DCC sound decoders for US diesels

[Gary H]

And yet...

I have noticed this, and am aware of the former comment, too.

 

Why is this, as it seems contradictory? I am just a layman, so please avoid being over-technical in the response!

It would suggest that this would be an issue with the decoder rather than the actual physics and constraints of a speaker assuming that the speaker is properly baffled to start with!

Air horns by their very nature are more in the way of midrange and treble frequency's and so are more easily re-produced by our tiny speakers compared with the deep rumble of a diesel engine! (AKA bass!)

Ive noticed that on especially Soundtraxx Tsunami decoders in particular that the horn sounds are always below par compared with say a QSI with the same horn. Its especially noticeable with a K5LA on an SD70ace. Personally, I live this shortcoming in the Soundtraxx as I think the diesel engine sound is more accurate than a QSI.

On a slightly different note, I listened to the horn sound on a ESU V4 fitted SD40-2 earlier and again, the horn sound atleast  totally eclipsed the Tsunami version, but the owner of the model noticed that the diesel engine sound was louder on the Tsunami even though both models in the test used the same speaker set-up!

All very surprising I thought. Horses for courses and all that.

[torikoos]

I've just posted an instructional video on installing a sound decoder, mainly for those who've never done it before.

 

I haven't recorded any sounds to listen to, this is mainly 'one of many' how to's around, and hope it's useful to someone.

There's a learned lesson in there too, I'm not afraid to make mistakes and admit them so others can avoid it. :-)

 

And yes, that wiring will be cleaned up when I replace the bulbs for led's ;-)

 

Koos

 

[Gary H]

I've just posted an instructional video on installing a sound decoder, mainly for those who've never done it before.

 

I haven't recorded any sounds to listen to, this is mainly 'one of many' how to's around, and hope it's useful to someone.

There's a learned lesson in there too, I'm not afraid to make mistakes and admit them so others can avoid it. :-)

 

And yes, that wiring will be cleaned up when I replace the bulbs for led's ;-)

 

Koos

 

That's a shame about the bulbs, although the Athearn bulbs are always lambasted by everyone, its a shame you blew them! Wish you had of asked as I'd could have given you a heads up here!!

Mind you, Athearn bulbs being what they are, they seldom last more than a few weeks anycase.

If you want to replace them, ive a fair few that Ive removed for LED's myself. All Il ask if you want them is the cost of postage for half a dozen of them.

For an ESU newbi like myself, that was an informative video!

Its interesting about the 'bridgeable' pads for 12V bulbs or LED's as NCE boards use the same thing on some of their DCC decoders! Its certainly a useful feature!

You may need to check the output voltage on those light pads as with the resistors in use for 1.5 volt bulbs, I don't think that will be enough for a LED??!!

Ive ended up bridging the pads (like you already done) and used my own 750 OHM resistor for each LED.

[torikoos]

Cheers Gary, it was just me wrongly asuming they were, a lesson in 'RTDM' first.... It was stated in small print on the back of the package. 

The onboard resistors are scaled that LED's should work directly according to ESU (They are 2200 ohms)  (also see the SELECT manual downloadable from the ESU website on page 12 for the schematic, and page 14 for a description) The actually say to by pass the terminals as shown, but use a separate resistor for the Athearn bulbs.

Where there are two in parallel (headlights) use a resistor between 360 and 510 ohms, and one bulb (roof beacon for example) between 680 and 1000 ohms.

The factory resistors will limit the current to 8 till 10mA, enough for most LEDs, they might not be as bright in some cases but probably adequate.

Thanks for the offer of the bulbs, but I'll install LED's. I much prefer those over the Athearn bulbs, they are already ordered .

 

Cheers, Koos

[Gary H]

No problem!

Would like to hear the sound on the GP15 sometime too.

[torikoos]

If I get a chance, I'll record the sound of this loco at the coming weekend and post it up. I guess however that my LED's might have arrived and I'll be busy replacing them first, before I get to assemble the loco again, but when it's all done, I will definitely record it on video again.

 

Koos

[torikoos]

HI all, here a short video of the loco in action, however the background chatter might drown the engine noise a bit.

 

[torikoos]

And here's one taken on the layout.

 

[Alcanman]

Nice videos Koos. The second video demonstrates the sound really well.

 

Mal

[torikoos]

Thanks Mal, the warning beep at start up of the engine is a little loud though, sounds like that carry a long way (similar as a bell sound, which I normally set up with a fairly low volume as they can quickly become very irritating otherwise). I think the sound recording might have been made under the hood (inside the engine compartment). 

That this sound carries a long way was demonstrated as my wife thought it was a smoke alarm going of at the neighbours, until she realised it originated in our loft :-)

 

The other engine sound on board is of an 8cylinder engine (good for smaller switchers like the SW1500 etc) (which you can select by changing a CV) , which does not have the beep, but has a hydraulic starter 'buzzing' for a while before the engine kicks in. Personally, I like that effect much better but the engine does 'chug' a bit more as every revolution there's 4 cylinders less in comparison to the 12 that a GP15-1 would have had. 

 

 

Koos

[torikoos]

I've now got two of these lovely engines. And both have now been equipped with a Loksound decoder. Here are the sisters in action.

Koos

 

[torikoos]

Did I mention the Loksounds have a useful auto tune function, which get's performance fairly close to where you want it. (these two locos have just that applied, I haven't yet fine tuned them any further).

Very easy, use Program on the main. (make sure you have a few yards of empty track available for the loco to be tuned), set CV54 to 0. Then press F1.

The loco will rapidly move away, slowing down to a stop, all the while the decoder reads the feedback of the motor and sets the trim values close to perfect. make sure you do it somewhere where it can't run off the layout meeting your carpet etc.. (or use a rolling road). 

That's it, loco tuned and slowspeed behaviour will be fairly good straight away.

 

Koos

[Gary H]

Im guessing you've seen this one Koos?

104 looks to be an ex CNW loco with the gong bell on the nose.

Edited December 9, 2013 by Gary H

[torikoos]

Hello Gary, yes I have. All CFNR GP15-1 were ex CNW engines. All retained their nose gong bell.  (ESU is working on releasing a soundfile with this gong bell included, and I'll be uploading those to my decoders once available).

They (CFNR) operated them in California until the Gensets were purchased. They then disappeared (apart from one or two stationed near Davis CA) of their home rails and can be found as far away as Texas. Now that parent company Rail America has been purchased by the Genesee&Wyoming some GP15-1 have made it back on those rails, be it in orange GWRR livery and I also am not convinced they're ex CFNR units.

I still like the CFNR colours the most on them.

Edited December 9, 2013 by torikoos

[Gary H]

Hello Gary, yes I have. All CFNR GP15-1 were ex CNW engines. All retained their nose gong bell.  (ESU is working on releasing a soundfile with this gong bell included, and I'll be uploading those to my decoders once available).

They (CFNR) operated them in California until the Gensets were purchased. They then disappeared (apart from one or two stationed near Davis CA) of their home rails and can be found as far away as Texas. Now that parent company Rail America has been purchased by the Genesee&Wyoming some GP15-1 have made it back on those rails, be it in orange GWRR livery and I also am not convinced they're ex CFNR units.

I still like the CFNR colours the most on them.

yeah, its an attractive livery for sure..

[Alcanman]

Does anyone know if the new ESU Loksound Select 21 pin sound decoder mentioned at the beginning of this discussion is available yet.

 

 

 

Mal

[Scottish Modeller]

Hi Mal,

Does anyone know if the new ESU Loksound Select 21 pin sound decoder mentioned at the beginning of this discussion is available yet.

 

 

 

Mal

Tony's Trains are listing them as available...

 

http://www.tonystrains.com/products/esu_loksound_select.htm

 

Thanks

[Alcanman]

Thanks Phil.  Looks like the the decoder in question is the 4th one down which states 'image not available'. It has 6 functions.

 

My next question is how do you fit a 21 pin decoder when most dcc ready locos have 8 pin sockets? Is there an adaptor similar to the ones available from Bachmann for their UK models?

 

 

Mal

[Gary H]

Thanks Phil.  Looks like the the decoder in question is the 4th one down which states 'image not available'. It has 6 functions.

 

My next question is how do you fit a 21 pin decoder when most dcc ready locos have 8 pin sockets? Is there an adaptor similar to the ones available from Bachmann for their UK models?

 

 

Mal

Mal, is this any good to you??

http://www.dccsupplies.com/shop/product_info.php?products_id=1250

[Alcanman]

Thanks Gary.

 

However,  this adaptor appears to be for opposite situation. ie Fitting a 8 pin decoder into a loco which has a 21 pin dcc socket.

 

What is required for the Loksound Select 21 pin decoder is an adaptor to fit into a 8 pin dcc socket, I think.....?

 

 

Mal

[Prof Klyzlr]

Dear Pete,

 

Nice tickover, but what is that distortion on the horn sounds? Not bad on my monitors in general but how do they eq it on the "master", though? Ultimately it has to come out of tiny speakers.

 

Best, Pete.

 

Um, arguably this is how it _should_ be done, and no, no DCC decoder manufacturer is going to do it for you...

 

 

 

 

'nuff said...

 

Happy Modelling,
Aim to Improve,
Prof Klyzlr

 

PS in the same way that children listening to "Doof Doof" on earbuds and 1" "iPod speakers"
(which are essentially the same as the "high bass" speakers we use in our locos!!!),
appear to get "bass" from "can never produce BASS" sized drivers/enclosures,

there is no reason why we can't use the "Radio Mix" 200-300Hz "faux bass" EQ hump as used by Top 20 music producers the world over to get "apparent diesel rumble" from our scale locos...

[trisonic]

Interesting.

BUT if you take the earbuds out of your ears and listen the "bass" magically disappears..........or should I say that in reverse?

I have to say that I'm unimpressed by Ipad speakers...

 

What do you think of mini speaker enclosures being described as "Bass Reflex"? I  say that they appear to be just randomly "ported". I hasten to add that I've not seen/heard all of them by any means.

 

Best, Pete.

Edited December 22, 2013 by trisonic

[Prof Klyzlr]

Dear Pete, Sound-serious modellers,

 

Maybe the pics are a little too cryptic. To explain:

 

- the first pic shows an Athearn switcher loco, with installed speaker (type and format not germain) wired directly to a 3.5mm minijack. This is not a photoshopped image or trick, the speaker/enclosure installation is 100% "real" and as-intended for the final DCC decoder sound design.

 

- First test, really simple, plug the "speaker enclosure on wheels" to an MP3 player or similar, and play some songs or other full-range signals you're familiar with. This gives the modeller a real-world reference as to the acoustic performance of the speaker, enclosure, and loco shell/mounting. It's not a Audio Precision grade analysis thing, it's a "Trust Your Ears" thing. However, it gives a very "real world" basis on which to explore various combinations of speaker/enclosure/mounting for the loco in question.
(We are no longer talking hypothetical or "I believe..." type assessments of any given onboard speaker performance, which are commonly sprouted as "fact" by various modellers with literally no actual hard Audio Science analysis or background,
but rather we're using known, full frequency, comparatively uncompressed high-res audio as a basis for _real_world_relevant_ aural performance assessment...)

 

- Second test, a little deeper. With some relatively-relevant general aural idea of
"what's missing",
"what's exagerated/over-emphasised",
"what radio-mix 'percieved bass' tricks and imaging effects we can apply"
and if there are any acoustic issues (rattles, buzzes, resonances, etc), within the loco shell and mech itself,

 

we can plug the loc...oops...I mean "speaker enclosure on wheels" into an audio editor, (Audacity is free),
and actively go hunting for the frequencies in question.

 

This is a lot easier than it sounds,

(pun not intended ;-) ),

 

and can be approached in 2 ways
(I do both, but either will give some interesting results)

 

NOTE! Before proceeding, its important to recognize that
- there is no reason or point in trying to make a scale model loco reproduce the 120+dB SPL volume of a prototype loco
- a given signal that you "know how it sounds" at x dB SPL will sound _Different_ when played at < x dB SPL.
(both high and low frequencies will appear to be significantly-less-obvious, whilst the mid frequencies will appear louder.
It's called "Equal Loudness Contours" and/or "Fletcher Munson Curves". Google is your friend...)

 

Technique 1 - Play the same song/soundfile you tested the loco with above, at a _Sane_ playback volume, and play with the EQ filters in the Audio Editor until the "issues" identified are minimised, compensated for, or if the physical speaker/enclosure/mounting is suitably chosen/assembled/optimised, completely eliminated.

 

Technique 2 - Use Audacity to generate a series of test tones, fire them in sequence thru the speaker, and make a 1-5 "subjective assessment" of the comparative percieved-loudness of each tone. Once the subjective levels have been noted, the EQ filtering could be applied to "equalise" the relative levels of each tone.

 

FWIW, Peter Ross from NZ created a LokSound decoder project with integrated test-tones which are triggerable by F-keys, for exactly this purpose.

 

The end result of either/both tests should be an EQ curve which represents the good, the bad, and the ugly of the speaker installation in question, as judged by the only ears that matter, those of the modeller/owner of the loco in question. (In the photo above, you can see a 31-band graphic EQ curve on the laptop screen, which rolls off the extreme highs and lows, and adds a "low-mid" bump around 200Hz to provoke a enhanced sense of "bass". FWIW, this nicely matched the 13mm exciter installed in the locos shown, and gave a nice "flat" percieved reproduction for all frequencies 100Hz-16kHz @ 72dB SPL-A @ 1m).

 

Now, what to do with this info?

 

Well:
- given that the TSU decoders have onboard 7-band EQ, the "compensating EQ values" established in the Audio Editor can be literally transcribed/recreated into the TSU.
(Check the TSU manual, the relationship between CV numbers and EQ frequencies is clearly set-out. JMRI is not required, although arguably can make this programming process easier).

 

- for User-loading decoders (thinking Lok particularly) the resulting established EQ curve can be applied to all component sound files in the editing/sound-production process. These "loco-specific optimised" files can the be uploaded into the decoder with the knowledge that the files will sound _significantly_similar_ despite the extrusion thru the decoder upload process.
(Note! For such situations, applying the EQ via Parametric rather than Shelving EQ may be a useful strategy. Finer, more surgical EQ application, allowing better "cleanup and compensation" while not overly-processing too much of the available frequency range).

 

Pete, I take your point that small domestic speakers don't really hold a candle to high-spec studio monitors or concert PA bins. However, a quick check of the specs for even modest drivers show a frequency response down to 200Hz, and some seriously threatening to push 100Hz.

 

http://www.parts-express.com/tang-band-w1-1942s-1-full-range-neodymium-driver--264-924#lblProductDetails

 

http://www.parts-express.com/tang-band-w1-1931s-1-neodymium-full-range-driver--264-923#lblProductDetails

 

http://www.parts-express.com/tang-band-w1-1828sa-1-neodymium-full-range-driver--264-883#lblProductDetails

 

http://www.parts-express.com/tang-band-w1-1815sa-1-neodymium-full-range-driver--264-882#lblProductDetails

 

http://www.parts-express.com/tang-band-w1-1070se-1-full-range-driver--264-860#lblProductDetails

 

Now while AES/SMPTE definition, the term "Bass" starts at 100Hz and goes _down_ from there, the reality is that the vast majority of modellers wouldn't recognize AES/SMPTE-spec Bass unless they were in an earthquake. Equally, when as modellers we say "bass", what we're actually expecting/describing is the kind of 200-300Hz "doof" common evident in the "radio mix" versions of Top 20 dance tunes, IE "sounds like bass" when played back on typical radio or iPod style speakers. Given the above speaker choices and signal source optimisation technique, such "perception of bass" low-mid processings is more than achievable, and with a modicum of "thinking with our ears" 100Hz is definitely within reach of a HO install...

 

Happy Modelling,
Aim to Improve,
Prof Klyzlr

 

PS Yes, I have to laugh at some of the terms used in the model RR sound arena to describe various elements. "bass reflex", "tuned port", "close coupled array", and "hyperbaric" are all terms which have been (mis)used to varying, in some cases spectacular, degrees in recent times...
(Dr Amar G, the team at JBL, John Meyer, the original Disney Enginears, and many other audio pioneers would turn in their graves...)

[torikoos]

Hi Prof, an excellent tutorial. I'm no stranger to sound engineering myself, but I think your approach above is quite a good way of determining what your speaker installation is capable of.

Not all DCC decoders have elaborat EQ controls, but amongst others Tsunami's do, and you can fairly crudely match the sound response as you demonstrate with the 31 band EQ in Audacity. 

 

As you say, trust your ears , play around with the EQ bands, and set it as it sounds good to you.

I've found a boost of the frequencies around 125Hz (which I believe is one of the low frequency steps available in a Tsunami) will work. You might as well leave the even lower frequencies untouched, or minimise them (roll off in technical terms), as they will take up precious energy of your on board sound amplifier, yet you can't hear them from such a small speaker, and could only introduce some distortion. Similarly I minimise the high EQ frequencies too, to prevent the sounds to appear to scratchy or shrill, again I judge by ear, if things start to become too muffled, I increase these a little. 

 

You cannot expect such small speakers to move huge volumes of air needed to produce convincing low bass tones. In the words of Scottie :' You can nee change the laws of physics'. 

 

Koos

[trisonic]

Damn, is that true? I've been trying to cram a Celestion "Blue" into an HO SW1500.......but John is right .

 

Most American diesels I've heard do tend to be Mid range rich (depending on your relationship to them Hello Mum!) and as we know high freqs tend to dissipate first with distance. Close Low freqs tend to make themselves known through the soles of your feet (well, they do mine)  presumably due to direct contact. So we can ignore those too.

 

Best, Pete.