My user post to Freiwald Softwares international “Traincontroller Forum” June, 15, 2011.

Railroad and Company 4D Sound: A New User’s Personal Experience, Observations, Questions, and Comments


I have been getting information and gleaning insight regarding computer control of model railroads from this forum for years. I feel it is now time for me to give a little something back in return. I hope this is useful to some of you.

Last month, I purchased Railroad and Company’s 4D Sound software and a modest 7.1 surround audio setup for my old computer.

Initial Assumptions:

After some research, planning, and thought, I concluded that a $30 7.1 Surround USB-linked sound box driving 8 cheap low power desktop speakers was all that was needed for my N Scale layout. I reasoned that since the N Scale trains are very small and since viewers observe the miniature landscape from a distance in an almost God-like overview, a high volume, high fidelity audio system would sound unrealistic and way too “in your face.” So far, my experience with the 4D system bears out this assumption, and I find that both moving train sounds and stationary sound effects are convincingly realistic at low volume levels. In addition, the low volume levels help the viewer/listener distinguish between sounds emanating from various places in the room. I am convinced that, with a louder output, each sound would tend to fill the room, and thus, multiple sounds from different parts of the layout would quickly create confusion as to its source location within the overall sound field.

Finally, I selected 7.1 surround as opposed to 5.1 because I wanted dedicated center rear speakers so that the multiple speaker setup along the long rear wall of the room would sound as much as possible like the 3 speaker (front left, front center, front right) setup along the opposing, long, front wall.

The Problem with Surround Sound:
In my mind, the ideal surround setup for 4D Sound is one in which all speakers are “discrete.” That is to say, Left Front, Center, Right Front, Right Surround, Right Rear, Left Rear, and Left Surround all operate independently from one another. With such a system, if the 4D software assigns a sound to the left front corner of the room, then that sound would come only from the point source of the left front speaker. Similarly, a sound assigned to the front center of the room would come only from the front center speaker, and a sound from a train traveling between the left front corner of the room and the front center would be tracked by panning between these two speakers – and so on all the way around the room.

Unfortunately, this is not exactly the way Surround Sound works. Most surround algorithms are designed to fill the center of the room with sound in order to enhance the listening experience of a listener seated in the exact center of the room. To do this, they employ matrixing schemes which, in certain cases, involve multiple speakers. For example, in both 5.1 and 7.1 surround sound, sounds assigned to a front center location come not only from the center speaker but also from an equal mix of left front and right front. Similarly, in 7.1 surround sound, sounds assigned to the center rear of the room come not only from both rear speakers but also from the left and right surround speakers. Furthermore, it appears to me that with 7.1 surround, any sound assigned to the rear speakers comes from both rear speaks equally as well as from the left and right surround speakers. Rear wall positioning panning is accomplished in the same way as on the front wall of the room, but with four speaks in play, the sound image positioning is even less precise unless one places the two rear speakers together in the center of the rear wall.

Another problem is that surround sound is designed to “hype” the listening experience at the exact center of the room. One result of all this “hype” is that sounds assigned to the listening position in the exact center of the room are much louder than sounds assigned to locations around the perimeter of the room. This is fine if your layout hugs the perimeter of the room like mine, but if your layout passes through the exact center of the room, you can only use the block attenuation control and/or “cheat” central block locations away from the center of the room to adjust the volume of a moving train. The volume of stationary sounds assigned to the center of the room can, of course, be adjusted within the Train Animator.

The result of all this matrixing and “hype” is too much sound. The ideal system for 4D model railroad sound would produce sound that clearly emanates from precise point-sources around the room. While surround sound is designed to reflect myriad sound source locations in the room, it is also designed to flood the center of the room with sound. With today’s modern surround sound algorithms, a listener seated exactly in the one prescribed ideal central listening position will begin to lose the intended spacial sound source perception within the sound field as he or she moves away from the center of the room. The 4D software only outputs x,y,z spacial coordinates. It cannot control any surround algorithms. So, although 4D works quite well in most cases, it does so using a less-than-ideal audio interface, and it is fully effective only if the listener remains in the center of the train room.

I should note that my $30 7.1 surround USB box is perfectly capable of putting out the desired 8 discrete speaker output without matrixing or “hype.” This is clearly demonstrated by the nifty speaker test mode that comes within the unit’s setup software. However, this box, like all the others, is designed to reproduce 5.1 or 7.1 surround sound, that is to say, partially matrixed room–filling sound, and unfortunately that is all it puts out in any mode other than the test mode.

Finally, I might also note here that there are two new, sophisticated 7.1 surround algorithms that (I think) do not employ rear channel matrixing. These are Doldy TrueHD and DTS HD Master Audio. Although these new surround formats undoubtedly put out room-filling “hyped” sound, I believe they are capable of truly discrete 7 speaker output. Unfortunately, they are only available on high end audiophile surround sound receivers. To my knowledge, they are not available on any sound card or sound box with low power speaker outputs. Has anyone tried these new surround algorithms with 4D?

Speaker Positioning and Set Up:

With all of this in mind, I placed my little speakers on the floor under the layout behind the thin fabric skirt that conceals the area under bench-work. (I plan to experiment with raising them up and attaching them directly to the underside of the bench-work soon.) Left front, right front, left surround and right surround were set in the corners of the room; the center speaker was placed in the center of the long front wall, and both rear speakers were placed together in the center of the long rear wall. I did not use a sub woofer. All volume controls on the speakers themselves were set at about 80% to leave room for tweaking. Happily, I found that most .yra sound files I created for this setup worked best when I set the volume in the Train Animator between 60% and 90%, thus leaving room for further volume tweaking of individual sounds later on.

As I mentioned my layout runs around the perimeter of the room, so with this setup, I get ideal, fully discreet sound for all sounds assigned to locations near the corners of the room, and I get un-matrixed panning from front to rear. The multi-speaker 7.1 surround matrixing only comes into play with sounds assigned to locations near the center of the rear wall. I was able to lessen the undesirable effects of this matrixing by turning the volume controls on the rear center speakers up just slightly – just enough to help obscure the fact that sound is also coming from the rear corners of the room, but not so much as to call much attention to the fact that things in the middle of the front and rear wall are louder.

The results of all this are, I must say, quite satisfactory. Especially when you take into consideration the fact that, when one is viewing a moving train, one tends to associate the train sound with the train location even when the matrixing or your off-center listening position might be throwing things a little off.

Train Sounds:

I am not very far along with this. Most of my locos are diesels, and so far, I have only managed to build simple yet acceptable .yra profiles for two trains using only train-rattle and diesel-hum loops. I am yet to mess with sound for a steam loco - I have enough trouble, keeping the damn little things running and on the track. In summary, building these .yra files is very tedious, but, I think, do-able, and I do not doubt that, with time, I will develop a pretty good library of sounds all of my trains.

A Few Problems and Questions:

One problem I am experiencing is that when I use “train operations” associated with either a schedule start or a section “continue” to turn on the sound for a train, I get full volume train sound for a few seconds as the train starts up. Then, as the train begins to move along, all sounds settle down to the correct volume levels that I set in the associated .yra file. I tried programming in a delay in operations and I also tried delaying the train start using waiting time in the schedule specific block menu, but neither of these tactics seems to help. What am I missing here?


In summary, I am generally delighted with the 4D Sound software, and I am sure that, with practice, I will develop even better sounds and more realistic sound illusions. This is all very complex, very tedious, very challenging, a lot of fun; and it really enhances the overall experience of The Altamont and Blue Ridge Railway.

Danke sehr, Herr Freiwald.

All comments are invited.

Pete

April, 2012 - Note: Not too long after I posted this article, Freiwald Software introduced an update to the 4DSound software that did away with the center of the room volume problem mentioned above, and I found that the start up volume problem I was having had been addressed in the release of a recent, more sophisticated version of the program. This is a great company.