Mixed sub models

[Jon Dent]

We currently have a single JBL 4642a in our theater being driven by a bridged DCA1422. While the power availability is sufficient for running at '7', the sub itself looks like it would be right at the limits of it's abilities, as to hit 113 dB at 2/3rd of 50ft, it would need to produce peaks of 133dB (roughly). That's under it's peak limit of 136dB but just barely, and its over the 131dB 'continuous' rating). We don't run at that volume, ever, but in the interests of having a system that could do it comfortably if needed, we're looking to add a second sub.

Instinct says 'go with another 4642a' but our installer is suggesting a Dolby SLS sub instead (I'm assuming the CS218XL), saying that the different specs won't matter (and they do look very similar on paper) and it'll save us some cash.

What the opinion of the pro gallery here?

[Steve Guttag]

In a 50-foot room, you'd need about 2600-watts for a 4642A subwoofer...it can handle 1200-watts continuous. So, no. You definitely want a second one. With the second one, your power requirements drop to 1300-watts and you'd have 2400-watts of capacity Your DCA 2422 (presuming a typo) would have just enough power too (providing not too much on the EQ.

As to mixing the subs...NO!!!! Don't mix things that need to play as a team.

What do you really know about the SLS thing? I don't see any published response. It only has "usable range" (whatever that means...we'll presume -10dB) down to 30Hz. The JBL is ported at 25Hz and is "usable" down to 20Hz. Without a response you have no way of knowing how that 102db 1w/1m spec came about. They rate the subwoofer from 30Hz-300Hz...their sensitivity spec is bound to be for non-subwoofer duties. Who cares if a subwoofer is sensitive above 80Hz? That isn't where you want the response to be particularly good.

Here is another thing, JBL makes some really good drivers. The 2241H that is in the 4642A has a long track record of doing very well. What noname thing is in the SLS? I'm not even telling you that the 4642A is the greatest subwoofer (it isn't) but I wouldn't pair it up with a different subwoofer...even if a different JBL (say a 5749...not listed as a subwoofer but it is better, uses the same box but with a 2242H driver or even the 5628). You want the subwoofer channel operating as a team. Keep the matched. About the only thing you could match to the 4642A, other than another 4642A is a pair of 4641 (same driver, essentially the same response, same volume, same porting). I would strongly question why your installer would propose dissimilar equipment.

[Jon Dent]

If you don't mind filling me in: where is the 2600 watts coming from? I admit I'm not well versed in this stuff, I was just referencing the infamous 1994 Dolby guide that showed a 4642 needing about 1000 watts to hit 113dB in a 50ft room.

Oh and I meant DCA 1622.

Capture.PNG

[Steve Guttag]

It comes from math and experience. Dolby's nomograph is off by about 3dB. Note, when that document came out, it was a shock to many people because that was STILL about twice the expectation of what most were used to. Furthermore, for digital cinema, the targets are 2dB higher. And even furthermore, dB SPL is nonsense for a subwoofer (band limited) level. When a full range change has an 85dB level, it is dBc, which defines the band. Furthermore, the frequency response is known (based on SMPTE 202)...as such, when a studio sets their level up for 85dBc, it should pretty closely match to a cinema. And, if one is using ⅓-octave on their RTA...the pass band between 100Hz and 2KHz should have all of those levels averaging around 72dB each.

A subwoofer, naturally, will have a frequency response with just a portion of a "C" weighted range, which is why one doesn't measure 95dBc on a subwoofer and, more typically, might measure 89-91dBc...it all depends on the range of the subwoofer...but you wouldn't want to raise a subwoofer's output to meet a single number like dB SPL because it would have less capable (narrower bandwidth) subwoofers being turned up while wider bandwidth subwoofers would be turned down. This is why they are set 10dB higher of in-band level than the in-band level of the Center speaker. So, again, in ⅓-octave analyzers, those ⅓ octave bands will be bopping around 82dB on the RTA. If you do the logarithmic math, you'll see that a -20dBFS pink noise generator will not and should not measure 95dBc (or 93dBc).

You should also note, the JBL 4642 is not the same as the 4642A. It was an unfortunately naming scheme by JBL (as was the 4645B versus the 4645...they were, in no way, descendants of the other).

The DCA-1622 will not be sufficient to play at reference in a 50-foot deep room, even with two 4642A. You only get 500W/CH into 4Ω. You'll be 1.5dB lite. Not to worry. If you stay below 6.5 on the fader, that will get you the 1.5dB.

BTW...Dolby's current math is a closer match to my own...using DARDT, they too poo-poo a single system...even using a DCA 1622 bridged.

Screen Shot 2022-04-05 at 8.33.02 PM.png

Moving up to two units and running the DCA 1622 in stereo: The amp becomes the limiting factor:

Screen Shot 2022-04-05 at 8.39.31 PM.png

However, dual cabinets, with a DCA 2422 (or two DCA 1622 or two DCA 1222 bridged with each amp driving one cabinet, you have the level and can deal with some EQ and a little bass boost on the bottom, if necessary):

Screen Shot 2022-04-05 at 8.40.44 PM.png

My numbers and Dolby's (current) numbers do not exactly match but they are close enough. There are unknowns in the system like room gain, how well coupled are the subwoofers to the floor, any sort of baffle wall/bafflete, clustered, position relative to side walls/center...etc.

[Jon Dent]

Sub isn't on the floor due to occasional water seepage (front of the theater is a few feet below ground). It's on a scaffold about 4ft off the floor and at roughly the 2/3rds point between the center and right screen channels. Side wall is maybe another three feet past the right channel. There's very little clearance behind the sub to the wall (maybe a few inches). No baffles.

The current level was set with an RTA and multiplexer (thanks Mark!). The pink noise signal from the CPA750 had information all the way up above 200Hz but I ignored everything over 120Hz and used the bands under that to set the level. Minor shave at 40Hz (maybe 2dB) with a narrow Q to correct a room bump.

And as far as the installer..... This isn't the first time I've been skeptical of their suggestions, but our owner is nothing if not loyal to his suppliers. Because of that I still have to go to them first. I'd rather just do it all myself but I don't have the time or money to get all the vendor trainings and certifications I'm sure I'd need.

[Pete Naples]

Absolutely don't mix subs, unless you're using a processor which can address them individually and discretely (you would need more amplifiers to do this too)

And even then, I wouldn't advise it. I'd suggest how they interact in the room will be less than predictable.

[Steve Guttag]

I think it's great that the theatre owner is loyal to their supplier. I just don't think it is wise to mix certain equipment, like subwoofers that should form a single (larger) speaker. You'll always have a different response out of the two subwoofers so they will sum in a non-favorable way. I'm also serious that JBL makes a good subwoofer (well, the drivers...their cabinetry is more "delicate" than others). The 4642A is a solid performer with a very well known track record.

Having the subwoofer raised 4-feet is unfortunate. This moves it closer to a 4π space. Normally, having the subwoofer at or near the ground picks up some free efficiency. Though your subwoofer may be close to the rear wall behind the screen, you have to add in the depth of the cabinet itself. So, you are, at least, 2-feet away from the wall. You will have a cancellation frequency around 160Hz...the further away from the rear wall you get, the further away you get from a 2π space, more like a 4π and the lower the cancellation frequency you get (where the sound wave reflects off the rear wall and then mixes with the wave front coming off the driver(s). For instance, if the subwoofer is 1-foot off the rear wall, this frequency is closer to 100Hz...which is putting cancellation in the area of interest for subwoofers.

I just did an Atmos room where the use of a baffle wall literally cut the power/quantity of subwoofer requirements in half (which helped pay for the baffle wall). The response and timbre from the screen speakers is also significantly improved in both efficiency and keeping acoustic reflections off the screen from bouncing around behind the screen.

Presuming you cluster the two subwoofers, you are still going to come up lite on the power. As mentioned previously...you'll have just enough subwoofer but not quite enough power behind them with a DCA 1622. Get another DCA 1622 so one amp/cabinet, with the DCA 1622 bridged into 4Ω...you'll have enough to tune to spec and play any movie at "7." You use an SLS with the JBL and all bets are off and be prepared to have an exaggerated mid-bass because it won't be helping the cause below 35Hz-40Hz.

[Jon Dent]

I understand and agree with regards to the elevated sub. It wasn't anybody's first choice but it's better than the enclosure sitting in water (even if briefly).

A baffle wall isn't in the cards at this point, as we're trying to pool money for a future expansion and construction costs are astronomical.

[Jon Dent]

Found the newest version of that DARDT worksheet. Nifty little piece of Excel wizardry there.

We've decided that when the time comes we'll look for another 4642a or get two of the SLS subs. Our current setup works fine at the 5.0 setting we usually run at. I know that's blasphemy to some but , I can understand the dialog just fine and my head isn't ringing when the movie is over.

[Steve Guttag]

The SLS subwoofer is likely going away. Dolby just announced their own Dolby branded subwoofer (dual-18").

[Torsten Juergens]

How important is the sound of a loudspeaker system in the cinema?

The calculations of sound systems in cinemas are mainly characterized by efficiency and power consumption.
The components are then selected on this database.
As also discussed here now.

How big the compression and distortions actually are in reality is not measured.

The mechanical strength and excursion of the membrane under a certain power consumption of the bass chassis for dedicated frequency ranges is not even specified in the data sheets.
Of course, provided there is enough space, you can set up as many subwoofers as you need to avoid this low-bass power problem.

Today's surround loudspeakers in particular should be able to fill the room with extremely high sound pressure across the entire frequency range with constant directivity.

When I then look at the "rudimentary" data sheets from the manufacturers, I sometimes ask myself, how do they do it?

How much linear excursion does the chassis actually have? In what percentage range are the distortions with such a load?

But everything is great on paper.

In my experience, many large halls sound very good up to moderate volumes, but beyond that it suddenly becomes very unpleasant and the entire sound image simply collapses.

Do I just have bad ears or is that just of secondary interest?

[Marco Giustini]

Some manufacturers are better than others when it comes to datasheets. I can think of one speaker I encountered which was rated differently depending if you look at the website, the datasheet or the back of the driver itself. That is why - I suppose - some speakers are more expensive than others!

And that is why 3dB headroom is always recommended to avoid stressing your speakers and getting too much distortion/compression etc.

The sad reality is that sound systems are seldom carefully selected by looking at SPL, power, sensitivity etc. So I guess the goal here is to make sure your speaker doesn't break while in use.

Re. the large hall, I think acoustic is also a big player here. At lower levels the reflected sound won't be so audible while at higher level it will come back to your ears at significant level.

I do remember doing an EQ into a very large hall - with very much little to no acoustical treatment - using an AP20. Once the Dirac sweeps were done I went to test the outcome. The older Dirac tool had an A/B comparison button which enabled/disabled Dirac with a click. There was also - if memory serves - a gain field to try to balance the two as Dirac can only apply negative EQ so the EQ'd sound is usually much softer than the original.
But on that occasion I did not think too much about that so I was comparing non-EQ'd sound at 85dBC with Dirac'd sound at a much lower level. The Dirac'd version was incredibly better - to then realise that the difference was vanishing when the SPL was matched between the two.

[Steve Guttag]

And that is why 3dB headroom is always recommended

I saw that Marco!

It is true that it is a game one must play with manufacturers that also have to prop up their fine-fine products to look better than they other person's spec sheet. And yes, I too have encountered spec sheets that either contradict other publications from the manufacturer or are worded to give the impression of something better than it is. A common trick is to move away from 1W/1m sensitivity and use a voltage...which means there is now an impedance to consider in the evaluation. The sensitivity numbers rarely specify the range that was considered too. As mentioned before a 50-120Hz sensitivity range is a worthless, if not misleading sensitivity for a subwoofer when the area of interest is 20Hz-80Hz (primarily).

I always seek out the frequency response of the speakers (on and off axis, polar, plots...you name it) and go into suspicion mode if they are absent. I take the position that companies hide what isn't favorable to them.

I have, with rare exception, been able to specify equipment that can play whatever the movie has recorded on it, without blowing up.

As to systems that sound harsh as the volume goes up...well yeah. Typically, as the volume goes up, so does the distortion of the driver and so goes distortion (harmonics) goes the pain/discomfort. I have absolutely noticed that if the room is good and the speakers are good, what is considered "acceptable" volume level, as seen on the fader is ALWAYS higher. Your SPL meter and RTA may show that two rooms are identical but the better room with better speakers will be able to play higher and not feel too loud.

Some of this just isn't going to show up on a spec sheet (actually much of it). A lot of this has to come from experience with the product, unfortunately. There is a speaker model that I hate from a brand that I like that I got work with as an outside contractor (I didn't specify or sell it...just tuned it). It sounded horrible (and it didn't tune particularly well either). Because the client just put in what this manufacturer said was right for the room, I got to experience this make/model in several rooms...each time, it sucked! I got to try it with the manufacturer's supplied crossover/tuning, my own supplied analog tuning, my own supplied DSP tuning...hand tuning it afterwards (after the normal B-chain tuning) based on actual content...there is nothing I could do to make this thing sound good...yet the client somehow liked them. What are you going to do? In any event, I got the benefit of that experience to NEVER specify them to a client.

Money isn't always the key either. I've heard some really expensive speakers that I just didn't like at all. I've rarely, if ever, heard cheap speakers that I've liked, particularly at higher volumes. On subwoofers, driver construction/design and cabinet construction/design play a huge factor into it. Circling back on topic...the JBL 4642A is a good systems that does what it says it can do. It is quite predictable. I prefer the 2242 driver for subwoofer work (that is what it was designed for and its T-S parameters show it if one cares to model the 2242...it wants a ported box doing subwoofer work, the 2241 is a large LF speaker that has essentially the same motor as the 2226).

If you want to play with a driver that has a large linear excursion...play with the JBL 2269...it's scary. That is used on the 5628 subwoofer (which looks identical to the 4642A...same box because...JBL likes that volume/driver with a 25Hz port).

[Jon Dent]

I'm one of those weirdos that is trying to learn about how SPL, power, sensitivity headroom and reverberation all play together with our sound system. Thankfully aside from the single subwoofer, the rest of it is correctly sized for our room(s). We don't have bi amplified screen channels though.

[Paul Finn]

The Audio Engineering Society (AES) in recognizing issues with speaker performance specifications, has just adopted and published AES75-2022: loudspeaker maximum linear sound levels using noise - measurement standard. If used by speaker manufacturers, this should help level the performance evaluation playing field.

Paul Finn