Reverb + EQ + raw "Direct" dry signal = overload/distortion
Dear Ken,
Sorry for the lag, work's been crazy...
Quote:
The reset made the crazy sound disappear.
Which tells us a few things:
- soldering the pickup wires etc helps from a reliability standpoint, but is not directly related to this issue.
(If the decoder was not showing any other symtoms of bad pickup,
such as flickering lights or stuttering motion, I would tend to focus on the decoder settings for expediency)
- virtually confirms a decoder setting issue as the culprint
It also points to a question that needed to be asked very early on,
"...has the unit always sounded bad from Day 1 out-of-box, or did it go bad at a given point-in-time?..."
Quote:
That being the case I reloaded the settings from DecoderPro3. Sounded bad again.
Great, "bad settings" captured live in time and space. If you can get at the Decoder Definition (Roster?) file for the loco-in-question, anyone with a compatible version of JMRI should be able to open it, and inspect the questionable values.
Quote:
I went into the sound settings and turned off the reverb and set the equalizer for the correct size speaker.
Did you turn off Reverb,
Retest
then reset the EQ
and Retest again?
IE did you seperate out these 2 settings to isolate if either one was "more involved" than the other?
From an Audio P.O.V.
(Skip if not deeply interested... 
)
- you start with an audio signal
- that audio signal is at a given digital/electrical/physical dB level
- Summing 2 or more signals together gives a cumulative dB output level higher than a single signal on it's own (logical)
- the resulting "sum of the signals" has to pass thru a D/A stage, a power amp, and to the speaker,
where each of those elements in the signal path have a limited (max) dB signal they can pass cleanly
(IE without audible distortion).
If we add excessive "Reverb" to the source signal (raw prime mover + horn + whatever else),
it is entirely possible that
- we can overload the Input to the Reverb process, creating "distorted Reverb"
- we may be feeding clean signal to the Input of the Reverb process,
and may even be getting a clean Reverb signal Out,
but the SUM of the Reverb effect PLUS the non-effected "Direct Sound" could create a Cumulative distorted result.
EQ is a powerful audio processing tool, and if you turn any significant number of the frequency band settings (CV 153-160) UP to their maximum (255), you are actually adding a significant ammount of dB boost to the overall signal.
Some numbers (I don't expect anyone to remember them) to give context:
- 0dB is considered "Unity" (what comes in, goes out again exactly the same)
- a dB value with a - in front is a lowering or "cutting" of the level
- understandably, a dB value with a + in front is a raising or "boosting" of the level
- +/- 3dB is considered the minimum change of level that the average human ear can reliably detect
- +6dB is a doubling of percieved volume (literally "twice as much signal" or aurally "twice as loud")
Now that we know what the correct terms are, you'll appreciate that each EQ CV in a TSU has
- CV value 128 = 0dB in audio processing terms
- CV Value 0 = -12dB (a Drop in audio level, resulting in 25% of "normal")
- CV Value 255 = +12dB (a Raising of the level, equivalent to 4x the "normal" signal level/volume)
Given that the Master Volume of a TSU is factory default set at "near maximum" (CV128 = 250 IIRC),
and just has enough enough "headroom" (spare signal range) to cleanly pass the Prime mover and horn without overloading (IE audibly distorting thru the D/A and power-amp stages),
adding the extra Reverb signal
OR
boosting a few EQ bands
is certainly enough to create a cumulative signal larger/louder than the restricted range of the D/A converter and Power amp can handle...
(NB that AFAIK, NO DCC Decoder Manufacturer has Ever Publicly Released the Real Audio Specs of their signal paths/circuitry. As such, we have No Idea what the actual Dynamic Range or Signal/Noise specs of the digital mixer, DSP, D/A, and power amp stages are).
I'm glad you got it sorted in "brute force" mode,
but I think that some consequent playing and "interaction observation" of:
- the Master Volume CV 128
- the EQ settings
- The Reverb Setting
- and possibly the discrete sound volume settings
would pay off in nailing down exactly what was happening...
Happy Modelling,
Aim to Improve,
Prof Klyzlr
PS Hopefully the astute and observant will have also deduced that by cutting the EQ bands in the mid frequencies, you get the aural equivalent of turning up the Bass and Treble bands,
(the "graphic EQ" shape still looks like a smile),
but with the benefit of actually increasing the available total headroom.
(Instead of boosting the frequencies you do want,
and threatening to punch into Overload/distortion territory,
you're cutting the frequencies you do not want,
freeing up headroom to reproduce the remaining wanted sounds).
This was discussed in DuckDogger's "SD40 sound" thread from a month or so ago...
(search "scoop the mids").
