If you want to know a good way to contain/direct/deliver the ejection gas while maintaining its effectiveness, read on. Otherwise, ignore.
Pressurized gas will flow best over short distances if you give the gas path a slightly expanding duct. By slightly I mean that the cross-section should probably increase by about 5 to 10% per foot. If you start with a BT-50 (with an inside diameter of 0.95 inches) you have a flow cross-section of 0.71 square inches. So a foot downstream, your flow area should be 0.75 to 0.78 square inches, or 0.97 to 1.00 inch diameter. But building a tapered duct is a pain in the ***. What difference does the taper really make? Probably very little.
In a theoretical world, the flowing ejection gas will form a boundary layer against the inside of the duct walls just like the external airflow does on the airframe external surfaces. Over a long enough distance, the boundary layer can build up and begin to block the flow. I dont have numbers for you, but I dont think that will pose a problem for anyones model rocket (or mid-pwr, or high-pwr either). Simple is also good, and a constant-diameter duct will carry the ejection gas just fine in the real world. Ejection charges seem to have plenty of power to spare to reach through a passage and still kick out a recovery system.
Complications begin when you allow the ejection gas to expand into an internal space and then try to re-direct the flow into a smaller duct. You would have this situation if you build your multi-motor mount to exhaust the ejection charges into the main BT and then use a centering ring to close down or converge the flow to fit into a stuffer tube. The gas will cool (talking theoretically again here, and cool only with respect to gas conditions inside the front of the motor case; the expanded gas will still be plenty hot) as it expands into the BT volume and the pressure will drop slightly. Facing the obstacle of the CR/stuffer tube, however, the gas will lose energy as it gets pushed into the smaller passage.
Over in the real world, we are beginning to run the risk of ejection problems IF we combine such a mount/stuffer design with a large/long rocket, a long/small-diam stuffer tube, and a large recovery compartment. You are not always assured of having all ejection charges go off at once (actually, you are pretty much guaranteed that the ejection charges will be staggered). For practical purposes, we really are talking about using the FIRST charge to deploy recovery; subsequent charges may serve as back-up. But for any and all motors in the cluster that are intended to provide ejection, you are relying on each individual charge to be capable of pressurizing the entire convoluted gas path and popping out the recovery system.
For most model-rocket-sized designs, deployment is probably not a problem; each ejection charge will probably be big enough. The more serious problem is over-pressurization of the compartment ahead of the motor mounts and facing the convergent CR (how many Ds do you have stacked together?). The ejection hot gas will go into the main BT compartment and stall momentarily while it tries to flow out the forward stuffer tube, effectively pressurizing this structural compartment. (One aspect of these pressure loads that I have never heard addressed is that they are applied very quickly, almost explosively; the severe onset rates of these loads can also cause problems if construction is not up to the job.) Structurally speaking, it does not matter whether this pressure lasts for three days or three micro-seconds, it will still act against the sides, front, and rear of this internal space. If you dont build this compartment to withstand the pressures, temperatures, and abrasive flow, you may have a surprise deployment of your motor mount. Or a surprise collapse of your internal structure and stuffer tube.
That is why micromisters comments are extremely important to study and understand. It may be necessary to use a bit more beef in constructing the compartment. Note his experimentation with foam-core bulkheads and centering rings; that stuff is pretty darn strong and still very light.
Another option would be to build extensions of the individual motor mount tubes projecting forward through what would otherwise be the stuffer tube zone. You would still need a good bulkhead at the front of the extensions, to seal off the internal volume of the center airframe and minimize the compartment volume that must be pressurized for recovery deployment. This will deliver ejection gas most efficiently to the front of the rocket, but will add complexity to the structural guts of your design. This step is probably NOT necessary if you can build a good, solid compartment around the front of the motor mount.
Yet another option would be to make a short internal cone-shaped duct to funnel ejection gasses down to the stuffer tube. The rear of the duct would have to cover the front of any/all motor tubes that are to provide ejection. All connections would require solid attachment and good reinforcing fillets. The internal face of this duct should be coated with epoxy or CA to improve its fire resistance. The outside of this duct might even be wrapped with reinforcing materials.
Keep in mind, most of the above discussion is really only important for situations where you think the ejection/deployment might be marginal. For model rockets with a gas path of up to three or four feet, and probably for airframes up to BT-60 size, you probably dont have any problem at all.
And yes, many old men like me know quite a bit about passing gas.