I intend this as a theroretical question only.

There was a thread last year about normalization. The Mk2 has a mike input. Could a variant of that software be used to increase or decrease volume based on the level of road noise present?

-- Glenn

There was a thread last year about normalization. The Mk2 has a mike input. Could a variant of that software be used to increase or decrease volume based on the level of road noise present?

I've been wondering about this for a while. One question is where do you mount the mike to measure road noise? Sounds doable, though.

Borislav

In car audio systems with active feedback built in, the sensor mikes are usually built in (in pairs) in the driver's headrest. Lotus did this with their early systems, and then went to a multi-mike system for their active noise cancelling stuff that used the amplifier to inject anti-phase noise. Very successful, never once implemented by commercial manufacturers (the plonkers).

Maybe that's a better idea - instead of active level compensation, where you're always fighting ambient, just inject an anti-phase component from the sensed environment. Then you just cancel it out without affecting your audio!

(You would have to consider patent issues if anyone tried this commercially, but I suspect they may well be coming up to public domain soon)

One of the few remaining Mk1 owners... #00015

In car audio systems with active feedback built in, the sensor mikes are usually built in (in pairs) in the driver's headrest.

Notice that this thread is not about noise cancelation (which we've discussed at length and the consensus was that it's way too hard to do in a random vehicle) but about something much cruder: increasing the volume as road noise increases (and maybe boosting the bass?). This means that you don't actually want any feedback from the player, or it'll just play louder and louder until your speakers blow up (or your eardrums).

Inside the main compartment or the trunk/boot it'll be getting feedback from the speakers. In the engine compartment is probably no good either, engine noise would vary too much when you step on the gas and is probably not a good measure of "road noise". I was thinking that the mike should ideally be outside the vehicle but still be protected somehow. Hidden under the rear bumper? Rain will still get to it. Something along those lines, I still haven't found the perfect spot.

Borislav

Wouldn't it be more efficient to sense the speedometer for speed and adjust the volume accordingly. I would go crazy if the volume went up and down when passing large trucks. What to think about railroad crossings?. It would be sufficient for me if the volume goes down when stopping in front of a traffic-light (it would stop people staring and pointing to their heads )

Frank van Gestel

Wouldn't it be more efficient to sense the speedometer for speed and adjust the volume accordingly. I would go crazy if the volume went up and down when passing large trucks. What to think about railroad crossings?

Good points, it probably makes more sense to connect it to the speedometer, except that's likely to be much more vehicle specific and require some additional hardware (I'm sure some people on this board can come up with something like this while brushing their teeth). Just hooking up the mike should be much easier from the installation point of view. If it reacts really slowly (say 1db/10sec) then it won't be affected by temporary changes in noise level yet still adjust properly when getting on/off the highway etc.

It would be sufficient for me if the volume goes down when stopping in front of a traffic-light (it would stop people staring and pointing to their heads)

Hasn't happened to me yet, I probably need to turn up the gains on my amp.

I would go crazy if the volume went up and down when passing large trucks. What to think about railroad crossings?

I would have thought that you would want the volume to go up whenever the ambient noise increased, whatever the source. The problem with mount a mike outside is that it doesn't take into account the (huge) difference between windows up and windows down. I'd imagine this difference would be even greater for those of us whose rooves (sp?) come off.

It seems to me an in-cabin mike could be better, but I don't have a solution to the feedback problem. Also, it would be interesting if you were trying to talk to a passenger above the music, but every time you raised your voice, the music got louder to compensate! :)

Richard.

Hi,

Interesting discussion this one ...

We know what the empeg is currently outputting, so the mike could subtract this from the input.

As not all cars and speaker setups are the same, there could be some sort of calibration option, where the mike measures the sound level in a "silent car" while the empeg outputs music/sound.

This will give a crude idea of what the noise level is.

Road noise seems to me to be in a certain range of frequencies. So with a FFT that range could be isolated, and the volume adjusted.

With the same method the volume could be decreased if someone speaks (which also would make it impossible to be singing in the car :)

Just some quick thoughts ...




Lars

I agree - but I only suggested antiphase noise cancellation since I can't recall having read a discussion on this board about it, and I've been reading since the beginning. I'm pretty sure I would have remembered something like that on the board, as it interests me - I have been looking at doing exactly this with a Mk2 (if I ever get one ). Did I miss something interesting? Can you point me to the thread?

I think you would be better with speed-sensitive level control, with a base level setting; this would avoid ambarrassment when you are stopped by the coppers in heavy traffic ("What did YOU SAY OFFICER? Sorry about this IT JUST KEEPS GETTING LOUUUDDDDEERRR!") - a closed loop, positive feedback situation. IIRC, there is supposed to be a wire definition (rarely used) in the ISO connector definition for level control for just this - allowing a unit to sense speed.

Even if you implemented this, then surely you want the level to be set relative to what the driver can hear at his hearing position - so I'd stick by the microphone in the seat's head restraint as being the perceived audio level relative to which you want to set the volume to get a fixed dB(A) delta. This does strike me as somewhat dangerous to the health of your hearing and also your amp...

Plus, if you boost volume in response to ambient noise levels, then what happens when you try to use the long-awaited VR? You speak to your Empeg, and it boosts the volume as the VR software is trying to recognise phonemes, thus reducing the S/N ratio and making it progressively more difficult to tokenise the longer you speak.

And what about just talking to someone in the car - it would get progressively louder as it integrates with time and you'd end up screaming at the person next to you if you can't find the remote in time....

One of the few remaining Mk1 owners... #00015

Then you just cancel it out without affecting your audio!

Ummm... how would you generate the anti-phase audio? If you are using your stereo to generate it, I suspect there would be considerable impact on the sound quality when the stereo was trying to do "double duty". Or would you have a separate amplifier and speaker set for the anti-phase sound?

Getting the anti-phase sound pickup to differentiate between extraneous noise generated from outside the vehicle, and the in-cabin noise generated by the stereo system might prove difficult. (No problem with my Mozart, but with something like Nine Inch Nails I'm not sure I could tell the difference myself, much less have it done with a microphone and a computer algorithm.)

tanstaafl.

"There Ain't No Such Thing As A Free Lunch"

Doug, what I want to know is...

Have you ever even heard any Nine Inch Nails?

I only ask because your description is uncannily accurate (random noise is a big part of Reznor's productions), but you don't strike me as the type who would own a NIN album.

___________
Tony Fabris

Have you ever even heard any Nine Inch Nails?

If I have, it certainly wasn't on purpose!

No, my impression of NIN is derived strictly from the posts I have read on the bbs.

tanstaafl.

"There Ain't No Such Thing As A Free Lunch"

I only suggested antiphase noise cancellation since I can't recall having read a discussion on this board about it ... Can you point me to the thread?

There were a few good replies when I brought this up in August. "At length" was a huge overstatement, I thought there was more than that.

Even if you implemented this, then surely you want the level to be set relative to what the driver can hear at his hearing position

Yes, but what you want even more is no feedback (you gave some good examples why not), thus the suggestion to put the mike outside the vehicle and assume that the noise the driver can hear is a linear function of the noise outside.

Plus, if you boost volume in response to ambient noise levels, then what happens when you try to use the long-awaited VR?

You won't be able to use both at the same time. That's OK with me, I'm not terribly interested in VR (says he while donning his asbestos suit - flames are guaranteed to come when you utter such a heresy) so I might as well put the mike to good use.

Borislav

The problem with mount a mike outside is that it doesn't take into account the (huge) difference between windows up and windows down.

The absolute difference you can adjust for manually when you roll the windows up or down. That still leaves the adjustment speed - I suspect that ideally it should be more sensitive with the windows down. I don't know, you can either make a compromise that works relatively well in both scenarios or you can use a non-linear function (i.e. if the volume is already high then you probably have the windows down, so make bigger adjustments). Hmm, finetuning this might be more difficult than coming up with the perfect equalizer setting.

Borislav

I think that the Speedo sensing with a per vehicle map of volume response would be a practical implementation of the intent I had.
There could be a graph displayed where the X axis has incremental speeds (say every 16.09kph ) each increment would have a delta volume. The owner sets those according to his percived needs. Another way, maybe better, would be to associate a specific eq with each speed.

Either way the software interpolates between fixed speed increments.

I must confess I don't have a clue how the road speed could be collected from the total variaty of cars on the road today.

The gingerbread is: An empeg visiual of a road endlessly appearing from over the hill with road signs showing track info flashing by at X miles and hour.

-- Glenn

Speedreading could of course be done with a GPS.
Has anybody gotten to use the serial-port for something
other than the standard interface? I guess /dev/ttyS0 (or
whichever is used) needs to be replaced by a pipe (or a symlink), while a new /dev/ file is created for the serial port. But. How do we get a command-prompt? telnet? USB? IrDA?

Marius (Escort Cab + Mark II)

(says he while donning his asbestos suit - flames are guaranteed to come when you utter such a heresy)



One of the few remaining Mk1 owners... #00015

Ummm... how would you generate the anti-phase audio?

Using a DSP and a lot of math. I suspect that the SA1100 would not really be up to it, to tell you the truth.

If you are using your stereo to generate it, I suspect there would be considerable impact on the sound quality when the stereo was trying to do "double duty".

It does depend somewhat on the power output of your amp (you wouldn't want to use a 2x 20W RMS amp, for example ), but Lotus showed it did not require too much to do it.

Getting the anti-phase sound pickup to differentiate between extraneous noise generated from outside the vehicle, and the in-cabin noise generated by the stereo system might prove difficult

Well, all cars have a definite accoustic "Fingerprint". At different engine and road speeds, the fingerprint varies somewhat, but not a significant amount. It is mainly in the amplitude that the fingerprint varies across the range. Before you attempt cancellation, you have to characterise the vehicle's signature, then after that you simply subtract it from the cabin noise by inverting it and playing it out. The extent of the cancellation depends on the amplitude of the antiphase, which is (as you can guess) dependent on your amp.

However, you don't need to totally noise cancel, as only a relatively small level reduction of particularly intrusive fingerprint components can produce a subjectively large improvement in listening environment quality. Hence, no really big special amp necessary.

It would be truly excellent if the Empeg had the processor capability and the signal path needed for the injection of antiphase. This would really make it an absolute killer app for the road.

If the Mk3 included a DSP in the signal path (not the output processor that exists now) for signal matching and injection (plus other unusual effects ) then my cash would hit their coffers so fast they wouldn't know what hit them....




One of the few remaining Mk1 owners... #00015

I'm not terribly interested in VR

OK, I've got some gasoline and a stake... anybody got a match?

tanstaafl.

"There Ain't No Such Thing As A Free Lunch"

corner of the headliner/a pillar,like where they mount cell phone mics. In fact, someone could USE a small cellophone handsfree mic as a prototype, couldn't they? :)