Jump to content
 

CSX Transflo, NY ...another small ho switching layout


Recommended Posts

Hi Mal, nice video.

 

Just curious does that U30B have a QSI decoder? It sounds a bit like it does, it has some characteristics I seem to recognise.

The 'sound of power' effect tends to distort a little (not the fault of your speaker, it either overdrives the build in amp or it's the sound file ) when accelerating, and the 'pitch' of the engine seems to change somewhat unnatural. I have similar effects with an EMD 567 engine sound on an SD9 I own with a QSI decoder, and tends to demonstrate my remark in your other thread on decoders.

 

Koos

Link to post
Share on other sites

Thanks Koos.

 

The Atlas U30B has a factory fitted QSI decoder and speaker.  I'm a complete novice when it comes to sound, however, the engine sounds rather high pitched to me.

 

Sounds are better when the engine is idling, as it seems to have more bass.

 

However, despite some of the shortcomings of dcc sound, I'm completley hooked, although I do find that having only one loco on the layout is sufficient.

 

Even 2 locos in sound can become irritating after a while.

 

 

Mal

Link to post
Share on other sites

Ah see I thought I recognised a QSI. :-)

 

That high pitch is a QSI 'issue', they simply 'speed up' the sound file (a bit like an old vinyl record from 33 rpm to 45 rpm etc) to make the engine sound as if it is running faster (you would get that chipmunk effect if it were a voice) . In a real engine although it runs faster , it is just more repeats of the pistons moving per time unit at the same low rumble, therefore you shouldn't loose any bass.  This is what Soundtraxx does very well, and also ESU does well in their newest sound files.  

 

I agree with more than one engine it all can get a bit much in a quiet room, and I generally turn the volume down at home (usually engine and bells I turn (much) lower than a horn).

 

Anyway, this answer really belonged in your other thread. :-)

 

Koos

Link to post
Share on other sites

If the sounds are digitally recorded  that should not happen. I can speed up my regular (music) digital recordings without any increase in pitch - though I do have the option to do so if I wish. If they are doing this then they must have made a conscious decision to do so.

 

I'm, still not sure just how any company records the sound of locomotives - the only true way would be to drive alongside (but not on) at precisely the same speed - somehow I don't see them doing that, frankly...

 

Best, Pete.

Link to post
Share on other sites

Then that would not be accurate either - if we assume that they are trying to model the sound of the locomotive from a bystander's point of view. Even close up we tend to look at model locomotives from about a scale 20 yards away. After all a sound of a Ferrari is different from the driver's seat to someone who it is passing (more induction noise vs exhaust noise).

Come to think of it maybe the higher pitch sound they use is to try and get (from the turbocharger) turbine whine - which certainly does increase in pitch under load. That does not apply to all locos, I know.

 

Best, Pete.

Link to post
Share on other sites

Having thought about it a bit more I would record a modern loco by setting a ten foot long boom with a wind shielded stereo mic projecting out at a 90 degree angle perpendicular to the length but pointing up at 45 degrees from the mid-point of the loco using some kind of Mac laptop with Garageband installed.

More chance of using the boom on an American loco of course due to more clearance and a walkway to set it on.

 

Best, Pete.

Link to post
Share on other sites

If the sounds are digitally recorded  that should not happen. I can speed up my regular (music) digital recordings without any increase in pitch - though I do have the option to do so if I wish. If they are doing this then they must have made a conscious decision to do so.

 

I'm, still not sure just how any company records the sound of locomotives - the only true way would be to drive alongside (but not on) at precisely the same speed - somehow I don't see them doing that, frankly...

 

Best, Pete.

Hi Pete, your music software uses some advanced processing, to keep the pitch at the same pitch (same with slowing down one) , however simply 'speeding up' a sample without doing anything else to it, does have the pitch shift effect, look at how early samplers (and digital music workstations) worked that way, and that's the kind of processing that happens on these fairly cheap sound chips.  Soundtraxx does make a variety of recordings, and recalls the right ones from memory with processed transitions for each speed step etc. I'm not sure which location they put the microphones. There are some video's on youtube showing how Rapido trains does theirs, quite interesting to see actually.

 

Koos

Link to post
Share on other sites

Koos,

I think you'll find that even the modern "el cheapo" recording software available nowadays that pitch change is not the default but an option; it is in the nature of current digital recording. It's more likely a cheap way to engineer speed differential.

 

What I really don't know much about is the "interactive" nature once the recording has been done and mixed. I'd hate to really do this by the way, it's a lot different from just passive listening to a recording.

 

Of course if they were still using analogue then I'd agree with you.

 

Best, Pete.

Link to post
Share on other sites

Hi Pete, 

 

I agree modern software does what you describe, but I doubt that the sound chips in some decoders have the same algorithms, of course technology keeps evolving, and it will surely come into reach of the newer generation of sound decoders (some might already have some).

I believe Memory is cheaper at the moment. It's therefore easier to record a series of (engine) sounds at various rpms, and interpolate between them between speed steps, which is what some manufacturers seem to do to my ear.

QSI speeds  it's samples up which is why they start to sound unatural and start loosing 'bass' at higher revs. I don't know if they still do that on the new Titan range, but on the 'Revolution' that can be clearly heard and I have no other explanation for that effect in those decoders other than speeding up the sample to emulate the higher revs (and also pitching up the low end sounds at the same time, not what they wanted I guess).

They should (and perhaps have done) process those recordings again as you rightly say, you can nowadays do time stretch/compression , in other words, speed up/down without altering pitch, and use that. 

 

Does anyone here have a QSI Titan decoder to listen to? :-) 

 

How they (manufacturers in general) process the sounds etc is probably a secret, but they'd be using sophisticated audio recording software (Logic Pro on a Mac for example, together with processing plug ins) to edit them to make them fit on the sound chip and work on the various volumes etc It would be interesting to see.

 

Koos

Link to post
Share on other sites

Mal

Like the video,that U30B sounds like my B30-7,it's got QSI decoder,it just dosen't sound right to me,not a lover of QSI,like Soundtraxx for engine sounds,but poor horns.

 

Jon Grants got a lovely sounding U18B,think it's a QSI, you can here the pitch change,but when it's chugging through the layout it sounds dam good.

 

I agree about to many engines with sound just becomes a din.

 

Ray

Link to post
Share on other sites

From comments posted here and on my other thread regarding dcc sound, it would seem that the consensus is that Soundtraxx sound is superior to QSI, and I'm inclined to agree.

 

I thought it might be interesting to 'tweak' some of the sounds on my Tsunami fitted Athearn Genesis MP15AC. I increased the mixer volumes slightly and then added some 'notch up' at lower speeds to simulate a heavy load.

 

Here's how it sounds

 

 

Mal

Link to post
Share on other sites

Tsunami's are definitely very good sounding, and with a lot of patience, you can get your locos to behave for low speed use very well too, as you show in this latest video (a nice one at that :-) ) . However, I've just installed a ESu Locsound Select in a GP38-2 loco, and that sounds very good too, it's very easy to set up (the auto tune function seems to work in 95% of the cases) and it operates (runs) superior to any other decoder I own. They really seem to have it covered in that respect, and I find that probably more important than superior sound. 

 

Cheers, Koos

Link to post
Share on other sites

Thanks Koos. I've noticed that in addition to Atlas and Bowser, Intermountain are also starting to use ESU Loksound in their factory fitted locos.  They are also selling the full range of ESU products.

 

 I had to do a lot of 'tweaking' to get the MP15AC to run slowly and smoothly for switching. I agree that this is more important than sound.

 

I have a few UK locos with ESU decoders and they sound good and run smoothly, so I'll definately be trying out ESU in the future.

 

 

Mal

Link to post
Share on other sites

To get slow speed control with Soundtraxx decoders it's neccessary to alter CVs. Primarily CV 209 and CV 210. This has to be done by trial and error until you are happy with the results.

 

There is plenty of helpful info to be found, such as this website http://www.mrdccu.com/curriculum/soundtraxx/tsunami.htm

 

Just scroll down to the heading 'I can't get my loco to run well slowly'

 

 

Mal

Link to post
Share on other sites

Thanks Koos. I've noticed that in addition to Atlas and Bowser, Intermountain are also starting to use ESU Loksound in their factory fitted locos.  

 

Yes I have too, and others have made the same observation. ESU is defenitely starting to make a push in the North American market. It is probably a combination of factors: There's more sound files for them available in their range now, the release of the Select Direct series, compact size, no real cooling needs, and their high quality standards. 

 

Koos 

Link to post
Share on other sites

I thinks its time that a DCC wheel squeal freight car was made  .After all we have a mechanical reefer sound now ..All great stuff and love that gnotching up .

Martin

I'm not sure if an on-board sound for flange squeal is the best solution -- it comes from outside the train, basically. ITTC electronics http://ittproducts.com/products.html makes an HQ105 track sounds module that includes a squeal -- you can test it on their site. I've found from experience that he'll load a custom sound onto a module at no extra charge if you provide the wav file. I use his module for a detector message and am very satisfied. The modules will run once with a pushbutton or continually with a permanent contact switch -- if you use a detector for a block, you can play the track sounds continually while the train is in the block; the sounds will go off when the train leaves the block. Or, as in the case of a detector message, you can use a momentary block detector like the TRAK-DTT2 from Dallee Electronics.

Link to post
Share on other sites

Dear RMwebbers,

 

OK, I'm hearing a some Bad Audio Science, and some Educated Guesses, but little solid info. To wit:

 

- RE Recording

 

A quick look at the recent MRH provides a good review of the SoundTraxx location recording process.
http://model-railroad-hobbyist.com/magazine/mrh-2013-10-oct/di_getting-the-sound-in

 

FWIW, they appear to be using a ProTools + Command-8 rig for the post-production and editing

 

QSI have long had a "how to record loco sounds" PDF available, which gives some (veiled) insights into their particular production workflow.

http://www.qsisolutions.com/products/techinfo/record-sounds.html

 

 

 

- RE Pitch-shifting for "speed"

 

While statements RE the audio processing capabilities of even free audio editing apps such as Audacity are basically true
(IE that a sound can be sped-without-pitch, pitched-without-speed, and both pitch+speed changed as required),

 

the issue is not related to PRE-decoder-loading software, but rather the playback DSP limitations of the Playback hardware of the Polyphonic Digital Audio Player
(which is effectively what a DCC sound decoder is).

 

From waaaay in the mid 1970s when the Aussie Fairlight CMI pioneered digital audio recording and replay,
http://en.wikipedia.org/wiki/Fairlight_CMI

 

the issues of pitching/speeding-up/down a digital audio stream in real time (and associated samplerate-math conversion processing) have been dogging musicians and engineers alike. It is entirely possible to engineer a hardware digital audio replay system which can perform realtime speed-without-pitch changes. However, to do that in a Decoder-sized circuit, while using known poor-performing generic microprocessor and PMW D/A output stages, (no Outbound SRC or filtering!), not to mention actual DSP processing programming which is both known and very seriously locked-down patent-wise is a big hill to climb.

 

http://en.wikipedia.org/wiki/Sample_rate_conversion

 

http://www.eetimes.com/document.asp?doc_id=1204860

 

http://www.kvraudio.com/forum/viewtopic.php?p=4526522

 

http://www.dspdimension.com/admin/time-pitch-overview/

 

http://homerecording.com/bbs/general-discussions/mixing-techniques/fixing-mix-pitching-tempo-change-changing-samplerate-35742/

 

 

 

- RE "Flange squealing cars"

 

Instead of the ITTC modules, I would strongly suggest reaching for something a bit more potent in digital audio performance terms, with built-in model-rr-designed Trigger IN and Control OUT capability, and that can be eaily user-loaded without needing custom reprogramming at point-of-manufacture. Let's say, something like the Pricom DreamPlayer LITE?
http://www.pricom.com/Trains/DreamPlayerLITE.shtml

 

http://www.pricom.com/indexold.shtml

 

Such hybrid car/layout-borne effects have been thoroughly discussed and "feased" by the LayoutSound YahooGroup. If anyone in interested in persuing such, I'd strongly suggest checking in with the team there.  http://groups.yahoo.com/neo/groups/layoutsound

 

NB that there are some US-based modellers who have also levered the User-loading capability of the cost-effective Digitraxx "Soundbug" range to add DCC-controlled sound to otherwise non-powered rollingstock (thinking RCPs, mech reefers, stock cars, flange squealers, frame-creakers, draftgear-stretchers, etc). They give away a lot in terms of sheer audio performance (8bit 8kHz max), and User-friendly programming/uploading, but as a realtime DCC controllable digital audio playback device, it fits the mission-parameters relatively well...

 

I hope this helps...
(If anyone wants to talk sound furthur, suggest starting a new thread, or taking the conversation over to the LayoutSound list...)

 

Happy Modelling,
Aim to Improve,
Prof Klyzlr

Link to post
Share on other sites

 

-

 

 

 

 

 

 

 

 

 

 

 

 

I hope this helps...

(If anyone wants to talk sound furthur, suggest starting a new thread, or taking the conversation over to the LayoutSound list...)

 

Happy Modelling,

Aim to Improve,

Prof Klyzlr

Thanks to the Professor of Sound!

 

 I recently started a seperate thread 'DCC sound decoders for US Diesels'. Perhaps this thread or something more general like 'DCC Sound for US layouts' could be pinned?

 

Mal

Link to post
Share on other sites

I too am overwhelmed by the good Prof's stats..........

 

We have to move to more simplistic explanations and fully understand what exactly is required to have a model sound like the real thing (and even that depends on the listeners perspective relative to the model with respect to doppler and other similar effects).  What may sound right from one location will not sound right to another listener who's orientation is different.

 

I'm not sure that we are making things too complicated! Take Mal's latest  recording. It sounds to me that some kind of reverberation is present in the recording that makes the loco sound distant even when listened to at close range. Some things are very difficult, if not impossible, to scale.

 

Similarly I'm always amused at the description of "Bass Reflex" speakers to fit into a loco's body - when they are all (the ones I've seen and heard anyway) simply "ported".

 

Interesting discussion, though.

 

Your wish is my command, Mal.

 

Best, Pete.

Link to post
Share on other sites

Thanks Prof, I didn't really want to venture in the depths of audio processing. :-)

you mentioned the Fairlight CMI, one of my all time favourite instruments (and sadly way out of my league, not to mention I haven't got the spare room to put it into :-) )

I've got to make do with an old trusty Kurzweil K2000RS, but that's great fun too :-)

 

Koos

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...