absolutely clear whether rjm was talking about analogue or digital signals, but even with analogue signals 60' unbalanced with decent cable shouldn't be a problem. The source impedance is the important factor and that is typically less than 100 ohms in half decent equipment, even computer sound cards. 100 ohms source impedance with about 2000pF capacitance of 20M of decent coax gives a -3db rolloff of 800Khz which is good enough for the most bat eared. Even if you use cheap coax with 320pF/M capacitance and a 200 ohm source impedance the roll off is -3db at 125KHz which is still much more than adequate. Interference is more likely to be an issue, but good quality cable, carefully routed in a reasonably electrically clean environment shouldn't be too problematic. Of course if he is sending SPDIF signals along decent coax there won't be a problem at all.

Balanced audio would be nice, but the advantages are only significant when working with small signals (mic level), or very long runs (much more than 60'). Not really significant with 2V audio signals fed from a low impedance.

[MerlinWerks]

Here is a discussion at HA: http://www.hydrogenaudio.org/forums/index.php?showtopic=53640
It supports my thought that USB audio is not anyhow synced before it is processed by the chip in the external sound device .

I don't know Alex, it seems to me they are assuming that most usb audio devices are using asynchronous mode when in fact very few do, most use the adaptive mode. Additionally, I think they are assuming that a USB audio device uses the same transfer protocol as say a usb hard drive, when typically they do not. Hard drives and other peripherals use bulk mode which has error correction and most usb audio devices use isochronous mode. I believe isochronous mode is considered more of a real-time streaming protocol and is optimized for minimal latency, etc. and has no error correction. I'm certainly no expert at this, but from what I've gathered so far I do believe in adaptive mode the PLL built in to the usb receiver chip still relies on the data packets "Start Of Frame" or SOF as timing reference from which it "reconstructs" the clock, unlike SPDIF that has the actual clock encoded in the bitstream.

FWIW, here's an audiophile company that's a J River fan, scroll down a little on the page...Wavelength Audio

A little too rich for my blood though 

[JimH]

MC14.0.31 added the ability to play files from memory.