JMRI Virtual Sound Decoder

[GForce2010]

I have been researching ways to add sound to my layout without having to spend £100+ per loco for decent sound decoders.

I googled and discover the Virtual Sound Decoder section on JMRI where you can assign a sound profile to a loco and it will play all the sounds that a actual sound decoder could play.

Has anybody used that before and can anybody advise where I can get the required sound files for different loco's?

Thanks.

[009matt]

That sounds quite a neat trick. I take it that you put a small speaker in the layout somewhere to amplify  make the sounds produced come to life. It may be a good approach for me in 009. I will have a play myself when I get time. Will be interested to see how this thread progresses.

 

Matt

[wiggy1]

As I'm no expert with JMRI I have seen this were the loco plays a sound at a set point around the layout according to what the locos is doing.

But the sounds were coming from the locos sound decoder not from speakers around the lay out.

However you could get whistles to work from a speaker around the layout triggered by the loco.

Lets see what other replies we get.

Wiggy.

[Harlequin]

Sounds like a very good idea! The tiny speakers in locos are never going to reproduce sounds very well, no matter how clever the installations become.

 

So how about this:

Never mind placing a small speaker somewhere near the layout, place several large-ish speakers around that can reproduce both bass and treble properly.

 

Then get some software to steer the sounds around the speakers. It would need a map of the speaker positions, some knowledge of the physical arrangement of the layout and would use train detection and knowledge of train speed to gradually adjust the mixing between speaker channels.

 

Hmmm, could be expensive!

[Nile]

https://github.com/JMRI/vsdecoder

[Nile_Griffith]

Sounds like a very good idea! The tiny speakers in locos are never going to reproduce sounds very well, no matter how clever the installations become.

 

So how about this:

Never mind placing a small speaker somewhere near the layout, place several large-ish speakers around that can reproduce both bass and treble properly.

 

Then get some software to steer the sounds around the speakers. It would need a map of the speaker positions, some knowledge of the physical arrangement of the layout and would use train detection and knowledge of train speed to gradually adjust the mixing between speaker channels.

 

Hmmm, could be expensive!

 

 

 

Even overcoming the expense it would still be a highly elaborate system. You could cut the number of speakers down to a manageable level, but you are still left with the problem that a DCC based system would have to communicate a loco's position to a computer (or software running on the same DCC operating computer), where the software would then have to select the appropriate sound file for that loco, take into account any loco acceleration/deceleration changes. Then from the positional information map that sound to a discrete output on a audio hardware device which either has onboard amplification sufficient to drive a loud speaker or in turn feeds a multichannel loudspeaker amplifier. The real heavy work would be the software required to do the mapping and the amount of "fiddling" required to come up with some sort of convincing soundscape, when the whole system is meshed together.

 

Slightly simpler would be some kind of control system that overcomes the limitations of on board stored sound that current DCC sound chips have i.e. relatively small memories. Unfortunately the way to do this would be to transmit the audio data to the loco itself from a library of sound files stored on a host computer and DCC isn't bidirectional.

 

Bachman and BlueRail have their current Bluetooth control system, which at present doesn't offer the facility to transmit sound to a loco. But the technology itself does have the capacity to do so in the future.

 

Personally I feel that DCC sound is a bit of a "nicety". You can't scale sound and there all sorts of physics that go with creating convincing artificial soundscapes that current DCC technologies just can't overcome. 

[sjrixon]

I had a play with this a few months ago, but didn't persevere as it was more of a distraction at the time to actually building the layout.

 

 

This was done with standard stereo speakers, what you can't tell from the video is the big issue with that. The sound doesn't move, even cheapest of TTS sound chips move with the loco. Running a loco out of my fiddle yard, into the station is all at the same volume, so it just didn't quite work for me. To add sections to a layout that doesn't need them for any other reason felt like a huge expense, for no good reason. You also get some very generic sounds, not the high fidelity that come from the DCC sound manufactures. 

 

But.... It's certainly worth kicking the tyres and see what you think!

 

What I will add.. With sound on, you drive the loco a little better. Sound the whistle, pull away to match the sound, it does add to the experience. I looked at a guards whistle and a few other sounds too, just all adds to it. 

Edited January 7, 2018 by sjrixon

[009matt]

Thanks - that is really interesting. For my 009 cameo layout, that is 4'6" viewing, it may be useful, although even then, the sound might feel a bit static, but I will have a try. 

[Trip]

 you are still left with the problem that a DCC based system would have to communicate a loco's position to a computer (or software running on the same DCC operating computer), where the software would then have to select the appropriate sound file for that loco, take into account any loco acceleration/deceleration changes. Then from the positional information map that sound to a discrete output on a audio hardware device which either has onboard amplification sufficient to drive a loud speaker or in turn feeds a multichannel loudspeaker amplifier. The real heavy work would be the software required to do the mapping and the amount of "fiddling" required to come up with some sort of convincing soundscape, when the whole system is meshed together.

 

 

This is more or less what JMRI's virtual sound decoder tries to do. It uses a 5.1 or 7.1 surround system connected to the PC running JMRI. You can associate a collection of sounds and a config (together called a VSD file) with a JMRI roster entry. JMRI tracks the loco round the layout using whatever block detection system you like and knows what the throttle positions and speeds of the loco are (the same way a sound decoder does) and plays sounds through the speakers.

 

I've only just started playing with it, but it looks promising.

[SimonME46]

I’d like to work on this soonish. Getting the sound files seems to be a problem, however MSTS has kinda been replicated by OpenRails, and so we have open source sound files and quite good too. The issue will be workin* out the various triggers for the loco sounds, especially steam locomotives.

 

I note your comment about the sound not varying as the loco moves, that will be something to watch out for..

[sjrixon]

Interested to see if you persevere with this.. Are you planning to have sections in JMRI so the sounds will move between them?

[Pandora]

The JMRI vitual sound decoder is replacing expensive (£45 to £100 per locomotive )hardware (the sound part of the decoder) with potentially low-cost  software,a  most welcome development  and a step in the evolution of the DCC system. 

Edited March 30, 2018 by Pandora

[GEMLKEL]

Re-opening a old tread but though it best to keep everything JMRI VSD related together.  I am building a new N Gauge layout based in East Anglia set around 2000 to 2010 and heavily influenced by Ipswich.  

 

I will be using the JMRI virtual sound decoder system to provide sound for the 15 or so locomotives as it is impractical to fit sound to each of them.    

 

Very early days but I have started to learn how to create new sound files, the first being a Class 66 using samples taken from various YouTube videos - results so far here - 

 

Edited March 6, 2021 by GEMLKEL

[Alan Oliver]

I have been looking into the vsd set up but cannot get it to work, have read the instructions but cannot find the menus in jmri that it says you need.

 

thank you 

[FranksLad]

I use VSDs on JMRI. It’s a good compromise. Sourcing sounds isn’t a problem if you also happen to have Train Simulator on your PC and you only wish to use the sounds for your own use 😉

[Alan Oliver]

How do you use the sounds from train simulator? It’s all new to me using VSD.