Suggestions for Updating Projection Room - LOW COST

[Steve Guttag]

Agreed. Furthermore, I would never put Center channel on the same amplifier as Left or Right (or LC or RC, in a ATMOS/70mm system). There is often a means to use remaining channels to keep a theatre up while a channel has failed but that is sabotaged if the back up in the same failed amplifier. This is true for multi-way systems as well, I'll never let any part of Center be on any part of the other stage speakers. The only channel Center can share with is the subwoofer (LFE).

I do have ONE exception coming up...all stage speakers are passive and the "stage" is being run with a single 4-channel amplifier. In that case, there is a full standby amplifier, at the ready to go on line with a moment's notice (already mounted in the rack with only a plug-in speaker connector needed to make the switch.

[Sean McKinnon]

I am also not on the single amplifier for an entire auditorium bandwagon. Like Steve I will typically not put any component of Center channel on an amp with another screen channel where possible. I will use the QSC 8 Channel amplifiers for surrounds. These days I will discreetly amplify all my surrounds or at most parallel a max of 2 on one channel.

I have been to multiple Dolby Atmos training classes where they have showed off the 32 channel amp but I personally would not use it for anything but surrounds. I would keep the screen channels separate.

[Steve Guttag]

Since the Dolby DMA comes in 16, 24 and 32 channel varieties. If I were using it, I would tend towards two amplifiers per auditorium to ensure, again, Center doesn't share the same amplifier with Left/Right. Another possible alternative is to again, have a spare on site...but swapping that beast in is more tricky than say a QSC DPA-Q. One can set up QSC DPA-Qs with Dynamic Pairing so the replacement amp, once plugged into the same port via RJ45 takes on the original amplifier's programming and the speaker outs are on a single plug in block. In my example in the post above, the amp can reside in the rack and, in the design, always have the signals being fed to it so all the user has to do is physically unplug the speaker block and plug it into the spare...bam...online. The DMA would require numerous speaker blocks to be transferred over and a crane to move the bad amp out and the new amp in.

That said, from simplicity...the CP950/DMA combo is about as fast an installation as you can get...plug a single Ethernet Cable in...no switches are even needed.

[Sean McKinnon]

I haven't had a customer that wanted/needed it yet but I have thought it would be trivial to provide an automatic fail-over for the QSC Q-Sys amps so that if an amp goes off-line it triggers a set of relays to switch the speaker lines from one amplifier to a back up that's always running much like the product QSC sells for the DCA series amplifiers. I suppose you could even take it a step further and have one amplifier that can have any other amplifiers output routed to it in the Q-Sys design and then multiple relays to connect and disconnect different loudspeakers an needed. It would take some sussing out to ensure that no two separate sets of speakers get connected to the back up amp at the same time but should be totally doable.

[Steve Guttag]

Yeah...you'd need to create a "Bussed" amplifier output and then a "1 of" logic circuit to keep it safe. The problem is...so now you put in a bunch of high current relays with at least 4 contacts or two 2-contact relays in parallel and have doubled your connection count and your cost for such an item is probably not going to be cheap as it would be custom and low-volume. It is certainly doable and also certainly doable from within the Q-SYS environment as there is an abundance of GPIO and status.

I'd like to hope that things are not so unreliable that we have to prepare for any eventuality. Having a back up to Center, for me is sufficient and the effort to move the speaker connector is minimal and doesn't add into the connection count.

[Sean McKinnon]

Yeah... I haven't had any failures with the Q-Sys amps anyway I could only think of a VERY high profile mission critical type situation that would ever call for something like this.

[Harold Hallikainen]

My concern is that like Steve, I think adding this failover in an attempt to make the system more reliable would actually make it less reliable. Complexity makes it less reliable. The fairly common practice of replacing failed center with L/R or failed L/R with center seems much simpler (since it's just software in the DSP).

[Steve Guttag]

Sean...I have a couple such mission critical screening rooms where most every show is a 1-off. The projectionist (high enough profile that there is one of those for every show too) can not only route around any stage speaker that is not working properly (for whatever reason) via the TSC-7 in a matter of seconds with just 2-taps (one to get to bypass, one to identify the errant channel). They also have the spare amp in the rack, already getting Center channel via QLAN and would only have to move the one speaker connector between amps to get that one back on line. This on top of redundant cores as well as redundant networks. Even if the ATMOS link were to drop, they have a DCIO-H to carry the show. They'd have to lose a LOT to drop a show and I think that is sufficient redundancy. I'd give greater odds that a projector could fail and one of my high-profile rooms has two projectors too with two servers...etc. They don't just have a spare tire...they have the whole car...so to speak.

However, I think we've REALLY drifted off the "low-cost" theme of this thread.

[Carsten Kurz]

We have amps that work day-in-day out since 1996. So I guess it is possible to build amplifiers that fail so rarely that any explicit countermeasures (except quick replacement/spares) are ridiculous or even counterproductive.

What I would like to see is equipment/procedures that allow to check all speakers daily. Speakers fail more often than amps. And for surrounds, or individual components of front systems, they are usually not easy to track down during daily operations, and patrons reports are typically vague or non-dependable. Sometimes, defective speakers make obvious noises, but if they simply fail, this may go unnoticed for a while.

In 35mm times, we detected multiple blown LF drivers in our JBLs only when we disconnected and tested every individual speaker. We were quite embarrassed when we noticed how many actually had failed.

I think Harold at one time thought about audio detection thresholds of the LSS-200 and in how far these would be able to detect individual failed speakers. Don't know if this evolved into standard monitoring procedures.

It gets easier of course the less you use parallel/serial (BOOO!) surround speaker wiring.

- Carsten

[Marcel Birgelen]

In a modern Atmos/MDA setup, where every speaker usually maps to their own distinct output channel, I guess running a "self test" once a day or at least once a week should be pretty straight-forward. It would also be great if e.g. Dolby could implement some easily accessible kind of "emergency routing" in case of a single-speaker failure, where the speakers next to the failed speaker take over temporarily for the missing speaker/output channel.

[Carsten Kurz]

I think both automatic testing/fault detection as well as dynamic remapping was mentioned in some of the early ATMOS white papers. I think both didn't make it into release software so far. Though I think manual remapping is possible.

[Harold Hallikainen]

Many QSC amplifiers include continuous speaker impedance measurement, so loss of a speaker can be detected very quickly on the amplifier. As Carsten mentioned, The LSS-200 can also detect failed speakers. This is especially the case if the speaker can be addressed individually (such as screen speakers). The LSS-200 can tell a CP-850 to put pink noise on an individual speaker, so failed speakers can also be detected in those systems. On surround arrays, it depends on how tightly you set the limits. If the limits are set +/- 3 dB from the original measured level, ideally the failure of half the speakers can be detected (though some are closer to the LSS than others, so the failure of a far speaker would not be detected due to the audio from a close one). If I calculated right, a margin of +/- 1 dB would detect the failure of 1/4 of the speakers. There is some variation in measured level with temperature and humidity, and the close/far issue would result in a fair amount of masking of the loss of a distant surround speaker. But, loss of a screen speaker is quite reliable since they are individually addressable.

In immersive audio systems, remapping to work around failed speakers SHOULD be fairly straightforward. There is a table describing the location of each speaker. This is used for object panning but COULD also be used for "panning" of bed channels to the available speakers. If a speaker is missing, the system works around it as though it was never there. Systems also provide for "zone gains." A zone gain of 0 indicates that no audio is directed to speakers in that zone. The system makes up for it with the remaining speakers maintaining position and SPL.

Harold

[Steve Guttag]

It should be noted on the LSS test...nothing requires a full-band test of a speaker. You could generate PN or tone of any bandwidth to check the various sections of the speaker. In fact, the available LSS200 tests include things like 4-way so, in effect each section is tested.

As for surrounds, I will generally parallel the speakers and do not run a separate amp per speaker. I can think of many more things to spend money on that will actually have a noticeable improvement. Most don't seem to realize that a DCA 1222 can run safely to 1.6Ω so 5 speakers may be run in parallel. This practice does not help identifying a single speaker failure. That said, I can't recall having much in the way of stage speaker failure unless the system was poorly designed. Historically, we've used Altec (A5 or A4), JBL 4675 or 4670 with some 4673 and, in more recent years 4722...they have not been a failure prone speaker (particularly the 4675 and 4670...they are probably the most reliable speakers). We've had zero QSC failures too...all 400 series. Now, on surrounds, that is a different story. The 8330/8330A have not fared as well in the digital world. The midrange speakers seem prone to failure. The 8340/8340A surrounds have a foam compliance (surround) that will eventually rot away. I'm currently seeking a "preferred" surround speaker. I really like the QSC SR-1290 and SR-1590 but they are going to be more pricey than most cinemas will put up with and are only 4Ω speakers...hence they go to immersive audio systems. The JBL 9300/9310 are decent enough sounding (though really lacking on the bass) but boy are they ugly (IMO); there are some rooms I couldn't put them in on aesthetic reasons alone.

[Sean McKinnon]

I agree with all your comments, like I said before it was mostly an academic exercise however you guys would probably not believe some of the requirements of specialty projects we get hired to do, definitely not your typical multiplex stuff for definitely not your typical multiplex customer.

I have had almost zero issues with QSC amplifiers and loudspeakers in the past 3-5 years that I can think of, maybe one failed DCA amplifier that was >5 years old.