Beyond mixed mode beyond multi tones

I added a post under the mixed mode thread, but I had a bit more to say on the general theme and thought I'd start a new thread.

when it comes to audio, recent detector design is moving in the direction of "hear what you want to hear" and "hear it the way you want to hear it". By this I mean assignable tones to different conductivities and variable volumes for each tone.

I expect that the detector design cycle which went from beep and dig to VDI to tones will continue to swing to more audio intelligence. For this to happen, I suspect that much greater processor power and more sophisticated algorithms for target analysis will be necessary.

I've been steadily impressed and won over by Keith Southern (and others) demonstrating the virtues of enhanced aural feedback from detectors. We have gome to digital machines and most of them to date have been beep or not - sometimes with multiple tones and lately with assignable tones and the ability to set individual tone volumes.

Keith has emphasised the extra intelligence furnished by "bleedy" "blendy" tones

We need more however. We need detectors that have pattern recognition capability - which they use to alter parameters on the fly - to optimize chances of success. Camera makers already do this with millions of "samples" of "targets" and lighting conditions stored in code - and then pattern matching the image being "seen" on the imaging chip in real time to this database to optimize settings like aperture, shutter speed and color balance.

This future detector would gather data, ground mineralization, targets encountered recently, user preferences for targets that they have recovered and "accepted" after examination vs. targets "rejected" after examination - and integrate it with the on board information to tighten the recognition of desirable targets. Accelerometers in the coil could assist in calculating target shape and size.

Why don't we have this yet? Probably because Nikon, Canon, Apple all have hundreds of engineers working on this kind of stuff and the TOTAL number of top flight metal detector engineers in the world probably doesn't exceed a dozen or two.

I suspect that Minelab's GPZ gold detector is doing some of this with it's ground balance system. That is a small first step in the direction I am talking about.

With even budget machines like the Fisher F44 offering highly customizable tones and volumes, detector makers are going to have to offer more to sustain the prices of their top line machines. More depth isn't likely - more INTELLIGENCE is likely.

Rick Kempf
Gold Canyon AZ- where there is no gold



Edited 1 time(s). Last edit at 02/29/2016 05:49PM by lytle78.

Great post!

"We need detectors that have pattern recognition capability"-The Whites spectrum xlt had a feature that was similar. It had different ways to analyze VDI numbers by, "average" and "Most reoccurring vdi number", I don't remember what they were called exactly, and I know you are talking way more advanced stuff, but that was a start.

Agreed, great post. Imagine what would happen if the metal detector companies "teamed up" with the big computer companies to produce small, efficient CPUs with reduced instruction sets specifically for metal detectors.

I can see a metal detector that "learns" based on what it finds as well as what the user wants, and there are so many other possibilities as well. Complex algorithms to help identify the target with more certainty. Advanced automatic frequency shifting based on EMI, ground conditions, or desired targets. Or maybe that would lessen the fun. Still, the possibilities are enormous.