Taking digital all the way!

[DigiPete]

Exactly. I have been searching for these for at three years. . .

 

As far as I know, no one has made this leap.

 

 

Well this is exactly what the WiSA association is trying to change through standardisation and certification.

 

 

Bang & Olufsen

has the Immaculate Wireless Sound system (WiSA) that seems to match your requirements with the components:

 

Beolab 17 - wireless speaker

 

Beolab 18 - wireless speaker

 

Beolab 19 - wireless subwoofer

 

Beolab 20 - wireless full range speaker

 

BeoLab Transmitter 1 - Making all of your signal sources wireless

 

 

 

Dynaudio

has their proprietary stereo setups:

 

Xeo3 - wireless speaker

 

Xeo5 - wireless speaker

 

Xeo Transmitter - Making all of your signal sources wireless

[Jud]

Minimum phase = no pre-ringing. This moves the ringing energy fully to post-ringing, it does not eliminate it.

 

1st order filter does not post-ring either

 

Got it, thanks. (Though I assume this means "hardly at all for practical purposes," since complete lack of ringing is physically impossible.) With a DAC filter this would be way too little cut, but for a first order speaker crossover, that would be just what you'd want.

 

But what about phase correctness?

[YashN]

How much could and should we actually expect from going digital?

 

Taking digital all the way - how?

 

Are you ready to take digital all the way?

 

A lot of the thoughts for going through this process coming from a traditional analogue Hi-Fi audiophile world are expressed in this article on HiResAudioVideo.com:

 

From Hi-Fi to Computer Audiophile: how to have better sound with Hi-Res Audio

 

 

Many of the good things learned within the analogue audiophile experience can be applied to good use even in the new digital Hi-Fi chain, a lot of cabling is still in use among other things.

 

I've already moved all my stuff to digital storage and playback and I look forward to having all if not most of my collection in DSD 2s or greater.

 

The personal and social experience when listening to music is fantastic and it's much more convenient to store files on drives rather than have a hefty vinyl collection using up much space. Not to mention, you can also stream your music to other rooms.

 

Sometimes, for fun, we do a karaoke at home, and part of my setup automatically fetched the lyrics of whatever song we choose.

 

I have managed to set my system up so that it sounds as good or even better than what I was hearing in vinyl form as a teenager. Clean sound, full bandwidth, great dynamics, and no crackling, popping, or scratches etc..., especially with DSD 2x.

 

Surround sound in digital for both movies and music is awesome and makes for a more immersive experience.

 

Digital may not be for everyone yet as it takes a lot of work. I think future media players will make the process easier and also when more DSD 2x downloads are available at lower cost.

[beanbag]

Got it, thanks. (Though I assume this means "hardly at all for practical purposes," since complete lack of ringing is physically impossible.) With a DAC filter this would be way too little cut, but for a first order speaker crossover, that would be just what you'd want.

 

But what about phase correctness?

 

If I may take up a bit of your time for pedantic blah blah...

 

A first order mechanical, or electrical, or electromechanical system does not ring at all.

 

A typical 2nd order system is usually modeled as a mass on a spring, with damping. The "Q" of the system is how much damping there is. If damping is low (Q is high), then the mass oscillates back and forth. If you add more than "critical damping", then this system does not oscillate either.

 

Going back, a 1st order system is like if the mass is zero, so there is no momentum that lets you overshoot, and therefore doesn't ring.

 

In the digital domain, a 1st order system compares the difference between the output and input, and makes the output fractionally closer to the input. It is like Zeno's paradox in that the output can't overtake the input, and thus doesn't post-ring either.

 

The thing about phase correctness is this: A first order crossover has the property that the high pass and low pass signals sum up to exactly one, so by definition the amplitude and phase are "correct". Some higher order xo's don't sum up to exactly one, but still pretty close, like Linkwitz Riley and etc.

 

This article is about subtractive xo, but explains blurbs about how the different types of xo's sum up.

Subtractive Crossover Networks

[Jud]

If I may take up a bit of your time for pedantic blah blah...

 

A first order mechanical, or electrical, or electromechanical system does not ring at all.

 

A typical 2nd order system is usually modeled as a mass on a spring, with damping. The "Q" of the system is how much damping there is. If damping is low (Q is high), then the mass oscillates back and forth. If you add more than "critical damping", then this system does not oscillate either.

 

Going back, a 1st order system is like if the mass is zero, so there is no momentum that lets you overshoot, and therefore doesn't ring.

 

In the digital domain, a 1st order system compares the difference between the output and input, and makes the output fractionally closer to the input. It is like Zeno's paradox in that the output can't overtake the input, and thus doesn't post-ring either.

 

The thing about phase correctness is this: A first order crossover has the property that the high pass and low pass signals sum up to exactly one, so by definition the amplitude and phase are "correct". Some higher order xo's don't sum up to exactly one, but still pretty close, like Linkwitz Riley and etc.

 

This article is about subtractive xo, but explains blurbs about how the different types of xo's sum up.

Subtractive Crossover Networks

 

Thank you again.

 

If I may presume on your knowledge and good nature with one more question (after which I will start RTF"M" for myself, I promise), why minimum phase if no ringing?

[beanbag]

Sorry, I don't understand the gist of your question.

 

Minimum and linear phase have precise technical definitions, but if I may overgeneralize, most mechanical and non-digital electronic systems are minimum phase, which effectively means that if the amplitude changes with frequency, so will the phase. Digital systems can also (but not required) be linear phase, which effectively means you can tweak the amplitude and phase separately. This is often used in digital xo's to "fix" overlap issues with the high pass and low pass sections. I think linear phase does not have to ring either, but I am not sure about that one.

 

Since your speakers are 1st order xo, they sum perfectly, there won't be ringing anyway, and you don't need to bust out the linear phase stuff to fix anything.

[Jud]

Sorry, I don't understand the gist of your question.

 

Minimum and linear phase have precise technical definitions, but if I may overgeneralize, most mechanical and non-digital electronic systems are minimum phase, which effectively means that if the amplitude changes with frequency, so will the phase. Digital systems can also (but not required) be linear phase, which effectively means you can tweak the amplitude and phase separately. This is often used in digital xo's to "fix" overlap issues with the high pass and low pass sections. I think linear phase does not have to ring either, but I am not sure about that one.

 

Since your speakers are 1st order xo, they sum perfectly, there won't be ringing anyway, and you don't need to bust out the linear phase stuff to fix anything.

 

Clearing up the mystery:

 

For DAC filters, min phase is used to "fix" pre-ringing. It pushes all the ringing energy post-impulse. It doesn't get rid of any ringing; in fact there's more post-ringing. With linear phase, ringing energy is evenly distributed pre and post impulse.

 

Minimum and linear phase don't affect total ringing energy, just the time distribution of what's there. Total ringing energy is determined by the amount of cut (filter slope/order).

[Sam Lord]

I want to supplement beanbag's answer to Jud's questions.

 

The phase correctness of 1st order analog crossovers happens this way in the simplest case, consisting of a two-driver speaker with a tweeter and woofer of equal sensitivity (rare!) and a crossover with only an inductor and a capacitor. We'll treat the impedances of the drivers as ideal, i.e. purely resistive:

 

1) The tweeter is in series (connected through) a capacitor. At the crossover frequency, the resistive impedance (resistance) of the tweeter equals the capacitive reactance of the cap, so equal energy travels through both and the output of the tweeter is almost exactly -3dB compared to its output at, say, 5x that frequency. But the signal through the tweeter leads the input signal by 45 degrees because of the storage action of the capacitor. At this point the tweeter's energy output is 50% and its amplitude is 70.7% of nominal.

 

2) The woofer is in series with an inductor which has a similar effect except that the phase of the signal reaching the woofer *lags* the input signal by 45 degrees. So the tweeter "hands off" the signal to the woofer. Only the *addition* of woofer and tweeter signals are in phase with the input signal. In such systems the individual trandsducers are never in phase with each other or the input signal.

 

3) At the crossover point is the combined woofer and tweeter output at a higher level (3dB) and in phase with the input; nearby frequencies are *close* to that, but not exactly. So makers can play a bit with the inductor coil and capacitor values to adjust the xover frequency as their ears suggest. And the transducers and wires and crossovers all form quite complex circuits when all reasonable aspects like voice-coil inductance and diaphragm mass and stiffness are considered.

 

4) Phase correction is not the same as time alignment, but one can address the other pretty well in many cases. In digital audio, time alignment can be solved almost perfectly by a simple register shift to effect a delay. Typically (almost always), lower frequency drivers have response delays relative to HF drivers for various reasons, so applying digital delay to the HF drivers is an excellent application of digital systems.

 

I'm not qualified to address the "how" part of digital filters.

[DigiPete]

I will refrain from trying to argue the bliss of digital filtering, I am afraid I'll get it wrong in more than a few ways.

 

I will however urge all of you to go listen to a Steinway Lyngdorf rig.

Even the smallest will do (Model S in stereo)

 

The sound is the best argument I know for how sublime digital amplification incl. filtering can be.

 

QED

[jcbenten]

Well this is exactly what the WiSA association is trying to change through standardisation and certification.

 

 

Bang & Olufsen

has the Immaculate Wireless Sound system (WiSA) that seems to match your requirements with the components:

 

Beolab 17 - wireless speaker

 

Beolab 18 - wireless speaker

 

Beolab 19 - wireless subwoofer

 

Beolab 20 - wireless full range speaker

 

BeoLab Transmitter 1 - Making all of your signal sources wireless

 

 

 

Dynaudio

has their proprietary stereo setups:

 

Xeo3 - wireless speaker

 

Xeo5 - wireless speaker

 

Xeo Transmitter - Making all of your signal sources wireless

 

The transmitters need more inputs: Phono, DVD/BR, Game System

 

I have been looking at the XEO since it came out but, for now, it is a bit to limiting.

[audiocat]

I'm willing to do it. What I've heard from Devialet, including class D hybrids like Micromega and Chapter makes me want to pursue it. Digital ime has taken on its on characteristics. There is a synthetic quality to the playback that i love. It's not studio grade playback that includes a natural timbre to instruments but a very processed and alluring sound. It works because noise and distortion have been engineered down.

[Kees de Visser]

I will however urge all of you to go listen to a Steinway Lyngdorf rig. Even the smallest will do (Model S in stereo)

Hard to imagine it can beat an analog Steinway

Digital microphones are slowly becoming more popular. If I wouldn't already have invested so much in analog mics I would seriously consider them. Wireless would be even better, but it's still a bit early for that (power being a problem).

And according to rumors Apple seems to be heading to active (digital?) headphones. It looks like the digital chain is getting longer. Since this is mostly PCM format, it might be not so good news for DSD lovers.

[InfernoSTi]

I'm taking my first plunge into taking digital "all the way" with an all digital amp that takes a SPDIF input from my computer and outputs a speaker level signal without any analog in between. I am looking forward to hearing this type of setup (no DAC, not analog amp). Its going to have to be very good to beat my Auralic Vega and Quicksilver Silver 88 mono block amps...

 

Best,

John

[Ajax]

I'm taking my first plunge into taking digital "all the way" with an all digital amp that takes a SPDIF input from my computer and outputs a speaker level signal without any analog in between. I am looking forward to hearing this type of setup (no DAC, not analog amp). Its going to have to be very good to beat my Auralic Vega and Quicksilver Silver 88 mono block amps...

 

Best,

John

 

Hi John,

 

What's the name of the amp/dac combination? looking forward the hearing your experiences / impressions.

[InfernoSTi]

Hi John,

 

What's the name of the amp/dac combination? looking forward the hearing your experiences / impressions.

 

Hi Ajax,

 

I'm giving the Core Audio Technology Kratos MKII a try. It's in my price range and it's recently been used by my speaker manufacturer with a set of speakers very similar to mine at the Capital Audio Fest 2014 as reviewed by Part Time Audiophile, here: CAF 2014: Core Audio Technology, Hawthorne Audio and GIK, take computer audio to eleven | Confessions of a Part-Time Audiophile My speakers are an MTM configuration of the main baffles in the article with the bass drivers doubled up in secondary baffles (as seen in my avatar).

 

Since I'm a tube guy, this is a big step in a new direction...I've only used self-built Bottlehead 2A3 and 300B and now Quicksilver Silver 88's with these speakers (or any speakers for the last four or five years). I spent a little time discussing the system with Darrel Hawthorne who is also a tube guy and also spent a lot of time back and forth with Ryan before decided it was worth a listen.

 

John

[orca99]

Look forward to listning to my new beolab 18/19.....soon

[orca99]

[quote

The new Magnepan 1.7 quasi ribbon speakers has however taken my fancy. They offer a transparency and openness that is difficult to beat. The passive X-ver are a know issue, and they do leave a lot to be desired.

 

My wish setup therefor includes digital X-over and power-DAC tri-amping for the stereo bit plus something for 5.1.

 

I suppose that I have gone completely mad!

 

I have been running a similar system for 10 years - but with modified Martin Logans, which have been given an X-oversectomy.

I use an digital integrated Amp which include x-over, Room correction, Power to the Panels, and a second identical amp which only provides power to the woofers.

I have the impression that the time correction in the room correction sort out any phase issues......anyway it sounds good.

The system will soon be deported to my "cave", and we will have something simpler in the living room. It will also be all digital - Beolab 18/19's full WISA 7.1 sustem to follow.

[Fitzcaraldo215]

A little late to this thread. I have for about a year been using the "digital all the way" route for as far as I can conveniently do so for right now. From an equipment standpoint, my system seems conceptually simple:

 

PC > USB > Exasound e28 DAC > 8 analog balanced XLR > 7 analog amp channels > 7 analog speaker cables > 7 speakers. My active, powered subwoofer connects directly via one analog XLR to the DAC.

 

It is not digital "all the way", as is clear. Someday, it might be nice to have digital running directly to active speakers with digital inputs to internal DACs and amps. Meridian has been doing much of that for over a decade from their Mch processor to their active speakers. Lyngdorf/Steinway also appears to have digital from their Mch processor to their digital amps. Those are likely to be specialized interfaces, hence possibly difficult to use with other gear. But, my "Mch processor" is the PC and for now USB out to a DAC with analog from there seems the best way to go, better than HDMI to an external prepro which I had been using before.

Beyond the hardware signal path I described above, there are some wrinkles and limitations. There is, of course, a NAS containing my library of mainly DSD rips of Mch SACDs, and also BD-A and BD-V. I do not play CD-based music. SACD cannot be played directly by the PC due to DRM and must be ripped offline via PS3. BD or CD can be played or ripped by the PC using built in optical drives and hacking software as necessary. I also have an external Ethernet connected TV tuner using cable card for cable TV. So, a normal external disc player or cable box cannot be used with the PC. Those along with my digital prepro have all been eliminated from my system.

Software in the PC is straightforward. I use JRiver because it supports all my media types and Mch audio. It also provides digital bass management and crossovers for my subwoofer. And, it has a very nice iPad remote playback control. Dirac Live in the PC provides excellent Mch room EQ up to 192k PCM. It, in turn, interfaces to the ASIO driver for my Exasound DAC.

I would not say that integrating all this was easy. It took time to learn a lot of new technology, but it was all doable. It can also be upset in operation by software issues, like OS updates. I am sticking with Win 7 Pro for now, which should have fewer OS updates. So, it is a bit less stable than separate non-PC components would be. I have learned how to deal with most issues that crop up fairly easily, though.

But, the sound is pretty awesome, the best I have ever had by a substantial margin.