OK guys -- as a new CTX 3030 user (about 6 weeks now), I'm really enjoying the unit. The transition from an Explorer has gone better/more smoothly than I anticipated, so far.

One thing I'd like to understand in a deeper way, though, are the "separation" settings -- ferrous coin, ground coin, low trash, high trash. I've seen these referred to (by Keith, among others) as "filters," but that is a term I'm only peripherally aware of, and don't really understand, from a "circuit" perspective. I've seen reference to "two-filter machines" and "four-filter machines," and if I'm not mistaken, this hearkens back to the "analog" era. But I don't know exactly what it refers to.

IN ANY CASE, I'd like to understand these four settings better, what each of them is theoretically "supposed" to do (not the simple, user-manual version), and how to really be able to use these settings to my advantage, depending upon site type, desired targets, etc.

Also, what advantages/disadvantages/characteristics of these settings have you all noticed, in your specific real-world conditions/sites/scenarios?

So far, I've run in ferrous coin exclusively, and if I'm not mistaken, it is this separation setting that allows multiple targets (ferrous and non-ferrous) to be displayed on the screen, with target trace. But I've read some old posts here, where Keith talked about "ground coin" being preferred at times, and perhaps being "deeper," and "less falsy" on iron, than ferrous coin? But why would that be?

Anything anyone would like to share, I'd be interested in pondering...

Thanks,

Steve



Edited 1 time(s). Last edit at 12/26/2017 08:52AM by steveg.

I adjust those settings bases on what the site conditions call for. Each of them are optimized for different conditions etc. Low trash for example works best in wide open areas where the trash density is less and targets are more sparse. It doesn't work wel k in high trash because it locks on to the strongest target and blanks out the rest causing you to miss other good targets around it. High trash works good in lots of soils and allows you to use a faster sweep speed which I lime the best. I hunt in this mode 90% of the time because my soil having changing conditions constantly due to patches of alkali handles it the best.
A lot of people use ferrous coin because it allows you to get separate signals for every different target the coil passes over which works good in areas of low mineralization and lots of iron targets.
Ground coin is the other mode I use the most because it gives me more consistant ID values in my terrible soil and I use this in wide open areas with less trash etc. I base what setting to use on what my ground setting say after I noise cancel. I usually run auto+3 and if I have a high sensitivity reading say 22 or above then I select ferrous coin, if it drops to 15 to 21 or so I will switch to high trash, anything lower than 15 I rub ground coin if less trash etc. You should select the setting according to your ground reading. Hope this helps and good luck!

Weglund --

Yep, it helps, thanks! So, you let your ground conditions (i.e. "preferred sensitivity" readout on the machine) be your guide. Interesting.

I guess one question is -- if the "low trash" setting supposedly "locks onto" the "strongest "signal under the coil (accepted OR rejected), whereas "high trash" locks on the strongest ACCEPTED signal under the coil, it would seem to only make sense to hunt WITH discrimination when using the "high trash" setting, right? Otherwise, since ALL targets are obviously "accepted" with open screen, it would seem that running "open screen" with the "high trash" setting would negate whatever advantage the "high trash" setting is supposed to have?

Meanwhile, what if you are running the "high trash" setting, and you run the coil over a nail, but you are discriminating iron IDs. It would seem to me that when you pass over the nail, whatever "high-tone chirps" the nail might be making would then get "locked onto" in the high-trash setting, no? (In other words, with the machine "seeing" the "falses" off the tip of the nail which are giving a reading in the "accepted" range, would the high-trash setting cause the unit to want to try to "lock onto" those falses -- due to the nature of the setting and what it's trying to do -- thus making it MORE difficult for the user to discern the "nail falsing" scenario from the "coin closely adjacent to a nail" scenario?)

These are the types of things I want to understand better. In general, I'm a fan of letting my own ears and brain do some of the work; when you have all these "artificial alterations" to the audio going on through algorithms, it seems there can be as much chance of CONFUSING the situation for the detectorist, versus HELPING...

Steve

steveg Wrote:
-------------------------------------------------------
it would seem to only make sense to
> hunt WITH discrimination when using the "high tras
> h" setting, right? Otherwise, since ALL targets a
> re obviously "accepted" with open screen, it would
> seem that running "open screen" with the "high tra
> sh" setting would negate whatever advantage the "h
> igh trash" setting is supposed to have?

no -

set your tonebreak to knock out the nails so you can keep the machine wide open/accepting all metals with no disc

Steveg,

One key setting is move tone break to line 32 ferrous using combined mode.
I have used CTX in horrid sites doing this.
Get ready though for a barrage of audio.
Train your ears.
Lots of resweeps too.
Lower conductors will escape you in these sites running auto sens even auto plus 3 sens setting.
Go manual.
How high?
You'll have to try different levels, start at 22, then work your way up.
Really needs to be higher than 22.
But hold on you are in for a ride!!!

Can produce though!!!

You rely on screen you can lose too.
Depend on the audio!!!



Edited 1 time(s). Last edit at 12/27/2017 02:29AM by Sod-buster.

MichiganRelicHunter --

I get what you are saying. Running open screen/combined makes a lot of sense to me, and I run that way about half the time (though I DO wish Minelab had given us a "multitone" option above the FE line, instead of just 4 tone bins!)

Anyway, what I was saying was more "hypothetical." In other words, if the big "benefit" of running a "high trash" separation setting is supposedly that it will "lock onto" the strongest ACCEPTED target, then that implies to me that it's a setting you use IF you are hunting with discrimination. I interpret it to mean that usually the CTX will lock onto the "strongest signal under the coil," period...BUT, in high trash separation setting, it INSTEAD locks onto the strongest ACCEPTED signal. So in a hypothetical scenario where you have, say, a shallow pull tab giving a strong signal, and then a deep coin, that is giving a weaker signal, both under the coil at the same time, theoretically in "low trash" separation mode it would lock onto the pull tab (strongest signal), and may not report the coin at all. In high trash, and open screen, it would also apparently lock onto the pull tab (strongest "accepted" signal), and again might not report the coin. BUT -- if you "disc out" the pull tab, THEN (theoretically) it would lock onto the COIN -- because at that point, with the pull tab discriminated out, the coin would thus become the strongest ACCEPTED signal. That's how I understand it "in theory," and thus "in theory," you could possibly benefit from running "high trash" separation IF you are also running with a discrimination pattern. NOW -- while this may be the engineer's intent, and may work "in the lab," I have serious doubts as to whether this is actually how it works "in the field." Thus, my question. As originally stated, I'm just trying to wrap my brain around what these "separation settings" are REALLY doing, and when/why you might choose one over the other.

That notwithstanding, I DO understand and appreciate your point about preferentially running open screen/combined tones, MRH. Thanks!

Steve

Sod-buster --

Clearly that's a much more "advanced" way to run the CTX -- trying to achieve maximum ability to unmask in iron (setting the tone break down to 32). Once I get more comfortable with the unit, and in an appropriately iron-polluted site, I will experiment with that, for sure. For now though, I dig enough nails when chasing after deep coins as is, with my tone break set at 20! LOL!

Same thing for manual sensitivity. I used to run my Explorer wide-open, hot-as-it-would-go most of the time, at 32 sensitivity. I got used to the noise, and unless it was EXTREMELY noisy/unstable, I'd keep it at maxed-out 32. However, so far I've largely run in Auto +3, but will often check a real "iffy" target in manual, to see if that cleans it up some, and gives me a stronger/better hit on the target. I anticipate that -- again, once I get more used to the unit -- I will be running manual sensitivity, as hot as I can run it...

Finally, I hear you on the audio. And I agree. Which is why I'm torn between running 50-tone conductive, and combined. It seems like most users of the CTX run combined tones, but I'm having a hard time convincing myself I'm not losing the "clues" that there are in the audio, by "dumbing down" the tones to only four bins...

I TOTALLY see the value of open screen/no disc, and setting a ferrous tone break (i.e. bottom bin). I just can't understand why it wouldn't be MUCH better to have the ability to run 50-tone conductive above the FE break line, as opposed to being limited to only 4. WHY IN THE WORLD Minelab chose to give us only four tones above the FE bin, I cannot understand. And what I cannot understand EVEN MORE, is why they don't put that in there as a software update!! Fine, they didn't do it at first. WHY NOT GIVE US THAT, IN AN UPDATE, MINELAB???

Anyone have any thoughts on multitone vs. combined?

THANKS, Sod-buster. Good info there...

Steve

Sod-buster Wrote:
-------------------------------------------------------
> Steveg,
>
> One key setting is move tone break to line 32 ferr
> ous using combined mode.
> I have used CTX in horrid sites doing this.
> Get ready though for a barrage of audio.
> Train your ears.
> Lots of resweeps too.
> Lower conductors will escape you in these sites ru
> nning auto sens even auto plus 3 sens setting.
> Go manual.
> How high?
> You'll have to try different levels, start at 22,
> then work your way up.
> Really needs to be higher than 22.
> But hold on you are in for a ride!!!
>
> Can produce though!!!
>
> You rely on screen you can lose too.
> Depend on the audio!!!

Steve,
You are quite welcome.

Actually I think Minelab should have offered possibility of 50 bins in combined mode if a person chooses to use.
Rather than the 50 multi tone using combined mode.
This way a user can still have their nickel window or whatever window they choose to meld with a tone of their choosing.

Much like what Rutus Alter 71 offers.

Minelab should offer a no tone option 0 hz option in the iron bin as well.

CTX, is by far a much better nickle killer in modern trash vs older fbs units. (Locating by use of ears)
Thanks to combined mode with its tone bin offering.



Edited 3 time(s). Last edit at 12/27/2017 05:33AM by Sod-buster.

I hear you, Sod-buster, and I agree. Offering more "customizable" bins than the 4 would also make sense. 50 would be the maximum, one for each conductive ID number, but if I wanted to make 10 bins, or 15, or 20, or whatever, and set up "windows," like you said, that would be great as well. Bottom line is, 4 is NOT enough!

The 0 Hz option for the iron bin makes sense also, IF running "disc" hinders depth at all. I have read here on the forum some speculation that running disc. on the CTX does not hinder depth, so if it really doesn't, I guess running a disc. pattern to "silence" iron is the same as "accepting" iron, but with a 0 Hz tone. But -- if disc. causes any hindrance of depth, then I agree, a 0 Hz option makes sense.

Steve

Hi Steve,

I think you have described the High and Low Trash separation settings correctly. My guesswork follows: the assumption is that in non-ferrous trashy areas the user will opt for more discrimination to eliminate the trash. But the more disc used the less sensitivity overall and weaker non-ferrous objects (coins) would may not register at all when using aggressive discrimination, especially with ferrous targets nearby. So, in High Trash some signal processing is used to tighten the disc pattern to minimize rejecting open bins near rejected bins, giving the open bins a better chance to register.

If you have an open screen (no or very little disc), then High Trash has no bleed-over to sharpen and may have little impact overall, as the sensitivity is at a maximum anyway when disc isn't used. I think the design intent was to partially mitigate the massive sensitivity loss when using a lot of discrimination, such as in trashy parks. The user can sweep quickly in High Trash and still get the non-ferrous target in an open bin next to a disc bin (at least in theory - I haven't tested it). However, in theory the disc'd bin will be more likely to sound off (i.e. be heard) in High Trash than Low Trash (Low Trash is more effective in keeping bins separated).

The Low Trash seems a bit puzzling, from a designer perspective. The goal was to have the machine respond to the strongest signal - which is simply a normal state of any detector, bigger amplitude signals are easier to obtain. Since Low Trash tends to blend adjacent non-ferrous (so ML says) it appears to be a smoothing processing of the signal, decreasing target separation ability. I suppose the design team thought eliminating weak non-ferrous signals and providing a smoother reaction in each sweep might be less stress on the user? The benefit of a longer averaged signal processing (or a broader response) might be an improved TID, keeping the TID centered on the strongest signal (removing the weaker signals that skew the resultant). Again, happier user (though they miss the deep weak signal from perhaps a rare non-ferrous coin, but it's a trade-off. The novice is not being flummoxed by constant signal pops from weak signals from deep gold coins). I kid - sort of.

It appears both Low and High Trash use signal processing to keep the open bins from being contaminated by adjacent rejected bin signals. Both allow the user to add discrimination and still get the open bin response to some degree, but Low Trash appears to be superior in that regard. That means Low Trash has a tighter rejection scheme, the disc'd bins don't bleed into the open bins as much. However, I think using Low Trash in areas with multiple adjacent non-ferrous or ferrous targets would be unwise (which is why ML recommends against using Low Trash in high trash areas). It appears that using High Trash in low trashy ground works just as well, making Low Trash somewhat unnecessary (and unwanted considering Low Trash's unwanted masking of adjacent non-ferrous ground signals). Since Low Trash allows for more defined disc'd bins with less contamination into adjacent open bins, this would likely be most useful if using disc to eliminate unwanted ferrous responses in bins next to wanted non-ferrous targets. It wouldn't help much if you are trying to isolate similar non-ferrous targets, say pull-tabs from nickels, as the signal response is blended and is self-defeating if the surface pull-tab is masking the response from the slightly deeper coin.

The only recommended target separation for high mineralized soil is Ground-Coin, otherwise, it looks like a decision between Ferrous-Coin and High Trash in low-med mineralized soils. High Trash probably has the edge. Ferrous-Coin is not as effective as mineralization increases (works best in clean soil), wouldn't be the choice for relic hunting (if ferrous targets are desired), may make the audio null over non-ferrous targets, and wouldn't handle non-ferrous trashy areas as well.

I use High Trash quite often and Ferrous-Coin at times. I've played with all the schemes and settings - I haven't come away with any real preference. If ML gave more information, then I might be more inclined toward one or the other. But, as it stands it is mostly just guesswork, I just pick whichever seems to work the best at the moment for the current conditions.

Johnnyanglo --

Great info, thanks. I went back a couple of days ago and read a bunch of old threads on CTX on this forum -- and they were filled with a bunch of info that you posted, which was very helpful.

One thing I'm failing to understand, in your description above...

You are saying that Low Trash uses a longer averaged signal processing -- thus giving a broader response to non-ferrous targets. This longer processing "smooths" the signal and gives more stable target ID, BUT it also thus decreases separation ability due to the tendency for "blending" of target signals (if there are multiple adjacent non-ferrous targets). This makes total sense, in and of itself. (I wonder if the longer processing would also mean more depth, in open ground?)

BUT, I can't then figure out how, as you said, Low Trash ALSO has a "tighter rejection scheme," where disc'd bins don't "bleed" into the open bins as much. This seems counter-intuitive to me. If low trash uses longer averaged signal processing (blending), then how can it have a "tighter rejection scheme," allowing for "more defined disc'd bins" with "less contamination/bleeding into adjacent open bins?" To me, it sounds like longer averaged signal processing would mean MORE "contamination" into adjacent bins, and so I can't figure out how Low Trash would be capable of doing both of these things at once (longer averaged signal processing, but also tighter rejection).

By the way, please don't take that as me "challenging" you, it's just me trying to get the details and my understanding of the details clear/solid in my mind...

Steve

steveg Wrote:
-------------------------------------------------------
> Johnnyanglo --
>
> Great info, thanks. I went back a couple of days
> ago and read a bunch of old threads on CTX on this
> forum -- and they were filled with a bunch of info
> that you posted, which was very helpful.
>
> One thing I'm failing to understand, in your descr
> iption above...
>
> You are saying that Low Trash uses a longer averag
> ed signal processing -- thus giving a broader resp
> onse to non-ferrous targets. This longer processi
> ng "smooths" the signal and gives more stable targ
> et ID, BUT it also thus decreases separation abili
> ty due to the tendency for "blending" of target si
> gnals (if there are multiple adjacent non-ferrous
> targets). This makes total sense, in and of itsel
> f. (I wonder if the longer processing would also
> mean more depth, in open ground?)
>
> BUT, I can't then figure out how, as you said, Low
> Trash ALSO has a "tighter rejection scheme," where
> disc'd bins don't "bleed" into the open bins as mu
> ch. This seems counter-intuitive to me. If low t
> rash uses longer averaged signal processing (blend
> ing), then how can it have a "tighter rejection sc
> heme," allowing for "more defined disc'd bins" wit
> h "less contamination/bleeding into adjacent open
> bins?" To me, it sounds like longer averaged sign
> al processing would mean MORE "contamination" into
> adjacent bins, and so I can't figure out how Low T
> rash would be capable of doing both of these thing
> s at once (longer averaged signal processing, b
> ut also tighter rejection).
>
> By the way, please don't take that as me "challeng
> ing" you, it's just me trying to get the details a
> nd my understanding of the details clear/solid in
> my mind...
>
> Steve


Steve,
I can understand this being more feasible vs the other way around.
I sure wouldn't expect faster processing to provide more accurate data.

Just like a meteorologist.
You can relate to this, Johnny too I think.

If you were given more time to study info presented, couldn't you give even a more precise forecast as far as the US landscape?
Granted it only makes sense to go only so far (depending on situation) with finite predictions surface wise- meteorlogically speaking.

And more time doesn't guarantee higher success with forecasting weather, nor does it when it comes to a metal detector.
But should be more accurate on average more times.



Edited 5 time(s). Last edit at 12/27/2017 07:26PM by Sod-buster.

Yes, Sod-buster, I agree. More time to analyze SHOULD result in more accuracy, on average. And, I'd think, more depth, in the case of "low trash" separation. If it is spending more time analyzing, I'd think it might be able to ID at depth a bit better, in a "clean" site...

Steve

Realize that this is conjecture, but based on what little info ML provides it does appear to me that the signal in Low Trash is processed so that the highest amplitude non-ferrous signal passes the gate within some increased length of time. That is, if the goal is to avoid multiple weaker non-ferrous hits. However, the detector I believe doesn't 'see' the ground longer, which is determined by frequency, which doesn't change, but rather uses a filter to remove all but the strongest signal within some set time period. In high metallic trashy areas this scheme would cause many weaker targets to be eliminated and probably most surface trash to (strongest signal) to respond, so ML doesn't recommend Low Trash for those areas.

In meteorology you might think of it as horizontal resolution. An infrared satellite image with 4km resolution will 'see' smaller features due to the smaller pixel resolution but move to a 8km full-disk image and the small cloud elements are gone - blended away by averaging across the pixel temperatures. Consider a detector that is 'sparky' like the F75, chattering away on tiny pieces of non-ferrous trash, it has that very high resolution (say 0.5km). ML doesn't have that high resolution to start with, but I'd figure the goal of Low Trash was to avoid the pips and pops to eliminate unwanted noise for the novice. They took essentially a say 2km potential resolution signal and processed it to 4km resolution - eliminating the fine details.

As to the target display bins, the lower resolution signal post-processing is now just one signal with competing signals removed. With variations eliminated by the filter the signal is actually small and compact and now fits nicely into a single display bin. The other Co-Fe signals that were received are not used, or blended away into a single measured value that fits nicely into a bin.

It would be as if the detector received a signal from the ground like this over some set time frame: 3.4532, 6.230, 1.22, 8.124 and processing it to be 8.0 by filtering the smaller amplitude signals out and cleaning it up by removing the fractional parts - there is little left to bleed into other bins.

The problem I think with Low Trash is that the 1.22 weaker signal was a deep non-ferrous target and the 8.124 was surface - generally the deeper target is likely to be more valuable than a recent drop (which has a high probability to be non-ferrous trash). I don't see the utility in using Low Trash for most coin hunting situations. The TID is more stable (post-processed signal is clear) but only at the cost of losing those other signals that make the TID bounce about (again, that's information lost).

High Trash, on the other hand, does appear at least notionally to have a better purpose. Eliminating rejected bins (disc'd bins) from responding even when the rejected signal is stronger is useful in reducing unwanted noise. Really, that's what any discriminating detector is supposed to do - the user inputs disc'd bins and they don't respond. Since the high-pass filter appears to be limited in High Trash, there will some pops and pips in the open Co-Fe bins too derived from an unwanted target, but ML probably filters a bit of the uncorrelated weak signals out, but not as aggressively as in Low Trash.

So, High Trash might be best described as the more normal operating profile of a discriminating detector. Since some processing has happened in High Trash too, some people may want to hear more of those transient pops and pips in their open bins as they could be deep weak signals that are valuable. Therefore, ML recommends using Recovery Fast On as it limits even more the averaging and filtering to a greater degree, perhaps by-passing the filter entirely, allowing the remaining post-processed signal to fall wherever it may. The problem then is a jumpy TID and more audio chatter - but you get more information. It's a trade-off. High Trash might be better for more experienced users and Low Trash for the novice.

--Johnnyanglo

'The Low & High Trash would be filter sets as you have on the other FBS detectors, E-Trac, Safari.
The Ferrous Coin & Ground Coin are specific to CTX and are 'driven' by the users needs to help properly identify coins.
In the case of Ferrous Coin, and with Target Trace on, it helps show non-ferrous coins with 'reds, orange & brown colors' while also showing the nail it's next to as 'blues, torquoise colors'
If a ferrous item is rejected outright it displays a red triangle.
With Ground Coin, and it should be utilised in mineralised soils, if using a big coil - 11" to 17", the overall footprint is 97% ground effect and 3% coin target so the ground effect can 'skew' TID's
So Ground Coin can hold the ground effect at bay and help the TID's of coins to improve and rather than showing heavily influenced FE TID's, it can lower them to within normal limits.

From my time spent testing all the Target Separation options, it helped to pick just one or two during a hunt because they are 'ground specific'
For example, if in an iron littered area Ferrous Coin could be used but then what can happen is the TID's can be off, at times by half the values again!
For example, a non ferrous target, coin, button or the like could display FE18 CO 38 (just an example) and if the Ground Coin is used to momentarily check, it can help display the correct FE/CO coordinates, and couid then display FE12 CO 36

That is the single biggest drawback about using FC: it will incorrectly display TID's a lot of the time, so I wouldn't use it all of the time.
I'd opt for one of the other options, Low Trash for example.

Johnnyanglo --

THANKS for the further info. It does make more sense now, what you are "conjecturing," about the low-trash setting. While it seemed "counter-intuitive" at first, I see what you are saying now. Makes sense. (The "satellite imagery" example helps, LOL -- you have some knowledge there obviously, as you are spot on with how the imagery works and you mentioned the term "full disc" which most would not know what you are saying!)

Thanks for the in-depth clarification. For a person unfamiliar with detector design/circuitry, could you possibly explain what you mean by "high-pass filter?" I kind of understand what "filtering" is -- "cleaning up" of signals in certain ways, removing parts of the signal, amplifying other parts, etc. -- and I presume what was done using electronic components in the "analog" days is now done in "software," yes? But even with this basic, general concept that a complex return signal is "altered" in ways that designers think will benefit the user, and that is called "filtering," what exactly is meant by "high-pass filter," or a "two-filter" detector, or a "four-filter" detector?

Des --

Thanks for the additional info there regarding the different separation options. I learned, on my Explorer, in my dirt, how the ground would "alter" the ID of a deep target...I didn't know much about the why or how, but instead learned "by experience" that ground mineral would "skew" the ID of a deep (weak) target -- very much like you were describing above in your paragraph discussing "ground coin." I took what I observed, over time, as to how the machine behaved, and then "worked backwards" to kind of figure out what must be happening...what the machine must be doing "behind the scenes."

You explained that "ground coin" is trying to remove that "skew" of the ID that is caused by mineralized ground (which I observed with the Explorer), and to return things to "proper" ID values. But -- and this is going to sound weird -- this is "throwing me off," moving from the Explorer to the CTX. Obviously, for someone who never used FBS, having settings available to help maintain "proper" or "normal" ID values (relative to air-test values) is in most circumstances a good thing, in that it makes using the detector easier for those who didn't have prior FBS experience, greatly reducing the "learning curve." BUT -- for someone who got so used to how ID values get "corrupted" by mineralization, etc., especially on deep targets, and learned to let their BRAIN do the "filtering" so to speak, I am finding it a bit of an adjustment.

As such, you now see the original reason for my post, asking about these "separation settings." I sense that the differences I'm seeing between the Explorer and CTX have a lot to do with "signal processing" happening behind the scenes. Things that I used to have to do "manually" with my brain...things which I had to learn through experience using the Explorer over several years, are now being done "in the software," prior to outputting the info to the user. And so, I find myself right now trying to understand just how the things I learned over time, nuances and subtleties, with the Explorer, are now "dealt with," behind the scenes, in the CTX. And Minelab is not exactly "forthcoming" with that info! LOL!

Some would say I am thinking too much, making things "too difficult;" that I should FORGET trying to figure out the "how" going on inside the control box, and just "hunt." I understand that point of view. They'd say "stop thinking about it and just trust the machine to do what it was designed to do." But, the way my brain and personality work, I really prefer to have some scientific UNDERSTANDING, before I just "trust." I hate "black boxes." When you blindly trust a "black box," you may be led in the proper direction much of the time (if the designer of the "automated system" was skilled/talented). HOWEVER, it's also true that you will ALSO at times be led astray, as no "black box" system that tries to do the thinking FOR you is flawless. And thus, UNDERSTANDING what's in the "black box" -- what it is trying to do, and how (so that it becomes less of a "black box,") means you will then ALSO acquire an understanding of the circumstances and scenarios where what the black box is trying to do in an automated sense will FAIL, and lead you in the wrong direction. And if you understand these "failure modes," you can then anticipate the scenarios where it is appropriate to "put less weight" on the black box's output, and instead inject your own intelligence into the situation to draw a more accurate conclusion.

The frustrating thing, of course, is that Minelab likes to keep their black box "black," which hinders thinkers from trying to learn how to apply their "thinking" to their detecting in a appropriate way, in different scenarios.

(Sorry for my rambling).

Steve



Edited 2 time(s). Last edit at 12/29/2017 12:39AM by steveg.

This gent does good videos.

[m.youtube.com]

[m.youtube.com]

[m.youtube.com]

This gent does good videos.

[m.youtube.com]

This gent does good videos.

[m.youtube.com]



Edited 3 time(s). Last edit at 12/29/2017 02:05AM by Sod-buster.

Thanks, Sod-buster.

Good videos (thanks, Albert!)

Steve

"Steve

I'm buying you a 'Miller Lite'....6pk...

You need to take a breath of else your head will explode!

We'll never understand the very advanced DSP involved...
Only way is, if you're hired as a physicist at Minelab and then you'd be tied to a very severe NDA

I worked there and helped develop the CTX but they never told me a single code word!!!

Des D

Des --

LOL! Thanks, buddy! I know. That's probably what I need. A six pack...

I realize I'm never going to know the "code," per se. And that's OK...I'm not trying to learn the code; what I'd LIKE to know, though, is WHAT was the INTENT of the guy who WROTE the code, when he programmed the "ground coin" setting, for instance. What real-world soil problem, relatively to detecting buried non-ferrous targets, did he know of, that needed to be dealt with when developing the CTX. What does ground coin try to do, and what is the result. What would happen if I'm in bad soil, and NOT running ground coin separation, and what changes, in terms of output, when I DO run it.

That's the kind of thing I'm talking about. Knowing those things about EACH setting on the machine, how the engineers TRIED to handle it, with various settings, and what changes in the machine's operation when you utilize the settings (and the "pros" and "cons" of the settings).

I will probably figure SOME of this out, eventually, just like I did with the Explorer. But Minelab really could EDUCATE, if they wanted to, about how an advanced user could really get the most out of the machine by understanding more of what's going on.

Steve