He doesn't like any of the sound decoders he has ever listened to for diesel because they have no way of telling if an engine is loaded climbing a grade or coasting down hill.  Maybe someday the manufacturers will come up with the technology to simulate sound correctly for diesel engines in all modes of operation, but right now he is passing on sound entirely because none of it is accurate.

You might try adding some momentum by tweaking the relevant CV for the QSI decoder. While I am not familiar with these QSI Revolution or Quantum decoders, for Soundtraxx Tsunami's adding momentum for acceleration and braking affects the sound.

John 

From my experience as a user of large sound systems, and from some of the discussions I've read concerning model sound installations, my first guess is that the speaker selection, and the enclosure within the loco, are the biggest determinants of good sound reproduction, especially with the lower frequency sounds.  I suspect the techs at Tony's or Litchfield Station (MRH sponsors) would have some words of wisdom here.

Click on the link to Rapido Trains web site and go to the page on the up coming Fp9 model.  Click on the latest issue of "Rapido Telegraph."  Watch the "cheesy movie" near the bottom of the page.  The movie is a short, light hearted, documentary on them doing the sound recording from an Fp9 at the Conway Scenic Railroad in New Hampshire.  They are recording the sounds of an Fp9 in passenger service, so they put another locomotive behind it with a load of about 100 tons (don't know if that was the weigfht of the locomotive or added weight of extra rail cars behind the extra locomotive.)  Anyway, they point out in the movie that a real Fp9 in commuter or passenger service will go from stopped to run 8 on leaving the station and then go from run 8 to 0 when arriving at the next station.  Most sound systems are recorded on locomotives without a load from 0 to about run 3 slowly accelerating from 0 to run 3.  The problem is that a real locomotive is very dynamic, and sounds different depending on how much load is on it, whether it is "cruising" on the level, climbing a hill, or coming downgrade with dynamic brakes working.  Model sound decoders just aren't recorded under all of those conditions, and without a cooperating railroad allowing that sort of recording over the course of a trip,  they aren't likely to get to do an accurate recording.

I have been completely impressed by everything Jason has done with Rapido.  Yes, they have had some quality control issues, but for a brand new company, that is not uncommon.  When the problems have arisen, they have gone the extra mile to make it right, in my opinion. 

I watched that video last night and I can't wait to get my new FP9 from them.  I only plan a diorama in HO, but I bought the HO Rapido passenger cars and have pre-ordered this FP9 (from MRH sponsor, Dallas Model Works) because it really is second to none in terms of the detail and effort put into making it "right."  They may just drag the rest of the industry up in terms of their innovations and attention to detail. 

I also have 6 of their N scale passenger cars also and really, REALLY hope they make the FP9 in N.  I think I would order 6 at least if they did.  I think Jason is probably sick of hearing from me ("when are you going to make this in N scale??") but I'm a big fan of someone who puts such an effort into their products.  And they are really a small company compared to Walthers, Athearn, Intermountain, Atlas, etc.  If anyone has not checked out Rapido yet, you must do so, especially if you are a Canadian modeller.

I just purchased a Proto 2000 RS27 with QSI sound built in.  I found the sound unconvincing when I first tested it as well.  But I made a few small CV adjustments and it sounded better afterwards.  I set the sound volume down to 75 and set the acceleration and braking momentum up to 40 and I turned the starting voltage down so that the loco wouldn't move for the first couple of speed steps.  I also turned off the stupid "talk back" feature.  The momentum and starting voltage allowed the engine to rev up realistically before the loco moved and then die back down once it was moving.  Turning down the throttle to zero when the loco was moving at about 30 mph allowed it to coast for about 18 to 24 inches before stopping, and it also made the sound drop to an idle long before the engine came to a halt.  This improved the slow speed sound considerably.

The decoder also has something they call the "Sound of Power" feature which makes the volume increase when the load increases and drops when the load drops off.  This makes the revs louder when the loco accelerates and quieter as it coasts and goes downhill.  Pulling a heavy train keeps the sound loud all the time.

The loco came from the factory with two large speakers mounted behind the cab facing towards the driveshaft.  I don't see how anything larger could be fit into the loco.  I found the sound somewhat mushy at road speeds, but I think that was partly because I turned down the volume and because it's an Alco sound, and they have a softer sound than EMD's.  I've also been operating it with the dynamic hatch opening uncovered because it's an undec and I haven't opened the parts packets (I don't want to open them before I'm ready to install all the details because I'm afraid of losing door handles....), so the sound is escaping out the top of the carbody and defeating the "sealed speaker" idea.

Before I bought this unit, I wasn't a fan of QSI, but I'm more impressed now.

I to saw the Rapido video. I just with it was an MILW FP7...

I purposely faced my speaker (w/enclosure) upwards so the sound would shoot out the rear see-through fans.  Perhaps this was a mistake based on what you've written.  It can't hurt to give it a try facing toward the tracks - I'll do that and report back.

Also, I had read only enough about the Sound of Power feature to know that it affected the volume of transitions; I didn't realize it went beyond that.  I'll give that a try as well (and it goes to show you should always read the manual thoroughly).

Good stuff; thanks.

The idea behind a sealed speaker enclosure is to make the speaker push the air in one direction at a time, so if you have the fans open above the speaker, it won't make any difference if you turn it over; it will still have sound coming out the top and bottom at the same time.  To make it worthwhile, you should close off the bottom of the fans so the sound only goes downward.  It really works best if you make a box around the back of the speaker which will let no air sneak around behind it.

Jurgen

The idea behind a sealed speaker enclosure is to make the speaker push the air in one direction at a time, so if you have the fans open above the speaker, it won't make any difference if you turn it over; it will still have sound coming out the top and bottom at the same time.  To make it worthwhile, you should close off the bottom of the fans so the sound only goes downward.  It really works best if you make a box around the back of the speaker which will let no air sneak around behind it.

It actually goes beyond air movement.  If low frequency sound waves from the rear of the cone are allowed to escape without being properly baffled they tend to cancel the low frequencies coming from the front of the cone.  There goes that low growling sound we all want to hear.

Jurgen has it right on the money when he says it is best to build a box around the back of the speaker.