Sunday, 18 November 2018

MUSINGS: Multidimensional Audio! (An old article... Subjective perceptual ability... And a rigged Atmos / DTS:X setup... :-)


Well folks, it has been a busy last number of weeks around here since I returned from the overseas work trip. No time to play with much audio or tech stuff as a result until earlier this week.

Today I want to just discuss/show a couple of things...

Consider the phrase "multidimensional audio". In today's world, I think if anyone were to use that phrase, one would be thinking about multichannel and surround sound as "multidimensional". Instead, that graphic above is taken from this article - "Multidimensional Audio" (1979) by Henning Møller of B&K.

It's a thoughtful article that harkens from a time when analogue audio was king, years before the availability of even 1Mbit (~128KB) memory on a chip, and a handful of years before digital audio took over (first CDs released in 1982). Reading a paragraph like this from the article:
"Nevertheless there is often open warfare between the "subjective people" and the "objective people". People that really have been listening intensively for years often say that measurements are absolutely useless because they judge from the few old-fashioned measurements they might know. And often the so-called objective engineers say that, for instance, reviewers are crazy because they judge contrary to the measurements."
Suggests that in audio discussions on some forums, evidently nothing has changed in the last 40 years :-)!

Nonetheless, the article marries both subjective and objective perspectives, along with a philosophy around it which IMO is just as relevant today as it was then. In the article, Møller lists his 6 objective "dimensions" worth keeping in mind when judging audio quality...

1. Steady State Distortion - represented as measurements like THD, IMD and perhaps correlated on a distortion-frequency graph. Good points there and these days we still use these measurements of course.

2. Transient distortion - discussions here especially relevant for amplifiers.

3. Audible effects of wow & flutter, rumble, tone arm resonances - functions of playback source devices. Clearly in the days of modern digital (especially hi-res digital), these effects are now essentially obsolete (cf. Thoughts on Vinyl LP Fidelity). We have way lower noise levels, minimal jitter compared to LP playback wow & flutter temporal anomalies, and no tone arm resonance to speak of in digital. Improvements these days are on the order of magnitudes unless one still primarily listens to vinyl of course.

4. Gating, Early Reflections and Box Sounds - effects and limitations of the speaker system. Still very much relevant these days although I suspect improvements would be clearly notable on equivalent tier devices if one were to measure the old and the new.

5. Frequency response in the actual listening room (using 1/3 octave pink noise) - ahhh yes... Still also very much relevant and a huge factor! Remember the importance of room measurements and I do believe it's essential to know about the rooms in which reviewers listen to their music. Nice to see the descriptions of the listening room on Computer Audiophile recently for example. 1/3 octave is rather coarse these days and at least 1/6-octave preferred (if not 1/12-octave).

6. Phase measurements and transient response - sure, of course... Again, remember the power of room correction DSP techniques capable of acting at the level of impulse/step responses as well as correcting significant frequency-domain anomalies.

While these measurement/objective dimensions are still relevant these days, when reading the article, what is of course notable compared to 1979 is the testing technology itself. Today, we audio enthusiasts have at our disposal software (often freely available like Room EQ Wizard) that will give us many of the parameters listed above quite easily with resolutions unattainable in 1979 (certainly not at home!). Computing resources and hardware devices like high quality ADCs and decent measurement microphones, can be purchased without huge expense.

Ultimately, what is most important from the Møller article is the underlying concept of bridging the subjectivist-objectivist divide by embracing all the dimensions of audio within reason. His use of the "apodization" concept beyond the technical definition to suggest:
The "truth" is always somewhere in between. Therefore the practical version of Apodization is to find the optimum compromise between the sharp extremes by smoothing things out.
seems appropriate. There is no need to hold on to "sharp extremes" of pure objectivism nor pure subjectivism. After all, "hi-fi" 2-channel sound reproduction is itself a compromise; a simulacrum of actual acoustic events in actual acoustic spaces which we cannot fully capture with 100% accuracy nor of course reproduce with exact fidelity. At best, we hope that each piece of the reproduction chain can do its part in producing the "local" objective intent such that as a whole, the "global" effect of the system represents subjectively "good sound".

Since this article tries to marry both objectivity and subjectivity, let's add one more "dimension". Let's add the dimension of the subjective listener:

7. High fidelity perceptual ability.

This is important because ultimately sound quality is only as "good" as the ability of the listener to perceive and judge such a thing. This dimension can to some extent by objectively understood as well because we can measure the hearing ability of the "subject" him/herself. The listener is the ultimate arbiter of "good sound" but we know that just like all human traits, there is a normal curve to human abilities - not all listeners are equal or desire the same thing. And I think it's just as likely that not all human beings can nor care to appreciate high-fidelity sound.

While we can calibrate measurement devices like microphones, ADCs, and oscilloscopes, the human "subject" is much more complex with all kinds of perceptual and psychological strengths and weaknesses we can never truly know, nor IMO fully trust when it comes to what is said or believed no matter how sincere. At best, we can measure hearing acuity and listeners can be trained, perhaps "certifying" them as capable listeners (like taking a test such as this old "Philips Golden Ears Challenge").

For me, the contention comes when I see supposedly "authoritative" reviews based only on subjective listening, especially when all the other objective dimensions tell us that the product more than likely cannot perform as the subjective reviewer claims! Classic examples of this are the countless cable reviews of remarkably expensive products (here's a recent example - remarkable!). The result of a purely subjective review can only be as good as the reviewing "instrument", the abilities and conditions of the person writing the observations if objective results cannot be used to converge on the same conclusion. We only hope that the reviewer has considered and accounted for questions such as:

- With what other gear was the review based on?
- How were the listening room conditions?
- How long did one listen?
- Was listening done alone or in the company of others and what opinions were added if so?
- What music was used? Was it music one liked? Was the mastering good?
- How was the listener's mood state during the sessions?
- Was there stress with looming deadlines or other life events?
- What is the hearing acuity at the time of the auditions (tinnitus, sinus conditions, recent loud noise exposure, presbycusis)?
- Were "substances" involved at some point in the audition to modify perceptions and mental state? :-)
- Did the physical appearance of the gear or price tag affect the listening impressions?
- What history does the listener have with the manufacturer, representatives, or previous products? Can the reviewer honestly state that relationships, perks, and financial support did not play a part in the product's perceived performance?
etc...

I agree with Møller that a broad set of objective measurements will correlate with subjectively "good sound". I've heard it asked whether a deaf engineer is capable of making good sounding gear based just on measurements. IMO, of course this can happen! After decades of research and generations of consumer audio, it would be an absolute failure of science and engineering if it were otherwise! The unfortunate part would be that without being able to listen, the engineer would ultimately not be able to verify the design and use the adjective "good" in a global, subjective fashion.

As I've said before, I think having a "more objective" bias when it comes to equipment reviews and methodology simply makes sense as providing much more information to the reader. In a related way, when was the last time you dug up your favorite subjective reviewer's articles from 1979 and found the described sound quality impressions confidently applicable in 2018?

While the subjective reviewer with all his/her complexities could be long gone from the audiophile scene, improvements in objective audio fidelity can be appreciated across time documented by the value and relevance of objective facts.

-------------------------------------

As mentioned above, the phrase "multidimensional audio" these days probably evokes the idea of surround sound. Back in early 2015, I mentioned the advent of two new sound formats - Dolby Atmos and MQA. While you know my thoughts on MQA, Atmos (along with DTS:X which came out a little later) is the "real deal" providing more dimensionality in the form of audio processing of "objects" in space. This past week I had some fun listening to a simple "rigged" Dolby Atmos set-up at home.

(By the way, remember that Dolby TrueHD, which is the lossless compression audio format typically used for these Atmos soundtracks is actually Meridian Lossless Packing [MLP] which Bob Stuart helped create, like MQA... See... I'm not biased against Stuart - only when there's reason to show that the technology is poor!)

Here's my audio / home theater room as of mid-November 2018:


Notice I've added a couple of rear height channels situated on top of my vinyl collection (there's also my old Onkyo receiver, Sony laser disc player, and Playstation 1 up there).

Since my Yamaha RX-V781 receiver (if buying these days, I'd consider the current Yamaha AVENTAGE RX-A1070 model) is capable of decoding 2 height/ceiling Atmos or DTS:X channels, and I had a couple of old Energy C100 bookshelf speakers hanging around, why not employ them to improve my surround sound system?

So I bought some generic 14G speaker wires, 8 of these Hudson Hi-Fi 1.25" hemisphere silicone footers:


And found inexpensive yoga foam bricks of the appropriate height to prop the speakers up to the calculated ~32° I needed for them to be angled at. The reflections off the ceiling would be directed just in front of those seated on the sofa. Here's a side view of the rigged Atmos speaker:

Note the two pairs of hemisphere footers stuck on at the base and upper-mid of the speaker to prevent slippage off the foam brick.
The last thing to do was to tweak the speaker settings on the Yamaha receiver. With decoding set to "Straight" mode so Atmos and DTS:X comes out unadulterated, I used the automatic YPAO measurement with a microphone and manually set some crossover frequencies afterwards:


Notice I've set them as overhead "Presence" speakers (instead of typical ceiling bounce speakers which I believe incorporates Dolby's HRTF). While I am doing some "ceiling bounce" with those speakers, they are physically placed higher up unlike the "Atmos Enabled" speakers like these where an up-firing drive unit is mounted just on top of the speaker cabinet slightly higher than ear-level (typically with only a 20° angulation forward). My speakers are set up as "Small" speakers with crossover at 120Hz. Some have recommended for reduced localization to increase the crossover frequency to 200Hz; something I'll try later. I have flat, smooth (not stucco or otherwise uneven), and reasonably reflective 8' ceilings which is exactly the type of conditions recommended for "bouncing" the sound.

Ideally, one should be going with the same speaker manufacturer and line of speakers for tonal matching. Notice that I have a pair of Paradigm Signature S8's powered by my Emotiva monoblock amps. The Paradigm Signature C3 center channel, Studio 80v3's in the rear are powered by the receiver. Now we add the 10-year old Energy C100 speaker pair to the receiver amp. The Energy sounds similar in tonality to the rear Paradigm Studio 80's. Perhaps not surprising since both are Canadian companies using research from the National Research Council and are not far apart in age.

No doubt my kind of arrangement isn't exactly what Dolby calls for with an Atmos 5.1.2 set-up:

Official Dolby Atmos 5.1.2 configuration.
But doing the official Dolby configuration would have meant that I carved out some permanent holes into my ceiling. Not something I'm willing to commit to at this point!

So how does it sound?

Pretty decent, actually :-).

The height and angulation worked quite well. Seated at the "sweet spot" on the sofa, when playing pink noise, the sounds from the "presence" speakers do appear like they're coming from directly overhead rather than behind when watching the screen. Atmos demo material (can be downloaded here) as well as some DTS:X demos (here) sound good showing off the extra dimensionality. I also checked out a few of my UHD Blu-Ray movies with Dolby TrueHD Atmos soundtracks: the start of Mad Max: Fury Road when Max hears the voices "worming their way into the black matter of my brain", the gory yet funny start of Deadpool 2 with contributions from Dolly Parton and Enya, and the fantastic overall surround mix for John Wick Chapter 2 come to mind.

The effect from Atmos has been described in a number of ways. Basically the intent is to create an impression of a height dimension to the sound. Most of the time, the effect is one of a sense that there is a "canopy" over the soundstage. Like there's more diffuse space in the room "up there"; the illusion that the sky exists in one's home theater. Occasionally, with more aggressive soundtracks, one will hear things like airplanes/vehicles overhead (eg. Blade Runner 2049). The Omaha Beach landing scene in the recent 4K remastered Saving Private Ryan with Atmos encoding sounds appropriately horrifying with bullets, explosions, and all manners of violence happening around and up and down. [Seriously, I hope this is the last time I buy another copy of Saving Private Ryan! :-)]

Yes, I'm sure mounting speakers on the ceiling and a .4 Atmos arrangement with both front and rear speakers would make the experience even better. For now, I'm content. Note that Andrew Jones has suggested (see video) that for small theater spaces with low ceilings, maybe a diffuse sound is better than pin-point discrete speakers in the ceiling. Maybe he's just trying to sell Atmos-enabled speakers :-). I noticed that for many movies, even if Atmos is employed, the effect is subtle most of the time. For example La La Land really doesn't need Atmos and I was surprised Incredibles 2, though sounding great in the surround presentation, wasn't using the Atmos channels as much as I expected even in the action sequences.

I'll have to play more with the angulation, speaker levels, toe-in/out and maybe try different crossover settings to see if I can optimize the sound further. I still have a number of other Atmos-encoded movies to check out (curious about Black Panther, Infinity War, the upcoming Mission Impossible: Fallout). Guys and gals, let me know if you've come across some good movies and scenes useful as "demo material" when friends come over!

I think live concert Blu-Rays could be very nice using Atmos or DTS:X. It'll help capture the ambiance of the performance space even more. On Amazon I see Brooka Shade's Galvany Street Atmos Mix is available; might give it a try for the audio experience...

Consumers like to debate format wars but as far as I can tell, there's no such war between TrueHD/Atmos and DTS:X at this point (nor needs be!). Both are capable of lossless, high resolution, object-based audio. Most good receivers these days support both. What we can say is that DTS:X has provisions to be more speaker placement agnostic and the Multi Dimensional Audio (MDA) format supposedly is open and royalty-free. Ultimately I guess we'll see which of the two is supported by the most studios and have the most number of shows released for the home consumer.

Hope you're all having a great time enjoying the music and movies :-).

4 comments:

  1. Great review, thanks for the insights.

    ReplyDelete
  2. Nice article.

    I went from 5.1.2 to 5.1.4 and still remain underwhelmed by Atmos outside the demo clips. What I mean by underwhelmed is that when I'm watching a movie I don't really notice the difference between native and upmixing. It is nice having a layer above your head for immersion but the only real content I've watched that made real good use of it is Daredevil on Netflix. That show made use of the atmos height channels to project a voice inside a characters head. It was particularly impressive.

    I was forced to watch Mission Impossible Fallout via upmixer as my Panasonic UHD player doesn't play over the network via Bitstream. It was an amazing sound track even with Upmixing. I doubt the Atmos track would add anything dramatically different unless you were A vs B in it.

    ReplyDelete
  3. "Now we add the 10-year old Energy C100 speaker pair to the receiver amp. The Energy sounds similar in tonality to the rear Paradigm Studio 80's. Perhaps not surprising since both are Canadian companies using research from the National Research Council and are not far apart in age."
    I haven't had such a good laugh in a while. :)
    BTW I do enjoy glancing at your measurements but I am somewhat puzzled as to why you started this blog. You invest money in new sound cards/adc and not in room treatment? $15 google chromecast audio sounds to you the same as logitech transporter so why not to spend money where you can measure a real difference such as room treatment? You room is dying for acoustic treatment while you are chasing minuscule differences in digital filters you cannot hear. Why waste money on monoblock amplifiers if it is irrelevant? I mean have you compared them to $300 receiver in properly setup double blind test? ;). Please continue on doing what you are doing as I do enjoy looking at graphs at someone's else expense, though I do not understand your motivation.

    ReplyDelete
    Replies
    1. Good comment Museatex,
      Let's talk about this in a blog post ahead...

      Delete