Auro-3D "Auro-Matic": Preset and Strength settings and effects. (Brief look at Dolby Surround Upmixer and DTS Neural:X.)

A number of months back, I talked about the nVidia Shield TV Pro box and explored Auro-3D which is another technology for reproduction of immersive sound with height layers. As noted in that post, I've seen comments over time about the use of Auro-3D's "Auro-Matic" surround upmixing algorithm for music playback as being the preferred setting compared to Dolby Surround Upmixer [DSU], and DTS Neural:X being the other two common options in modern AV receivers. As a reminder, while Auro-3D could use object-oriented tools in production, it is delivered as channel-based playback rather than the hybrid channel/object systems like Dolby Atmos and DTS:X.

Having now owned my Integra DRX-8.4 for about a year and listening for myself, I concur that for 2.0 and 5.1 upmixing, I tend to prefer using Auro-Matic to expand the 3D image into the height and surround channels. For most music material, Auro's method is noticeably "fuller" in presentation without sounding harsh. It also has more options as we can see in the settings menu for Onkyo and Integra receivers that have this feature:

Auro Settings section from Onkyo/Integra manual.

For non-Auro encoded content, "Auro-Matic" will apply its DSP in whatever "Preset" you choose (Small / Medium / Large / Movie / Speech) at whatever "Strength" you want to apply from 0 (disabled) to 15 maximum. Default for Onkyo and Integra receivers is quite high at 12.

I must admit that the "Preset" is a bit nebulous and doesn't seem to be well-described in Auro's documentation. The easiest test I could think of doing was just to check whether these Presets affected the frequency response when playing back a single channel (let's test the right channel):

As you can see, I've also included the dotted lines for Dolby Surround Upmixer (DSU) and DTS Neural:X both appearing almost identical across the spectrum. I've set the Dirac Live room correction target more or less following the Dolby Atmos Music Target Curve previously discussed. Clearly Auro-Matic has changed the frequency response of this single speaker, noticeably only in the bass frequencies. At Strength level 8, we see that Auto-Matic accentuates the bass up to around 150Hz. Whether I set it to Small, Large, Movie, etc. doesn't change the frequency response in any significant way. As such, we can assume then that these Presets are more likely affecting temporal characteristics like reverb, perhaps delay added to the height, center, rear channels when doing the upmixing.

We can see a hint of these temporal changes when we plot out the reverberation time:

In the graph above, I'm showing the Room EQ Wizard T30 results (a measure of RT60 or duration in milliseconds for sound energy to decrease by 60dB, calculated based on the slope of the Schroeder curve between -5dB and -35dB derived from the impulse response). Notice that Auro-Matic dissociates from DSU and Neural:X in the low frequencies significantly below 150Hz. We can see a nice correlation with "Large" adding more total reverb while "Small" least (shades of red). "Movie" setting is similar to "Large" (maybe a little less), and likewise "Medium" and "Speech" look similar. While I did not formally test the right height and bed speakers separately, my suspicion is that this is what's added to the height channel to create that subjective sense of space.

I've played with the Presets over the last year and generally have been happy with the default "Medium" setting in my room. Since it's only the bass frequencies affected, not all material will show a clear audible difference.

Strength 8 was enough to clearly accentuate the bass, let's take a look at how much bass is added as we vary the Strength from 0 up to maximum of 15:

I measured the frequency response starting from Strength 0 and even values except for level 15 (highest) and level 7 (dotted). Notice I put the cursor at 35Hz and 80dB SPL for reference. At the cursor's 35Hz, we can see the relative SPL levels of the curves in the legend at the bottom of the graph. This estimates the bass gain relative to the Dolby Surround Upmixer; lets round to the nearest 0.5dB):

Strength 0: 0dB      Strength 2: +0.5dB   Strength 4: +1dB
Strength 6: +2dB     Strength 7: +2.5dB   Strength 8: +3.5dB
Strength 10: +5.5dB  Strength 12: +6.5dB  Strength 14: +7.5dB
Strength 15: +8dB

While the bass boost is most obvious and audible, notice that across the higher spectrum, Auro-Matic does run a little louder than DSU - for example at 2.5kHz, Strength 15 adds about +2.5dB. I suspect what's happening is that as we increase the strength of the mix, the corresponding height speaker is louder and this adds to the total amplitude. When sitting in the listening position, as Strength increases, it sounds like the right speaker stretches upwards - spatially taller and a bit louder.

Therefore, let's run those calculations above but correct the bass gain by subtracting out the high frequency increase; let's subtract out the average of the gains at 1kHz and 4kHz, rounded to the nearest 0.5dB:

Auro-Matic corrected bass gain:
Strength 0: 0dB     Strength 2: +0.5dB   Strength 4: +0.5dB
Strength 6: +1.5dB  Strength 7: +1.5dB   Strength 8: +2dB
Strength 10: +3dB   Strength 12: +3.5dB  Strength 14: +4dB
Strength 15: +4dB

While I enjoy exciting bass, there's a limit for me of around +3dB which is the gain I found at Strength 10. I'll keep my Strength setting below 10 in typical listening.

Something I found interesting in Auro's whitepaper on Auro-Matic is this quote:

"Unlike most other upmixing technologies on the market, Auro-Matic® keeps the original channels of the input source intact instead of steering parts of it to particular speakers. The sophisticated adaptive processing rather places the original recording in a 3-dimensional environment that matches the acoustical properties of the recording, creating an immersive sound field that creates the impression of being in the same environment as where the recording was taking place."

This is an interesting way to upmix content if it's as claimed. From the days of Dolby Surround, the technology has been "steering" content to the speakers that best represent the location where the sound "should be" coming from. The easiest example is when we play a mono recording representing frequency, level and phase-identical content to each front stereo speaker to create the phantom center. With DSU and Neural:X, instead of a phantom center, most of that sound should be "steered" to the actual center channel if present while the two front (bed and height) channels go quiet.

Here's a demonstration of this. Let's record a sine sweep as 2.0 mono (L+R playing simultaneously) using Dolby Surround Upmixer and DTS Neural:X showing the relative output level when I disconnect either the center speaker or the left and right bed + height speakers:

Notice the upper-mid accentuation in DTS Neural:X between 2-5kHz. Could sound a bit harsh depending on your system. Maybe it could balance out speakers with the "BBC Dip"? (I've touched on this in the past as well.)

As you can see, indeed, both DSU and Neural:X are pushing the mono sweep into the center channel only. Note that I have my subwoofer crossed over below 100Hz which is why you see that low-bass is still there down to 20Hz regardless of center or L+R (+ heights) speakers playing.

Now check out what happens with Auro-3D "Auro-Matic":

Interesting! Most of the mono content is still "steered" towards the center channel but when the center channel is disconnected, the L+R bed plus height speakers still have content but with a V-shaped mid-range dip applied, centered around 1.5kHz. Make sure you have a good center channel as the very important midrange is primarily delivered by this speaker. As such, I'm not sure about taking Auro's literature literally when it says "Auro-Matic® keeps the original channels of the input source intact instead of steering parts of it to particular speakers"! Regardless, clearly what's happening is quite different from DSU and DTS Neural:X.

For completeness, Dolby Surround Upmixer has a setting called "Center Spread" which will, as the name implies, spread the central image across into the front stereo speakers when upmixing 2-channel content. Here's what it looks like when turned on:

As you can see, with "Center Spread", the combined amplitude of the L+R front speakers (+ heights) is now significantly louder than the center speaker. In doing so, you'll hear a wider front center image. This function does not do anything when the receiver is sent multichannel content.

Auro-3D: from data to playback.

Summary...

Auro-3D Auro-Matic is an interesting DSP technique for upmixing everything from mono to 2.0 to surround music when you have a multichannel system with surround and height/ceiling channels. It offers "Preset" and "Strength" settings for tweaking.

Here's are some take-home points based on my testing:

1. There is accentuation of bass up to 150Hz correlating with higher Strength values. Since I'm using Dirac Live with the Dolby Atmos Music Target curve, I'd prefer not to add too much bass. Using a Strength setting of 10 can increase bass by +3dB which is my threshold. I've been using 8 as my default, at most I'll push to 9 if I want a bit more height effect.

In my opinion, the default value of 12 for Onkyo/Integra systems sounds excessively bass-heavy. Of interest, I see that the Marantz AV-7704 defaults at 10 which is more reasonable assuming the bass gain is similar to the Onkyo/Integras.

2. Preset like "Small", "Medium", "Large" settings appear to be related to amount of reverb added to the bass frequencies below ~150Hz. Reverb time is highest with "Large" and "Movie" presets, "Speech" is somewhere between "Small" and "Medium". As you can imagine, which Preset you choose will be highly subjective and will depend on your room as well. In my listening, I found the effect is relatively subtle and the "Medium" preset sounds good without any objectionable bass overhang or muddiness.

3. It's interesting to see that Auro-Matic "steers" sound quite differently than Dolby Surround Upmixer or DTS Neural:X. Clearly, the mono sweep shows that it still will shift most content to the center channel but maintains significant front bed and height R+L content with a midrange dip. How well Auro-Matic behaves as it dynamically extracts surround and height channels with its claimed proprietary "sophisticated adaptive processing", you'll have to listen for yourself to determine if the psychoacoustic effect is something you enjoy.

4. Want a wider front image for 2-channel upmixing? Give Dolby Surround Upmixer Center Spread a try. This will not steer correlated content to just the center speaker but spreads out the sound to the left and right front speakers plus use the center channel to fill in the middle. There was discontent awhile back in 2020 when Dolby tried to remove this feature.

[There was also a time when Dolby wanted to restrict upmixing choice which thankfully did not go far. If customers want to use a competitor's product more, then the company needs to innovate and make theirs better, not restrict freedom. All proprietary technologies IMO should respect consumer choice; this of course goes for Lenbrook/MQA as well regarding their ongoing obsession with making money off low tap-length minimum phase filtering as if audibly meaningful! See recent post on QRONO d2a and Spotify Lossless. 🤣]

There was a time when I was more of a purist when it came to 2-channel playback. These days, I've been impressed by the sound quality achievable with DSP upmixing to my surround and height channels and more often than not, I've been using the Auro-Matic feature. Combined with DSP room correction, the results are highly enjoyable.

[It should go without saying that listening for the differences between DSP upmixers and trying the different parameter within Auro-Matic is way more interesting than pretending that one can hear a difference between interconnects, power cables, various component footers, etc. Even digital streamers and decent "perceptibly perfect" DACs which I believe nobody would notice a difference in controlled blind listening. 😯]

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

I mentioned in the comments last time, I really like the Sinners (2025) movie soundtrack. What an interesting, eclectic, mix of blues, Irish folk, pop, electronic sounds!

Have a listen to Miles Caton on "I Lied To You" using Auro-Matic upmixing:

Unfortunately, there's no native multichannel/Atmos version of this album. However, the multichannel upmixed playback sounds great starting off with straightforward front-and-center blues guitar and vocals, then things get wild from 1:20! 😯

For those who have a multichannel copy of Enigma's "Mea Culpa, Pt. 2 [Fading Shades Mix]" off The Platinum Collection (2009):

Given all that's going on in the the world, as the lady (Sandra) says: "Turn off the light. Take a deep breath. And relax." Have a listen to that upmixed with Auro-Matic for some surreal immersion. 😌

Hope you're enjoying the music, audiophiles. Off to the historic lands of the Habsburgs.