The Unrealized Promise Of Multifrequency

[PimentoUK]

I have to disagree a little with you, Steve.
The fact that multifreq machines allow for superior handling of the ground signal ( with improved target ID, greater depth and good salt-beach performance being the benefits ) is the PRIMARY purpose of MF. The ability to give a better response to a wider range of targets from 'low conductivity' to 'high conductivity' is very much a secondary feature. And it's not a feature at all unless the engineers specifically designed it into the machine, either.
I do think that part of the 'intelligent' aspect of ML's Multi-IQ is the fact that it is still looking for 'low-conductor' targets even when you're using a 'higher-conductor' mode like "Park1", for example. Whereas older machines like the Fisher CZ's and ML Sovereign's are only somewhat tuned to their two frequencies, and definitely favour the low one ( 5kHz and 3.125kHz respectively )

[Steve Herschbach]

1 hour ago, PimentoUK said:

I have to disagree a little with you, Steve.
The fact that multifreq machines allow for superior handling of the ground signal ( with improved target ID, greater depth and good salt-beach performance being the benefits ) is the PRIMARY purpose of MF.

That's what I said, so we are in agreement. I was being critical of the marketing people making it appear otherwise.

[Steve Herschbach]

On 3/28/2022 at 12:50 PM, ☠ Cipher said:

When it comes to the V series, they tend to be fine in mild ground. With the right coil, even great. But as conditions worsen, moderate to harsh, or rapidly changing, they just don’t perform as well as other multifrequency designs. One of the Engineers, Geotech, explains that it’s the result of poor filtering of the frequencies, leading to frequency crosstalk, making a good ground balance especially difficult. If it were a better performer in such conditions I believe the complexity would’ve been forgiven. I have one and I love it, but I also have 0-2 bar soil, and rarely have opportunity to get to a saltwater beach. As they are, complex with compromised performance in difficult ground, the V series lends to too much chasing one’s tail for many people. For them, more time is spent looking for the holy grail setup than detecting. 

At the end of the day people just want detectors to perform for them. They can be simple, complicated, or insanely expensive, and all will be forgiven, if the performance is there. I was a very large White's dealer, and simple fact is that few MXT users could genuinely tell themselves that getting a V series ended up being for the better, and many went back. It's pretty much that in a nutshell.

[EL NINO77]

 ....I think the multifrequency for different detectors can be divided into at least 2 generations ...

1st generation Multifrequency - used the frequencies used to detect the signal similar to what single frequency detectors do - with "simple VDI signal processing"
but with the bonus that it made it possible to eliminate the salt signal to a certain extent..and this way these detectors were able to detect well on beach and in seawater ...

Such a multifrequency will have a problem displaying the correct VDI target ,,,, in difficult mineralized terrain conditions, or at deep targets .. there is a strong shift of the ID .., or the loss of the signal from the non-ferrous target..moving the VDI to the iron ...

2nd Generation Multifrequency ... uses multifrequency mainly for excellent work of the detector in more demanding field conditions..while eliminating magnetic, ferous  or conductive ground signal response .., but also for advanced calculation of correct VDI target for such demanding conditions ..., which guarantee much higher ID accuracy than 1-generation multi-frequency detectors, or 1 frequency detectors ...

However, of course, the 2nd generation multifrequency detectors can use various types of multifrequency programs to improve the detection capabilities of diferents .... terain types ...

I would classify FBS multifrequency detectors as the 1.5th generation of multifrequency ...

There is still something ... in some low-mineralization, or very specific conditions, even best single-frequency frequency detectors can work at least as well .... as multi-frequency detectors ....

On the other hand, you will only use it on demanding Terain ... 2gen...multi-frequency detectors ... because I will work there reliably and accurately and correctly ... and in many cases you can trust the target signal significantly more even on the basis of a reliable VDI ... before digging.

...But the important truth is that multifrequency allows you to move the development of VLF detectors much further ...., as would be possible with single-frequency detectors ....👍

 

[JOE WATSON]

Hi just a thought, but do you think metal detectors could ever utilise chirp technology as in the new echo sounders?

[mudwhale]

On 3/28/2022 at 12:32 PM, Steve Herschbach said:

To be clear here I am talking genuine multifrequency. Not selectable frequency misleadingly labeled as multifrequency, so we are now supposed to say “simultaneous multifrequency” or SMF, to differentiate. Sorry, I’m calling BS on all that. Selectable frequency machines are still running in a single frequency. Just because you can pick the frequency is not new. Been around for decades! When I say multifrequency detector I mean multifrequency detector, a detector comparing two or more frequencies to get a desired result. I don’t need to change my thinking. Marketers need to stop calling selectable frequency machines multifrequency. It’s misleading, I’m not playing along any more, and handing out thumbs down to those that continue to do it. Machines are either single frequency or multifrequency. If you want to your your single frequency detector as being able to select frequencies, fine. But quit calling it multifrequency. Even Minelab never stooped that low, and they had single frequency detectors that allowed you to select frequencies, far before most did.

So what was the promise? Or more accurately, the hype, of multifrequency? It has been that multifrequency detects all targets, large or small, high conductor or low, all at the same time. As opposed to single frequency, where each frequency is strong in one way, but weak in another. To this day I see this said in marketing, and I constantly see users repeating it as some kind of fact.

More BS. As anyone who follows this should know by now, the underlying frequency mix still rules. Minelab calls it “weightings,” where each mix is weighted higher or lower frequency, depending on the desired end result. So we have high frequency weighting for gold prospectors, and low frequency weighting for saltwater, as the most obvious divide.

Multifrequency does have an inherent strength over single frequency. Comparing at least two diverse frequencies gives more ground and target information to the system. This, in particular, generally results in better target id capability. There also is a very real ground handling advantage on mineralized saltwater beaches.

But you can’t make multifrequency detector that runs in all frequencies at once, that will do the best on everything everywhere. If so we would not have all these different modes, Park, Field, Beach, and Gold. The end result is not much different than a selectable frequency detector. You have high frequency mixes, or medium, or low. Each serves a purpose. I still have to choose frequency modes, and it hardly matters if I am saying high single frequency for nuggets, or high frequency weighted mix for nuggets. If people were not told different, they would not actually know it is any different, except for the extra target id and ground handling capability.

So stop with the “multifrequency finds all targets across the entire range” nonsense. There is no multifrequency detector made that runs in one mode, and hits tiny gold nuggets, while also working perfectly well at the beach. That’s what we are being sold, and it’s simply not true. This is from an old post of mine about Equinox, and it’s what we really have:

Hottest (more sensitive to small targets, hot ground/rocks, and saltwater)

1. Gold Mode
2. Park Mode 2 and Field Mode 2
3. Park Mode 1 and Field Mode 1
4. Beach Mode 1
5. Beach Mode 2

Least Hot (less sensitive to small targets, hot ground/rocks, and saltwater)

Now seriously, just how different is that from a range of selectable single frequency, high to low?

What I’d to see is all this incorporated into a simple variable control. Instead of discrete modes and complicated explanations, give me a single control. Setting higher sets for progressively higher frequency mixes, setting lower goes to lower mixes. You simply set the control as high as your ground allows, which is basically they way we pick between the modes above anyway. Simplistically using my little chart above the control would have 5 - 7 levels.

I think attaching names to the modes misleads people anyway. There is no reason not to use Gold Mode to Relic hunt, or Beach Mode to hunt silver coins, but people get locked into thinking those mode names mean way more than they do, and I think some peoples detecting suffers as a result. Get rid of the names, and it frees your mind to a more open approach as to which modes work best under and given circumstance.

Steve

Great post. Fortunately for me the Deus II is both! SMF, or as they label it, FMF and it still has single selectable frequency mode.

[Steve Herschbach]

On 3/29/2022 at 1:53 AM, JOE WATSON said:

Hi just a thought, but do you think metal detectors could ever utilise chirp technology as in the new echo sounders?

So far GPR systems so far lack the resolution needed to hunt small targets, require lots of power, and are very expensive. But as batteries and processing, improve it opens possibilities there for the future. Since target density is involved, something like separating gold rings from aluminum would be a good place for any developer to focus on.

Back to the subject at hand, from Minelab:

“Find Every Target Type & Size with Every Sweep Generally, high transmit frequencies are more sensitive to small targets and low transmit frequencies give more depth on large deep targets. FBS simultaneously transmits and analyses a full band of multiple frequencies from 1.5 kHz to 100 kHz and is therefore sensitive to both very small and large deep targets at the same time. This means you only need to cover the ground once and can be confident you’re not leaving ANY valuable treasure behind.”

Wow, 1.5 kHz to 100 kHz, all at once! Not leaving anything behind!! Must not need anything else, right? I don’t know how many times I had to explain that the 100 kHz claim did not mean this detector was any good as a dedicated gold nugget detector, which it should be, if that’s all true. But fact is the FBS machines, while very good on most targets, are only just good on low conductors, and no match for dedicated nugget detectors, even ones running at moderate frequencies like 19 kHz. And certainly no match for Equinox on tiny gold, even though it maxes out at 40 kHz.

[Skookum]

I keep thinking the next level of MF is going to come from increases in processing power that allow application of AI algorithms within the machine.  What if a processor analyzes how a target responds to multiple frequencies or technologies at the same time and then it extrapolates that information to identify other targets at your location?  No doubt this already happens in R&D labs, but could there be fresh gains having the process occur in real-time, on the ground?  Perhaps, its another potentially cool application of artificial intelligence.

[mh9162013]

4 hours ago, Skookum said:

I keep thinking the next level of MF is going to come from increases in processing power that allow application of AI algorithms within the machine.  What if a processor analyzes how a target responds to multiple frequencies or technologies at the same time and then it extrapolates that information to identify other targets at your location?  No doubt this already happens in R&D labs, but could there be fresh gains having the process occur in real-time, on the ground?  Perhaps, its another potentially cool application of artificial intelligence.

I've had similar thoughts.

For example why can't the metal detector be programmed for the user to tell the machine what was found? After the user does this enough, the detector will have "learned" how to identfy targets in the ground based on what it was detecting and what was ultimately dug up.

It can be limited to just American coins, for instance. Let's say a user (I'm gonna used Minelab's MIQ scale here) finds a target that rings up as a 25 and is 4 inches down. The user thinks it might be a copper penny. The user carefully digs it and notes how deep it was, its orientation in the soil (if that's possible to discern...it sometimes is when the coin is neatly stuck to the side or bottom of the plug) and what the object actually was.

So in my hypothetical, it turns out to be a 2004 Zincoln, not a copper penny. The detector will make a "mental note" of the find and several variables that corresponded with it (including the GPS coords of the find). If a user is willing to take the time to teach its detector to do this, the machine would be pretty darn good after just a dozen or so digs in the same general location.

The problem I see with this approach is threefold. First, it requires a lot of time and patience. Second, there's the problem of "garbage in, garbage out." If the user is wrong in the data it gives the machine (gives the wrong depth of the coin, for instance, when the user swings the coil higher off the ground than is recommended), then the detector can't properly learn. Third, you're basically teaching a machine to do what a neurotypical human being who's really into metal detecting is doing anyways. Basically, this hypothetical machine isn't just a metal detector, but a notetaker and pattern finder. But humans are already pretty good at seeing patterns. They're just lousy with memory and tend to be lazy in their habits. So it's too easy for them to incorrectly remember: the find, the settings used and the ground conditions at that time.

A simpler process would just be to use the machine, tell the machine to note a given signal, then after digging it up, simply telling the machine if it was a "good" or "bad" target. This boolean "process" could also teach the detector about what's a potential good find or not. For coin shooters (gven how standardized the desired targets tend to be), this could be pretty effective. I think in its ideal form, it would still give a tone and VDI like any other detector, but it would have a "confidence bar" to go along with it. It's this confidence bar that would be constantly altered as the machine learns from the user.

[☠ Cipher]

15 hours ago, mh9162013 said:

I've had similar thoughts.

For example why can't the metal detector be programmed for the user to tell the machine what was found? After the user does this enough, the detector will have "learned" how to identfy targets in the ground based on what it was detecting and what was ultimately dug up.

It can be limited to just American coins, for instance. Let's say a user (I'm gonna used Minelab's MIQ scale here) finds a target that rings up as a 25 and is 4 inches down. The user thinks it might be a copper penny. The user carefully digs it and notes how deep it was, its orientation in the soil (if that's possible to discern...it sometimes is when the coin is neatly stuck to the side or bottom of the plug) and what the object actually was.

So in my hypothetical, it turns out to be a 2004 Zincoln, not a copper penny. The detector will make a "mental note" of the find and several variables that corresponded with it (including the GPS coords of the find). If a user is willing to take the time to teach its detector to do this, the machine would be pretty darn good after just a dozen or so digs in the same general location.

The problem I see with this approach is threefold. First, it requires a lot of time and patience. Second, there's the problem of "garbage in, garbage out." If the user is wrong in the data it gives the machine (gives the wrong depth of the coin, for instance, when the user swings the coil higher off the ground than is recommended), then the detector can't properly learn. Third, you're basically teaching a machine to do what a neurotypical human being who's really into metal detecting is doing anyways. Basically, this hypothetical machine isn't just a metal detector, but a notetaker and pattern finder. But humans are already pretty good at seeing patterns. They're just lousy with memory and tend to be lazy in their habits. So it's too easy for them to incorrectly remember: the find, the settings used and the ground conditions at that time.

A simpler process would just be to use the machine, tell the machine to note a given signal, then after digging it up, simply telling the machine if it was a "good" or "bad" target. This boolean "process" could also teach the detector about what's a potential good find or not. For coin shooters (gven how standardized the desired targets tend to be), this could be pretty effective. I think in its ideal form, it would still give a tone and VDI like any other detector, but it would have a "confidence bar" to go along with it. It's this confidence bar that would be constantly altered as the machine learns from the user.

It is fun to wonder what would happen if you could incorporate artificial intelligence machine learning into these devices. I would imagine such things aren’t the strong suit of metal detector engineers/programmers and they might have to utilize outside consultation, but some level of this could certainly be done, and enabled to be switched on or off.