Virtual Sound Decoders

[u18b]

The only thing I did experience that was a snafu is this.....

The decoder does not need anything for running (I don't think     ).

However, if is remember correctly from several years ago when I did this, it needs some kind of resistance for programming.
When I encountered this, the simple solution was to just mount a spare motor across the gray and orange leads.
So yes, I have a N scale motor running inside my sound electronic system!

But people who are more knowledgeable than I might know what kind of resistor you could mount across the orange and gray leads instead of a motor that
1.  will not damage the decoder
2.  not burn out the resistor
3.  have the resistor get so hot it damages your enclosure.


 

[chicken45]

Speaking of sound, you guys see the Rolling Thunder thing from BLI?

http://www.broadway-limited.com/rollingthunder.aspx

I'd buy it in a second, but I could see people not in to sound.
If you think sound on a layout is annoying, just imagine the prototype!

[jagged ben]

As someone who used to do sound design for theatrical productions, I've thought a bit about what it would take to do this well.  Unless one has a very small switching layout, it would be a lot of software development, in my opinion.  Personally I think I would rather just have speakers in the locomotives if it were not possible to accomplish all this.   Things I care most about are literally the bells and whistles.  (Well, horns not whistles for my diesel era modeling).    I tend to turn down the other sounds because it is just not that realistic for them to be so loud if you are not up close.  (Think about it, if you are 8 feet away from your N-scale locomotive that's a quarter mile in scale and in the real world you will not hear much rumble at that distance if it's not a 6 unit consist going up a steep grade.)   Those 'ambient' sounds are not so simple to thrown in realistically. 

Anyway, here are the components I believe one would have to make work together to do this well.

- Transponding in locomotives
- Block detection (at least your signal system also needs this), resistor wheelsets etc.
- soundfiles for each engine type that correspond to the all the decoder features
- at least 20 or 30 random sound effect files for the non-loco rolling stock effects: wheel scraping, flat spots, etc, different speeds, etc.  These should be distinctly edited effects, not just playback recordings of a train passing by.

The software would have to:
- Keep track of where each engine is and what its type is
- Follow block occupancy and speed of trains
- randomize the various sound effects to not be too repetitive
- fade effects from each speaker location to the next as trains pass by, according to their speed  speaker locations would probably need to be every 4ft or so on an N scale layout to feel realistic.
- fade effects in and out according to block occupancy, not merely soundfile length

One thing that seems to have changed since I was doing much audio is the affordability of multi-channel audio interfaces for computers.  It looks like a 16 channel output device would cost between $500-1000.  A computer that can handle such a card, probably the same.   So at least this idea has that going for it these days.  The development of the software is a whole other ballgame, I think. 

 

[GaryHinshaw]

I have been thinking along the same lines jb.  I agree with your outline and I've added some comments below.  I see a system like this as probably a supplement to on-board loco sound rather than a replacement, though it's hard to tell without trying it.  In any case, it would let you fill out the frequency spectrum, add ambient sounds, and -- when desirable -- boost volume. 

Sound is not currently a high priority on my layout, but it's fun to think about, and it might be, eventually.

Anyway, here are the components I believe one would have to make work together to do this well.

- Transponding in locomotives
- Block detection (at least your signal system also needs this), resistor wheelsets etc.

With contiguous block detection, you can define memory variables in jmri/PanelPro that will track a train's location by block under the assumption that the train runs between contiguous blocks. This works well in simple operations, but it is possible for the code to lose a train in more complicated situations.  The capability is free if you have block detection already set up and it potentially frees you from the need for on-board transponding decoders.  The information could be exported to a sound controller.

- soundfiles for each engine type that correspond to the all the decoder features
- at least 20 or 30 random sound effect files for the non-loco rolling stock effects: wheel scraping, flat spots, etc, different speeds, etc.  These should be distinctly edited effects, not just playback recordings of a train passing by.

Seems like you would want a set of actual sound decoders (including one or more with ambient sounds) to drive the output sound.  The job of jmri would be to send throttle info, train ID, and block location to the sound controller which then drives the appropriate decoder(s) and directs the output to the appropriate speaker(s).

The software would have to:
- Keep track of where each engine is and what its type is
- Follow block occupancy and speed of trains

Done - sort of.  See above comments.

- randomize the various sound effects to not be too repetitive

Especially for ambient sounds.  For loco sounds this could be an existing sound decoder.

- fade effects from each speaker location to the next as trains pass by, according to their speed  speaker locations would probably need to be every 4ft or so on an N scale layout to feel realistic.
- fade effects in and out according to block occupancy, not merely soundfile length

Definitely.  This seems pretty doable with the right mixer setup though.  My son is a sound engineer, so I plan to tap his expertise. 

[sirenwerks]

- fade effects from each speaker location to the next as trains pass by, according to their speed  speaker locations would probably need to be every 4ft or so on an N scale layout to feel realistic.
- fade effects in and out according to block occupancy, not merely soundfile length
 

I'm going to keep banging the drum for wearable speakers.  For certain sounds - the loco consist - it makes; in the current DCC sound situation the speaker travels with/in the loco and, usually the operator would be following the loco, so why not wear a speaker?  Something like a belt clip unit.  The speaker would be larger than what N scale locos could carry, allowing for better sound quality, and the travel issue is solved mechanically rather than technologically. 


Stationary sounds could be left to stationary speakers for ambient sounds and train-related locales for sounds, like curves where squeal could be a feature.  And if you use hypersonic speakers for these aimed at the location (say, ceiling mounted aimed at the layout so that the operator would travel between the speaker and layout, rather than stationed at the location pointing outward towards everyone), and set to reasonable levels then the sound-blasting-across-the-room issue could be resolved.

[peteski]

I still don't get this desire to make a 6" model sound as powerful as a real Diesel.  I can't even imagine trying to operate a room-size layout with 5 or 6 of these running at the same time. 

Just last night I test-ran a pair of Centipedes on friends room-size layout and even just those two were way too loud. The Diesel rumble (as much as can comes out of 4 small speakers) got really annoying after just few minutes.  Granted, that was at the factory volume setting (fairly high). I will eventually turn the volume down, but even with lower volume, with multiple sound locomotives running the noise gets annoying quite quickly.  A deep low frequencies of the type of sound system being discussed here will just add to the noise pollution.

[GaryHinshaw]

I think a number of us have tried to make it very clear that volume is not a primary goal.  Rather, it is full frequency response and (for me at least) ambient train sounds, which are a hallmark of mountain railroading.  It may be an unrealistic goal, and it may even turn out to be undesirable, but the technology is within reach and worth a try.

[peteski]

I think a number of us have tried to make it very clear that volume is not a primary goal.  Rather, it is full frequency response and (for me at least) ambient train sounds, which are a hallmark of mountain railroading.  It may be an unrealistic goal, and it may even turn out to be undesirable, but the technology is within reach and worth a try.

I understand the full frequency response thing (at reasonably low volume).

But low frequencies are non-directional, so the rumble will seem to be coming from every direction in the room. Plus low frequencies in a hard-wall room create standing waves, so as you stand in certain locations in the room the low frequency part of the sound can be either as intended, much louder, or even much quieter (if you happen to stand in a location where the waves cancel out).  I can easily hear that effect when listening to music on my stereo while moving around my room (which is not covered with acoustic deadening material). 

[Ed Kapuscinski]

I think a number of us have tried to make it very clear that volume is not a primary goal.  Rather, it is full frequency response and (for me at least) ambient train sounds, which are a hallmark of mountain railroading.  It may be an unrealistic goal, and it may even turn out to be undesirable, but the technology is within reach and worth a try.

Also, for me, adding operational sound to things like SW1500s without a milling machine, or whatever made Ant Man.