Article: Mitch Provides A Unique Perspective Of The Vancouver Audio Show

[mitchco]

View full article

[MikeJazz]

Brilliant...and this is why it's getting difficult to find a successor for my quad esl.

[The Computer Audiophile]

Hi @MikeJazz - Have you listened to the new MartinLogan Renaissance? I've heard great things about them. The bass is adjustable as well.

 

MartinLogan | Renaissance ESL 15A - Luxury Electrostatic Loudspeaker

[sphinxsix]

'Speakers and rooms typically have the greatest influence on the tonal response for any given sound reproduction system.'

This is the most fundamental truth as far as audio systems are concerned. If one doesn't understand it IMHO he either doesn't understand anything about audio or can't hear a ..damn thing..

A very good article written by a person who both understands and can hear a lot.

Very appreciated!

Got no time now but I think I'll be back with a longer comment soon.

[wdw]

I attended the show and so very happy to read your report...your experience of the MQA demonstration mirrored mine exactly. The presentation was lacklustre and did nothing to enlighten us as to the merits of MQA and seemed consistent with other reports of similar MQA presentations. I had no preconceptions and was anxious to hear all and everything but if that presentor is indicative of their attempts to enthuse us they could be in deep trouble.

As you describe; no comparisons, nothing.

[PeterV]

Hi Mitcho, thanks for this extensive review.

 

I have one comment with regard to the Time Coherence aspect of Loudspeakers and MQA: It is true that this aspect is of large importance for accurate music reproduction and indeed many loudspeaker do suffer from bade time-incoherent design... DSP is capable to repair a lot of those errors, but with MQA there is something unique involved and that is the capability to repair for both PRE-RINGING and post ringing effects of A?D and D/A converters. As you will agree, no loudspeaker is capable to produce PRE-ringing errors, so this part of time -incoherence is not present in the analog part of the reproduction of the sine wave. But in the digital processing of A/D an dD/A signals, this effect is present and audible. Especially since it is an unnatural effect, it causes fatigue. The MQA DSP is capable to recognize the way the original recording was processed, like Photoshop software is able to correct for lens profiles, chromatic- and geometric errors. The PRE-ringing of past A/D errors can only be corrected in the digital domain when this information is available or recognised by the MQA DSP. So there is a difference which, for the moment, only MQA is able to compensate for.

[PeterV]

With regard to the fact that MQA demonstrations are unable to demonstrate A/B comparison with original WAV or FLAC files, there is a logical explanation: Any MQA accredited DAC has been measured on a test bench what its intrinsic D/A time smearing effect is and it is being compensated for by the MQA DSP. I suppose (but never received prove or answers yet from anyone..) that the performance of a MQA certified DAC is intrinsically improved with regard to its D/A pre-and post-ringing effects.. this should result in an improved sound quality also when playing non-MQA encoded files. This phenomenon is what I have encountered myself with my NAD C390DD after upgrading it to V 2.89 MQA firmware. The DDFA DAC/Amp sounds even more analog, relaxed than when I prurchased it 3 years ago. When playing MQA authenticated files, the SQ increases a step further. So an objective comparison between MQA and non-MQA can only be done by using 2 the same types of DAC, one which is MQA accredited and one which is 'standard'

[4art]

Sphinxsix, I fully agree with you. Speakers and room acoustics are so tied together...

I remember a dutch article referring to Peter Walkers " To you speaks the Room '' ... So true.

[hvbias]

Always an education reading your blog entries, thank you Mitch!

[charlest]

So, LINN Product is right all the way

and Pitch accuracy is related to timing

[hvbias]

Mitch do you know if Floyd Toole has written anything about time domain speaker response in his books?

[dallasjustice]

Mitch,

Great show report!

 

I agree that the step response is a good metric. And with passive loudspeakers the step response is used to characterize those minimum phase crossovers to demonstrate crossover integration more so than absolute time coherence.

 

As you can see from a recent review of the Zu Audio speakers in stereophile, the Zu speakers exhibit a very vertical step impulse. But when one looks more closely at the lateral off-axis, it's clear that the Zu isn't a world beater. I've heard these speakers at shows and they aren't very impressive.

Zu Audio Soul Supreme loudspeaker Measurements | Stereophile.com

 

So I think using the step is helpful with DSP software like Acourate and much less helpful to fully characterize passive crossover loudspeaker systems.

 

Hvbias,

Floyd Toole would say that lateral off axis smoothness relates much better to stereo imaging and overall musical enjoyment than absolute time cohererence.

http://seanolive.blogspot.com/2008/12/part-3-relationship-between-loudspeaker.html?m=1

 

Michael.

[dallasjustice]

It's ironic that Meridian is now presumably interested in time coherence. Their active loudspeakers are anything but time coherent even tho it would have been easy to make them so.

Meridian DSP8000 digital active loudspeaker Measurements part 2 | Stereophile.com

[mitchco]

MikeJazz, sphinxsix, thanks for your comments. wdw, we will have to meet up at the next show!

 

Hi PeterV, thanks for your comments. Please note that when one is using DSP to correct the timing of loudspeakers, it is also correcting for any (pre)ringing in the DAC and anything else in the measurement/playback chain, and aligning it to a known standard, like the ideal step response in the article. One can also include the digital audio playback software, like JRiver in the measurement chain if desired. The DSP software, in this case, Acourate is working just like in your Photoshop example. Note that Acourate analyzes the full frequency range of the transient behavior and it is this result that is used as the basis for further calculations for frequency and timing corrections.

 

As a side note, DSP digital filters in loudspeaker correction can cause preringing using a linear phase FIR filter or post ringing in minimum phase FIR or IIR filters. In my eBook, I exaggerate the effects of preringing in the digital filters to orders of magnitude larger than what is measured in a typical D/A converter. With the transient nature of music, I find it very difficult to hear even the exaggerated preringing. You may note in the measured step (timing) responses above, there is no (pre)ringing at all in the step response as there is a preringing compensation applied in the digital correction filter, to not only correct the timing response of the speaker, but also the DAC. So it is not quite true that only MQA can fix a D/A pre and/or post ringing effects.

 

I would love to objectively compare the digital and analog outputs with/without MQA. My feeling it is likely to come to this:

Archimago's Musings: MUSINGS: Digital Interpolation Filters and Ringing (plus other Nyquist discussions and "proof" of High-Resolution Audio audibility)

[mitchco]

Hi Michael and hvbias, step response is one measurement attribute of several used for achieving accurate sound reproduction, if accuracy is one's preference. In my eBook, in addition to frequency and step response, I cover off; group delay, energy time curve, reverb time, polar response and interaural coherence coefficient (i.e., imaging).

 

Smooth off-axis frequency response, which is polar response in my eBook, is just as important as time coherence, in my opinion. I show in my eBook the smoothness of the response of my loudspeakers across a 6ft x 2ft listening area, which is largely contributed to the constant directivity waveguides I use from 500 Hz up. Michael, I did not see in the article linked that "Floyd Toole would say that lateral off axis smoothness relates much better to stereo imaging and overall musical enjoyment than absolute time coherence" From my read of the artilce, smooth on and off axis frequency response is what trained listeners would prefer, which I am in agreement with, and I presume you are as well, based on the loudspeakers you are listening to :-)

[dallasjustice]

Floyd Toole said,"In the design of loudspeaker systems, knowing the phase behavior of transducers is critical to the successful merging of acoustical outputs from multiple drivers in the crossover regions. Beyond that, it appears to be unimportant." Sound Reproduction 2nd ed. p.420.

 

The translation is that step response is a helpful metric to show how well a passive crossover behaves. I realize Mitch is also referring to absolute time coherence. In the context of loudspeaker design, phase coherence is a big deal due to its direct relationship with the frequency domain. All of the NRC/Harman studies can statistically tie listener preference to smooth frequency response off axis.

 

I've previously searched the AES library for any study which show the same listener preference for time domain and haven't found anything concrete. Of course, time and frequency are related to each other so step response is still very helpful. But it's not the holy grail. If it were, then Quads would be the perfect loudspeaker. Of course, they aren't, even tho I love how they sound. :-)

[mitchco]

Yup, agree with Floyd on that one, but that's not the issue I am trying to raise awareness about. I am talking about the acoustic signal arriving at one's ears matches as closely as possible to the waveforms stored in a digital media file. And in the case of this article, specifically, the time domain.

 

As you have found out, little AES or other research, other than anecdotal evidence, mine included, to suggest that time alignment does make an audible difference. I hope Dr. Toole updates his book to include audibility/listener preference studies around speaker time coherence.

 

As mentioned before, not many speakers are time aligned as it is technically very difficult to do properly. That's why I believe there is little research on this subject area, especially around audibility. With the advent of powerful DSP software, and new computer aided passive XO designs, like this Danley design, (see how many parts it takes to get a passive XO to be time aligned and many in the industry feel Danley's designs sound the best), I am hopeful that the speaker industry ups it's game around time coherence loudspeakers.

 

The thing is Michael, unlike many "tweaks", hardware or software, time alignment is easily measured objectively. Therefore, it is simply a matter of time that audibility tests will show whether real time alignment, and not just though the crossover region, is audible or not and what the benefits are. For me, I have spent the last 5 years researching and conducting my own audibility tests and I have my answer.

[dallasjustice]

I agree that if there were a way to test listener preference and the listeners could be trained up on soundstage effects I hear with time correction, the evidence would support your experience. I just think that comparing active crossovers and passive ones is really hard to do since there are so many excellent sounding passive speakers with imperfect time performance.

 

I do wish there were more exhibitors who used DSP in their setups. I try to encourage dealers and distributors to use DSP in the hotel rooms. I'd love to see a multi sub setup at a show with digital crossovers. I think listeners would be shocked at the dynamics; Maybe, for once, an audio show not dominated by boring 3 piece jazz combos or chamber music.

 

Yup, agree with Floyd on that one, but that's not the issue I am trying to raise awareness about. I am talking about the acoustic signal arriving at one's ears matches as closely as possible to the waveforms stored in a digital media file. And in the case of this article, specifically, the time domain.

 

As you have found out, little AES or other research, other than anecdotal evidence, mine included, to suggest that time alignment does make an audible difference. I hope Dr. Toole updates his book to include audibility/listener preference studies around speaker time coherence.

 

As mentioned before, not many speakers are time aligned as it is technically very difficult to do properly. That's why I believe there is little research on this subject area, especially around audibility. With the advent of powerful DSP software, and new computer aided passive XO designs, like this Danley design, (see how many parts it takes to get a passive XO to be time aligned and many in the industry feel Danley's designs sound the best), I am hopeful that the speaker industry ups it's game around time coherence loudspeakers.

 

The thing is Michael, unlike many "tweaks", hardware or software, time alignment is easily measured objectively. Therefore, it is simply a matter of time that audibility tests will show whether real time alignment, and not just though the crossover region, is audible or not and what the benefits are. For me, I have spent the last 5 years researching and conducting my own audibility tests and I have my answer.

[Account Closed]

Some of the reasoning in the posts above as to why there has been no direct A-B comparisons of MQA to WAV (or FLAC) of the same master at any of the shows so far leads me to believe that they are afraid to do it. Maybe the difference is not as great as some of the wet dream reviews have indicated. If the difference was substantial and repeatable, I would think you would want to flout it.

 

It may be a moot point for a long time to come if Apple buys Tidal and shuts it down as they have a habit of doing (see MOG/Beats and now Pono). MQA becomes a Knight with no Steed. MQA has always been about streaming something like hi-res with a quasi-DRM infrastructure. MQA downloads will be few and far inbetween as will all downloads.

 

MQA needs a streaming service (or at least a tier of one) that is dedicated to presenting music in full resolution. That would be expensive and require our community to support it in large numbers. MQA may be a bridge too far but only time will tell.

[simmobb2]

FFS just stand back and listen to yourself "I compare to a sound reproduction system that has been calibrated for accuracy" and "If I were to categorize from bass, mids to high frequencies, I listen for smooth bass response (+- 3dB tolerance)" and "My preference is for accurate sound reproduction so that the music arriving at my ears is as identical as possible to the music that is on the recording, regardless of format." This is writing for writings sake and I'd file it in the category 'the bleeding obvious'. Articles like this remind me of why I'm an infrequent visitor tbh. Back to the music for me I'm afraid.

[The Computer Audiophile]

FFS just stand back and listen to yourself "I compare to a sound reproduction system that has been calibrated for accuracy" and "If I were to categorize from bass, mids to high frequencies, I listen for smooth bass response (+- 3dB tolerance)" and "My preference is for accurate sound reproduction so that the music arriving at my ears is as identical as possible to the music that is on the recording, regardless of format." This is writing for writings sake and I'd file it in the category 'the bleeding obvious'. Articles like this remind me of why I'm an infrequent visitor tbh. Back to the music for me I'm afraid.

Wow. No good deed goes unpunished. can you elaborate on what you don't like about this article?

[Nikhil]

Floyd Toole said,"In the design of loudspeaker systems, knowing the phase behavior of transducers is critical to the successful merging of acoustical outputs from multiple drivers in the crossover regions. Beyond that, it appears to be unimportant." Sound Reproduction 2nd ed. p.420.

 

Time coherence is distinct from Phase coherence though. You can have a speaker that is phase coherent but not time coherent. A Time Coherent speaker will be both time and phase coherent.

 

 

"I compare to a sound reproduction system that has been calibrated for accuracy" and "If I were to categorize from bass, mids to high frequencies, I listen for smooth bass response (+- 3dB tolerance)" and "My preference is for accurate sound reproduction so that the music arriving at my ears is as identical as possible to the music that is on the recording, regardless of format." This is writing for writings sake and I'd file it in the category 'the bleeding obvious'. Articles like this remind me of why I'm an infrequent visitor tbh. Back to the music for me I'm afraid.

 

FWIW ... Mitch was just describing and defining time coherence.

Many talk about it - just like Meridien - but haven't the faintest idea. And it is true that once you get used to time coherent sound you really can hear "the box" when listening to speakers that are not time coherent.

 

 

@mitchco: Mitch, great article!

I also think that the Group Delay of the accompanying gear also add to the problem.

Would be very interested in your opinion on that aspect.

 

MQA is a borderline scam. I admire your restraint in the article.

[dallasjustice]

Phase is just a relative expression of time at a frequency. Here is a nice conversion tool:

http://www.sengpielaudio.com/calculator-timedelayphase.htm

 

If you've read the book I referenced, you would know Toole is talking about both phase and time. There are numerous quotes one could pull from his book saying the same thing. Amir pulls several here:

Interesting article by Mitchco. | Audio Science Review (ASR) Forum

You have to remember we are talking about speakers here which can be measured with a microphone. We aren't talking about the usual audiophile jibberish accompanying the latest worthless gadget or new imaginary technology.

 

 

Time coherence is distinct from Phase coherence though. You can have a speaker that is phase coherent but not time coherent. A Time Coherent speaker will be both time and phase coherent.

 

 

 

 

FWIW ... Mitch was just describing and defining time coherence.

Many in audio talk about it just like Meridien but haven't the faintest idea. And it is true that once you get used to time coherent sound you really can hear "the box" when listening to speakers that are not time coherent.

 

 

@mitchco: Mitch, great article!

I also think that the Group Delay of the accompanying gear also add to the problem.

Would be very interested in your opinion on that aspect.

[Tintinabulum]

It's ironic that Meridian is now presumably interested in time coherence. Their active loudspeakers are anything but time coherent even tho it would have been easy to make them so.

Meridian DSP8000 digital active loudspeaker Measurements part 2 | Stereophile.com

I'm not exactly sure what's meant by time coherence, but Meridian did introduce EBA to their DSP range a few years ago, enhanced bass alignment, ie DSP achieved time alignment of delivery of sound to the ear based on speed of travel of the various frequencies. If this is time coherence, then their speakers have been for a number of years?

[mitchco]

Time coherence is distinct from Phase coherence though. You can have a speaker that is phase coherent but not time coherent. A Time Coherent speaker will be both time and phase coherent.

 

@mitchco: Mitch, great article!

I also think that the Group Delay of the accompanying gear also add to the problem.

Would be very interested in your opinion on that aspect.

 

MQA is a borderline scam. I admire your restraint in the article.

 

Nikhil, thanks for the kind words. To expand on your first paragraph, in my case using Acourate DSP, I rotate the signal for 2 out of the three drivers, using one driver for time reference, by the number of samples measured in the z-offset. This lines up the acoustic centers in the vertical plane to form a single point of sound. Using a linear phase crossover, sums perfectly in both the frequency and time domain as illustrated in this article. Thus both time and phase coherent. My preference and I provided a link above to an advanced CA article showing how this is achieved, if folks want to check out the details.

 

One can use https://sourceforge.net/projects/rephase/ to adjust the amplitude and phase responses of the filter independently for one's loudspeakers to make them phase coherent, but not time aligned. See this article and Example 1: https://www.minidsp.com/applications/advanced-tools/rephase-fir-tool I encourage folks to try it out and see if you can hear a difference.

 

Nikhil, in the eBook, I dedicate a chapter on Group Delay. D/A converters analog outs, active preamps, amps, may all have some level of high pass filter applied, so as not to pass DC. Depending on the corner frequency, will add group delay in the bottom octave or two. My tests and research shows that my ears are not susceptible to GD below 100 Hz. The research starts at 500 Hz. This is the only article I could find on testing audibility of group delay below 100 Hz: group_delay

 

Cheers!