Having connected three home theater A/V receivers to expand my Dolby Atmos surround sound into a crazily convoluted 7.1.6 channel system, the next step was actually making it work. Sadly, that proved to have some unexpected challenges.
Because consumer home theater hardware (the kind we humble 99-Percenters can afford, anyway) tops out at a speaker limit of 7.1.4 channels for Dolby Atmos and DTS:X, adding extra channels in the height layer (which both immersive sound formats should at least theoretically support) turns out to be a very involved Do-It-Yourself project. It’s not impossible, but it is complicated and, as I found, difficult to get right.
In Part 1 of this article, I explained the concept behind using two additional A/V receivers in the downstream signal chain to derive new Top Middle speaker positions between the Top Front and Top Rear channels that the primary Atmos/DTS:X receiver decodes from a Blu-ray disc.
Although the speaker wire and cable connections required to do this can get confusing, the theory behind the process is pretty logical and straightforward. Each downstream AVR receives what it believes to be a stereo signal (actually comprised of one part Top Front and one part Top Rear), and then uses Dolby ProLogic II to extract a center channel between them from any mono sound information common to both. This is the same method by which ProLogic II in a traditional 5.1 system can pull movie dialogue from a stereo soundtrack to form a center channel.
After spending a very long time hooking everything up, you can use each receiver’s internal test tones to validate that all the speakers are working. The primary AVR should send tones to all 7 ground speakers and 4 of the heights (Top Front and Top Rear), but it cannot check the Top Middles. For that, you need to use each of the secondary AVR’s Left/Center/Right tones.
This will get you as far as confirming that every channel plays sound, but it will not prove that the 7.1.6 extraction process is actually working. We’ll get to that in a minute. First, we need to set some preliminary channel distances and levels.
When operating multiple AVRs together, all of the downstream receivers should be set for a reference level 0 dB master volume and left there. You will never touch the master volume on those receivers again. Volume control for a movie soundtrack will be handled by the primary AVR. When you turn your master volume up or down on that receiver, it will change the signal level to the Top Front and Top Rear channels through its analog pre-out connections.
You should also lower the speaker distances to their minimum setting (zero if possible) in each of the downstream AVRs. Speaker distance will also be calculated by the primary AVR. Increasing that setting affects the amount of signal delay the receiver applies, and you don’t want to double-up delays in both the primary and secondary AVRs.
However, you will need to calibrate your individual channel levels in each of the secondary AVRs. For that, use each receiver’s L/C/R tones and a sound level meter. One receiver will do the Top Front, Top Middle and Top Rear speakers on the left side, and the other receiver will do the right side.
Next, make sure the primary Atmos AVR is configured for 7.1.4 format with Top Front and Top Rear height channels. For the first wave of calibration, you want to set the speaker distances and levels as you normally would if you only had the one receiver. Most people will use Audyssey or whatever other auto-calibration program is built into the receiver (YPAO, MCACC, etc.). That should be fine, as far as levels and distances go. (Room correction EQ features are another matter, as we’ll learn shortly). Note that, in this process, it is normal for these auto-calibration programs to calculate very long speaker distances for the Top Front and Top Rear channels, regardless of how far those speakers actually are from the listening position. That happens because the ProLogic II processing in the secondary AVRs inherently adds signal delays that the primary AVR will assume are caused by far speaker distances. Don’t be concerned about this yet.
At this point, all of your speakers should be active and set for a uniform volume level. So far, so good, right?
Here’s where the problems start.
Why Doesn’t It Work?
The one thing we haven’t done yet is prove that the downstream A/V receivers can properly extract a Top Middle channel from the Top Front and Top Rear. The only way to validate that for certain is to use discrete Dolby Atmos test tones. Unfortunately, at present, the only source for test tones with 6 height channels is the Dolby Atmos demo disc that was handed out at the CEDIA trade conference in September 2015. This is not something you can pick up at Best Buy or order from Amazon. The disc was not intended for retail sale. I find it ridiculous that these important calibration tools are not available to the general public, and I encourage Dolby to expand their distribution. (Currently, VUDU streams a handful of Atmos trailers for free. I see no reason why Atmos test tones couldn’t be offered that way as well.)
I’m not going to get into the particulars here, but let’s say that you obtain a copy of the Atmos demo disc anyway (specifically, the September 2015 version). In the calibration section is a chapter for discrete 9.1.6 test tones. Assuming you don’t use Wide channels on the ground level (most people don’t), your receiver will fold the signal for those channels into the front Left and Right speakers. For our purposes here, we’re focusing on the six height channels: Top Front, Top Middle and Top Rear.
If your Atmos receiver is configured in a 7.1.4 layout as described above, it should redirect the test tones for the Top Middle channels on this chapter to evenly disburse them to both the Top Front and Top Rear speakers. If indeed you really only had those specific speakers, this would cause the sound (which is a mono signal split equally to the two channels) to phantom image between the speakers, tricking your brain into hearing the sound somewhere between the two locations. By inserting another AVR at this stage of the signal chain, however, ProLogic II processing should instead extract that mono sound from the left and right (Top Front and Top Rear) and pull it into its newly-created center (Top Middle) channel. When everything works properly, the test tone for a Top Front speaker should play from that Top Front Speaker, the test tone for Top Middle should play from that Top Middle speaker, and the test tone for Top Rear should play from that Top Rear speaker. It should sound like you have six accurate and discrete channels of audio.
Sadly, after connecting and configuring everything (and double-checking it all), my 7.1.6 system failed the test tone challenge. Although the Top Front tones played from the correct speaker, and the Top Rear tones also played from the correct speaker, the Top Middle tones were not discrete to the Top Middle speaker. Instead, that tone played from all three speakers on its side of the room (TF, TM and TR). The same thing happened on both the left side and the right side.
For quite a while, I was utterly stumped as to why this didn’t work. Based on the theory, it should have. After a lot of troubleshooting and consulting with other members at AVSForum, I eventually realized that I had overlooked something of critical importance.
In order for the ProLogic II processor to extract a middle channel, it must first be able to identify the mono sound information common to both the left (TF) and right (TR) channels. The only way that can happen is if the incoming signals to both of those channels are absolutely identical to one another. That means that the primary Atmos AVR must be programmed with, for starters, the exact same speaker distance, trim (channel level), and bass crossover settings. If any one of those parameters differs between the front and the rear channels, the secondary AVR interprets the signal as stereo and will not extract it to the center.
Matching up the speaker distance setting isn’t a big deal. Some fudging there won’t hurt anything. The bass crossover also won’t be a problem. If you have a subwoofer, all of your speakers should be crossed over at 80 Hz or higher anyway. Err on the side of bumping the crossover setting to the higher number if those speakers differed before.
Channel trim may seem like a bigger obstacle (you don’t want one speaker to be too loud or too low when playing a movie soundtrack), but you can compensate for that in the downstream AVRs. For example, when I did my original calibration, Audyssey set my Top Front Right speaker to +3 dB and my Top Rear Right speaker to +5 dB. For this project, I matched those both at +5 dB, which means that the Top Front channel was 2 dB too loud. To offset that inappropriate gain, all I had to do was set the same channel in the secondary AVR to -2 dB and everything sounded right again. Doing this will not affect the center channel extraction. The key point is that the ProLogic II processor must believe that it’s receiving a mono signal, not a stereo signal. The trims are applied after the decoding.
We’re not quite there yet. Even after matching the speaker distance, crossover and channel trim levels in the primary AVR, I still couldn’t get the Top Middle test tone to play from just the Top Middle speaker. The secondary AVR was still interpreting the incoming signal as stereo, not mono. I missed something else.
During my calibration, I had turned on Audyssey room correction and Dynamic EQ. These are features that I like and expected to continue to use. Unfortunately, they play havoc with the center channel extraction in the secondary AVRs.
As part of the auto-calibration process, Audyssey MultEQ XT32 (the version in my receiver) measures the room characteristics from every speaker direction and applies unique EQ to each speaker. This cannot be adjusted afterwards. Audyssey’s room correction is an on/off, all-or-nothing decision. There’s no way to make the EQ for a Top Front channel match a Top Rear channel. As such, the downstream AVR continued to detect differences between these channels and read the signal as stereo. It could therefore not perform the channel extraction correctly.
At this point, there was nothing left to do except turn off all Audyssey processing features. Once I did that, the secondary AVR finally saw a mono signal and did its job properly. ProLogic II pulled the mono test tone into the Top Middle speaker and anchored it there exclusively. Every test tone in the 7.1.6 system was fully discrete and accurate.
The trade-off to this is that I can no longer use Audyssey room correction or Dynamic EQ. I’m disappointed by that. Without any EQ at all, my room sounds a little flat. My receiver has a manual EQ feature that will allow me to copy the Audyssey “curve” as a starting point, but the EQ controls are rather rudimentary and it doesn’t copy all of Audyssey’s proprietary filter algorithms. Still, it’s the best I have at the moment. I turned that on for the ground-level speakers in my system, but left all the height channels with EQ settings of 0 (so they match each other and send an identical signal). So far, things in my room sound pretty good, but I’ll need more critical listening time to evaluate whether Audyssey is actually expendable or not.
Ironically, the process of going through this whole ordeal with the test tones served to enlighten me as to how inadequate and inferior my former “Zatmos” configuration was. It turns out that I had the same problem there regarding ProLogic II processing not working correctly, but I didn’t have discrete test tones to validate it and never realized the extent of the problem. Because I heard some sound from my extra matrixed channels, I just assumed that everything was working. Only now do I understand why I’d been left a little underwhelmed and unsatisfied with the results I heard at the time.
Is the new system really a significant improvement? We’ll talk about that in the next installment.