Sound Stage - Is it all in the timing?

[One and a half]

Reading the Nordost Heimdall thread, mention was made of swapping out a USB cable with a different one changed the soundstage for the better. On that statement, let's assume everything else is the same, source material, amplification and speakers.

If it is the cable what properties of the cable causes the change in the soundstage to occur? This is not the first time a cable made a difference to soundstage, other influences can change soundstage quite dramatically.

 

Recordings themselves have variable quality as we know, not all recordings are made optimally, but when they are recorded 'properly' sound stage is excellent.

 

I don't believe the cable can actually add anything to the signal to perhaps brighten it up. The soundstage information needs to exist in the recording in the first place, because playing MP3, or radio has generally no soundstage at all, there's only the mid point in the speaker position and that's it.

 

If the information is in the recording, what process do we need to optimise to the fullest and best possible soundstage ? USB cabling is one way, but it's not final. Playback software, file format, media players, timing of the arrival of the USB stream to the DAC, possibly noise contained on the cable, PSU noise, even DACs? These are items that can influence from a CA playback perspective, but please feel free to add more. I really notice the change DSD makes compared to the same commercially made 192/24 of the same album.

 

Now a few years ago, comparing an Audiophilleo USB/SPDIF converter and an adaptive USB DAC (Stello signature) made the world of difference to soundstage. Should jitter be controlled (and where) or a combination of all items mentioned in one soup mix... as each bit in the chain is changed does it affect the other in a negative way?

[sandyk]

One and a half

 

I believe that it may come down to a further improvement in S/N which assists with the preservation of the low level harmonic structure. Ambience etc is also at quite low levels, and can be masked or sound "gritty" if the S/N is inadequate.

This is particularly obvious with DTV, where the audio level is already around 20dB down.

 

Alex

[elcorso]

Maybe is "soup mix" where a lot of factors influence the resulting recorded soundstage. Or fake soundstage from digital music players that adds reverberance, non existent on the recording, giving you the sensation of a deeper soundstage.

 

Regarding DSD, I believe we can get a realistic soundstage thanks to better timing, deeper bass and deeper bass control (non fake).

 

But the above are only my guess from what I'm hearring, I don't have a 'graphic proof' or measurement...

 

Roch

[crisnee]

The soundstage information needs to exist in the recording in the first place, because playing MP3, or radio has generally no soundstage at all, there's only the mid point in the speaker position and that's it.

 

It is absolutely not true that MP3 recordings lack sound stage (I don't know about radio). I have some high bit rate MP3 recordings that throw an excellent sound stage.

 

Chris

[crisnee]

An interesting phenomenon re sound stage involves headphones. There is software which works to create a sound stage similar to that of a speaker set-up. If set-up correctly for an individual (it involves head size and ear size compensation/adjustment) it works for something like 85% of humans. For the rest, not dice.

 

Maybe a similar thing applies where it concerns speakers.

 

Chris

[elcorso]

An interesting phenomenon re sound stage involves headphones. There is software which works to create a sound stage similar to that of a speaker set-up. If set-up correctly for an individual (it involves head size and ear size compensation/adjustment) it works for something like 85% of humans. For the rest, not dice.

 

Maybe a similar thing applies where it concerns speakers.

 

Chris

 

Binaural recordings are enhanced for headphones. But I like it even for loudspeakers.

 

Roch

[kennyb123]

Reading the Nordost Heimdall thread, mention was made of swapping out a USB cable with a different one changed the soundstage for the better. On that statement, let's assume everything else is the same, source material, amplification and speakers.

If it is the cable what properties of the cable causes the change in the soundstage to occur? This is not the first time a cable made a difference to soundstage, other influences can change soundstage quite dramatically.

 

I had reported that. The difference I largely attribute to the absence/presence of low-level details including spatial cues. The better USB cables do a better job of allowing these low-details to pass. Lesser USB cables filter these out.

 

An analogous thing happens on the analog side of things too.

 

 

If the information is in the recording, what process do we need to optimise to the fullest and best possible soundstage ? USB cabling is one way, but it's not final. Playback software, file format, media players, timing of the arrival of the USB stream to the DAC, possibly noise contained on the cable, PSU noise, even DACs? These are items that can influence from a CA playback perspective, but please feel free to add more.

 

Reducing the noise floor is a huge part of it. And this big gains can come by addressing this system-wide. Power conditioning plus the proper selection of interconnects and signal cables can significantly reduce the noise floor and allow more low level details to be heard. Resonance control helps in this area too. I got benefit also by using shorting plugs on all unused inputs on my preamp.

[sandyk]

Lesser USB cables filter these out.

 

Do they filter them out, or allow them to drown in low level wideband noise/ RF/EMI ?

I agree with your last paragraph though. I also switch the earth side of all inputs too.

[kennyb123]

Do they filter them out, or allow them to drown in low level wideband noise/ RF/EMI ?

 

That's a great question. Noise is a big part of it, but not sure that's the only thing going on.

[sandyk]

That's a great question. Noise is a big part of it, but not sure that's the only thing going on.

Hi Kenny

I use a low noise +5V Linear PSU with a shorter USB cable which has the red +5V wire disconnected at the PC end's USB-A plug. With many USB devices there is a potential earth loop due to internal connecting together of the 0 volts (black) wire and the shield. With a PC that is connected to IEC mains earth, this causes an earth loop. Breaking the shield connection between the USB In and Out sockets of this external PSU results in a worthwhile improvement, as does disconnecting a Broadband Modem result in a further small improvement. Normal USB cables do not have separate shields between Data and Power leads . Noisy +5V SMPS USB power is also a problem as in a normal cable it couples rubbish into the D+ and D- ( green and white) leads, as well as affecting the USB device itself.

 

Regards

Alex

[One and a half]

That's a great question. Noise is a big part of it, but not sure that's the only thing going on.

 

Noise floor in current DACs are very low -90db and below for the really good ones. If music contains the finer details to allow the soundstage to breathe, at which point does the DAC need to extremely quiet... -120db... would we hear at this level. Depending on the noise level from a typical USB connection, and from other sources, PSU mains noise injections, at which point would the level of noise interfere with the soundstage. Anybody have some rules of thumb?

 

It seems pointless to have a very quiet DAC when it's being swamped with noise via the front and back door.

[One and a half]

It is absolutely not true that MP3 recordings lack sound stage (I don't know about radio). I have some high bit rate MP3 recordings that throw an excellent sound stage.

 

Chris

 

I didn't consider high rate MP3, only the 128k variety.

[sandyk]

Depending on the noise level from a typical USB connection, and from other sources, PSU mains noise injections, at which point would the level of noise interfere with the soundstage. Anybody have some rules of thumb?

 

It seems pointless to have a very quiet DAC when it's being swamped with noise via the front and back door.

 

That's the $64 question ! Until recently , most would think you were nuts if you dared suggest that noise via the front and back door could affect the DAC itself, provided that the binary data wasn't corrupted. Many are now finding that SQ is greatly affected by RF/EMI hitching a ride along with the data via a typical USB connection, and going to great lengths and expense to attempt to correct USB shortcomings. Async USB was the 1st major step, but it's far from the complete answer by itself.

[pooger]

I think that besides a very low noise floor, soundstage quality is strongly affected by phase coherency. Slight phase anomalies will make the harmonics (overtones etc) not appear to emanate from the same space as the fundamentals--so that one gets less a sense of space or air surrounding each instrument. Without a solid sense of where each recorded sound is coming from, the lower level room information will become just more mush. Some of the best soundstaging I've heard has come from tube systems, playing LPs, that had horribly audible noise floors. But they still transported you into the venue.

 

In general, it seems to be easier to build more phase coherent digital filters than analog, so this may be less of an issue with good digital equipment. But ultimately we end up needing to deal with the speaker network, including any resonances that get thrown in by enclosures, and even the resonances in component level vibrations, which is why one of the most common improvements in good vibration attenuation is "an improved soundstage."

[Paul R]

I tend to think that all the soundstage information we hear is related to phase differences we perceive. We learn to interpret those phase differences as soundstage. A trick similar to those point pictures with a three-d image imbedded in them they used to sell in shopping malls.

 

If true that would suggest that a USB cable that affects the timing in some manner would be able to affect the soundstage. Since the data is digital at that point, the USB cable can not "filter" anything, not in the sense an analog signal can be filtered. Hence timing, in terms of propagation speed, jitter, or other time domain issues.

 

Power supplies may serve to enhance the detection threshold for digital dats, and thus enhance the timing, but have no actual effect on the data carried in the signal, unless and until they vary far enough from nominal to cause errors.

 

All of which is to say, why USB cables can and do have such an effect on the sound is a bit mysterious. But then, S/PDIF connections are even more vulnerable to cable and power differences than USB cables. Go fighter...

[wkhanna]

A large portion of my professional time & experience is devoted to root cause failure analysis.

 

It is a rare occurrence when only one variable can be attributed to a failure (or reduction of quality for the previously validated output).

 

It is my opinion cables, whether digital or analog, act as filters in the worst case, and are perfectly transparent in the best case. Nothing added, only subtracted.

 

It is also my opinion based on my limited personal experience with computer-audio that improved timing (reduction of jitter) & isolation from power sources will contribute to expanded soundstage.

 

But I have no idea what other specific factors are likely to contribute aside from those that are also common to analog systems.

[sandyk]

It is my opinion cables, whether digital or analog, act as filters in the worst case, and are perfectly transparent in the best case. Nothing added, only subtracted.

E.E. John Swenson has been able to demonstrate that RF/EMI goes along for the ride with the Binary Data via a typical USB cable. That's why we need improved isolation from the source in this area.

I am not sure if John has found the time to post screen shots showing this.

[acg]

I believe PeterSt has posted plots somewhere to show the "noise" riding on the usb analogue signal. Not sure where, but I have seen them.

[sandyk]

I believe PeterSt has posted plots somewhere to show the "noise" riding on the usb analogue signal. Not sure where, but I have seen them.

 

Mischievous little devil. (grin)

[Paul R]

E.E. John Swenson has been able to demonstrate that RF/EMI goes along for the ride with the Binary Data via a typical USB cable. That's why we need improved isolation from the source in this area.

I am not sure if John has found the time to post screen shots showing this.

 

It is not at all that simple, and assuming that the USB cable is somehow filtering the data, as would happen in an analog transmission is - well - to be blunt, wrong. The data at this point is digital and is not compromised.

 

About all that can be effected is (1) timing and (2) injection of EMI into the receiving device. Neither of those two operators acts like a filter. A filter would have to change the digital data, and that is absolutely not happening.

 

The most likely operator is timing, which can be affected by power, conductors, geometry, and other considerations. We also know jitter, which is a form of timing error, can drastically effect the sound quality.

 

I am not being pedantic here, and I certainly know that cables can make a difference. But it does bother me when people start talking about absolutely impossible situations. The real reasons why power makes a difference, or why jitter is audible, or why USB cables can sound different is intensely interesting, and vital to the hobby. But poor logic based upon incomplete and faulty understanding is a blue herring.

 

Paul

[sdolezalek]

If we were to assume that a cable (a) does not change the digital content, but (b) might be able to pass some forms of noise downstream and © might affect timing (whether through jitter or otherwise); do we actually have tools capable of measuring (b) and/or ©?

 

I have seen various claims in white papers or otherwise, but I don't think I have seen any measurements that are agreed across several product manufacturers that would allow for published noise or timing coherency numbers the way that we see published amplifier power or THD (I know those are also subject to interpretation).

[sandyk]

Paul

Perhaps you should have quoted from post 16, not mine ?

Alex

[wkhanna]

 

But I have no idea what other specific factors are likely to contribute aside from those that are also common to analog systems.

 

E.E. John Swenson has been able to demonstrate that RF/EMI goes along for the ride

 

Agreed, as I meant for this to be implied when i said; "....factors......that are also common to analog."

[Paul R]

If we were to assume that a cable (a) does not change the digital content, but (b) might be able to pass some forms of noise downstream and © might affect timing (whether through jitter or otherwise); do we actually have tools capable of measuring (b) and/or ©?

 

I have seen various claims in white papers or otherwise, but I don't think I have seen any measurements that are agreed across several product manufacturers that would allow for published noise or timing coherency numbers the way that we see published amplifier power or THD (I know those are also subject to interpretation).

 

Normal Network Analyzers can give us very exact timing for cables, though I admit, I usually do that with two cables of the exact same length and look for the relative differences between them. I also don't know if there are any standards being applied in the audio industry, but it would not surprise me to find that the industry uses the same standards and techniques as used for coaxial cables.

 

One additional thing I would like to see is measurements of reflected power- that can affect both electrical and optical transmissions, and I can see several mechanisms where it might mess about with the sound. At least, in theory.

 

Surely someone is out there running these kinds of tests?

 

-Paul

[gmgraves]

Reading the Nordost Heimdall thread, mention was made of swapping out a USB cable with a different one changed the soundstage for the better. On that statement, let's assume everything else is the same, source material, amplification and speakers.

If it is the cable what properties of the cable causes the change in the soundstage to occur? This is not the first time a cable made a difference to soundstage, other influences can change soundstage quite dramatically.

 

Recordings themselves have variable quality as we know, not all recordings are made optimally, but when they are recorded 'properly' sound stage is excellent.

 

I don't believe the cable can actually add anything to the signal to perhaps brighten it up. The soundstage information needs to exist in the recording in the first place, because playing MP3, or radio has generally no soundstage at all, there's only the mid point in the speaker position and that's it.

 

If the information is in the recording, what process do we need to optimise to the fullest and best possible soundstage ? USB cabling is one way, but it's not final. Playback software, file format, media players, timing of the arrival of the USB stream to the DAC, possibly noise contained on the cable, PSU noise, even DACs? These are items that can influence from a CA playback perspective, but please feel free to add more. I really notice the change DSD makes compared to the same commercially made 192/24 of the same album.

 

Now a few years ago, comparing an Audiophilleo USB/SPDIF converter and an adaptive USB DAC (Stello signature) made the world of difference to soundstage. Should jitter be controlled (and where) or a combination of all items mentioned in one soup mix... as each bit in the chain is changed does it affect the other in a negative way?

 

 

Soundstage is mostly a function of the recording, and even here, it depends a lot on one's definition of "soundstage." To some it means the width of the ensemble across the playback room. To others it means being able to close one's eyes while listening and being able to pick out each and every instrument in the ensemble is space. I.E being able to hear that the trumpets are behind the woodwinds and are to the right of the violins, etc. That's my definition. You get that by recording the space that that the ensemble is playing in, not the individual instruments themselves. If the instruments are individually miked, they have to be placed in their relative right-to-left location using pan pots to mix the instrument proportionally into both the right and left channel. IOW, the more an instrument is dominant in the right channel and the less that is in the left channel, the more to the right it will appear on playback. If the instrument is mixed equally into the right and left, it will appear in the center on playback, etc. However, while this approach will give the listener a sense of lateral instrumental placement, it can provide no depth. Only by using techniques that capture the space that an ensemble occupies can one get a truly realistic soundstage perspective with palpable instrumental placement and the proper "air" around each instrument. IOW, miking instruments rather than space is not stereo, it's multi-channel mono and virtually all pop and most jazz recordings are made that way.

 

I mention this because format has little to do with soundstage. Even the playback equipment has little to do with it (although small speakers (like the Rogers LS3s and similar) will generally image better than most large systems). I've heard decent stereo soundstage from a so-called "rack system". With digital, it doesn't matter if it's MP3, 16-bit 44 KHz sampling or high-res., if the stereo soundstage was properly captured, it will be there on playback. Listen, on speakers, to a live Boston Symphony concert on WCRB over the internet some Saturday and you will hear wonderful soundstage; wide, deep, and with glorious image specificity and it's 128 bps MP3! The engineers who broadcast/record the Boston Symphony for WCRB mike that orchestra in real stereo, and you can hear it even over the Internet.

 

I can't imagine the USB cable having any affect, positive or negative on imaging. As long as the digital file being sent over that USB cable has proper separation (more than about 30 dB - and all digital has greater than 90 dB of electronic separation), then the stereo soundstage will be intact. Even DACs which switch between right and left (as opposed to having a separate D/A for each channel), do it so fast as to not mess-up the timing between the arrival (at the microphone) of the phase and timing information produced by the music.