Reply
 
Thread Tools Display Modes
 
Old 12-20-2014, 12:33 PM   #41
Guru
 
JDCAVE's Avatar
 
City: Lions Bay, BC
Country: Canada
Vessel Name: Phoenix Hunter
Vessel Model: Kadey Krogen 42 (1985)
Join Date: Apr 2011
Posts: 1,602
Charts on tablet, pc, laptop or LCD off chartplotter

Sailor of fortune said "Vector are more accurate". Why would you believe that to be true? 3rd party suppliers of Vector charts for Canadian waters would have to source their information from the Canadian Hydrographic Service, which produced the information on the raster charts to begin with. Just because there are zooming capabilities on the Vector charts, or because they have that pseudo 3d look, doesn't make them inherently more accurate. It depends on the original sounding data.

I chatted with one of the Individuals from the Canadian Hydrographic Service at the Vancouver boat show last January, and he mentioned one of the biggest issues with Cdn charts lies not with the soundings, but with where these are placed on the charts. They still rely on some soundings that were taken pre WWII, and the spatial coordinates may be quite inaccurate. Such inaccuracies "ground truth" the best Nobeltec Vector charts.

Vector charts do not provide the same level of land detail (height contours) etc which I appreciate when I'm running about. I have both vector and NOAA charts for Puget Sound on my Coastal Explorer program and find I go back and forth between the two when I'm traveling south of the fence.

Jim
Sent from my iPad using Trawler Forum
__________________
Advertisement

JDCAVE is offline   Reply With Quote
Old 12-20-2014, 01:43 PM   #42
Senior Member
 
Richard W's Avatar
 
City: Florida USA and Ontario Canada
Country: The 3rd Rock from the Sun
Vessel Name: anytime
Vessel Model: 2007 Chaparral 270 Signature LOA 29'
Join Date: Jan 2014
Posts: 303
Quote:
Originally Posted by JDCAVE View Post
I have both vector and NOAA charts for Puget Sound on my Coastal Explorer program and find I go back and forth between the two when I'm traveling south of the fence.
This exemplifies the confusion and explains some misunderstandings ...

The NOAA charts come in two versions ... you can have NOAA raster and NOAA vector chart. They both look similar and are equally accurate as are based on the same data.

One cannot compare chart technology (raster/vector) with chart source/content as provided by creator/publisher. These are two different qualities/criteria ... one is the technology or method used to render a chart, another is content, and specifically quality and accuracy of the data used to create a chart.
__________________

__________________
Best ... Richard
Richard W is offline   Reply With Quote
Old 12-20-2014, 01:45 PM   #43
Scraping Paint
 
City: -
Country: -
Join Date: Oct 2007
Posts: 13,748
Quote:
Originally Posted by ranger42c View Post
Not sure I can understand how a vector chart can be inherently more accurate than a raster chart.
They aren't. As I understand it, vector charts are created by digitizing data from paper charts. As such, there is a risk of inaccuracy if mistakes are made in the digitizing process. That's why if one prefers the flexibility and user advantages offered by vector charts it's important to use vector charts created by a reputable and proven company.

This is why we have selected C-Map charts for the plotters on all our boats. The plotter manufacturers we use are Furuno, Standard Horizon, and Echotec (no longer in business).

The reasons we prefer vector charts have to do with their flexibility of use as I've described earlier in this thread. To us, vector charts have rendered raster charts into the equivelent of stone tablets in terms of user-friendliness, variety of information, and ease of use.

The vast superiority of properly designed vector charts like C-Map is all about what they make possible operationally, not their accuracy which, if they are digitized correctly, is identical to raster or paper charts. They can be as accurate as raster charts, but I see no way they can be more accurate.
Marin is offline   Reply With Quote
Old 12-20-2014, 02:05 PM   #44
Guru


 
City: Full-time onboard
Country: USA
Vessel Model: Trawler
Join Date: Oct 2007
Posts: 937
The confusion is about accuracy versus precision. Both raster and vector have identical accuracy. They both come from the exact same source. But since the vector charts are comprised of data points and objects, they can be zoomed to a much greater level and still look clean and crisp - that's providing more precision.

The big danger with vector charts is that the added precision provides a false sense of accuracy - you see a much larger channel drawn so you feel like being in the center of it means you're really in the center. But as the vector data is zoomed past its accuracy limits, you're only fooling yourself about where the channel actually is. The visual center on the zoomed in vector chart can be (and often is) way off center.

This is one of the advantages of raster charts. As you zoom in too far, the pixel replication makes the image look worse and you don't get that false sense of knowing where things actually are located. They get blurry and give you extra information about accuracy that is quite valuable.

Vector charts, however, are much smaller and allow for object queries. They also allow the text to be scaled to your preferences (meters, feet, fathoms) and the underlying data can be used for collision avoidance - as you're approaching shallow water, a vector chart rendering product could warn you (some do) because they actually have depth regions and can detect as you're approaching a shallow one. The recent Volvo race grounding could have easily been avoided with a variety of products that use vector data to warn in advance of approaching shallow water. There's generally no depth data behind the pixels of a raster chart. They're just pixels.

Raster charts, however, are created with hand placement by cartographers with careers of experience in the best ways to provide as much data as clearly as possible. They've studied cartography and learned it as an art. And raster charts are wonderful works of art - you see tables, bags, wall hangings, etc all made from raster charts. I've never seen a vector chart wall hanging. Vector charts are drawn from a database by programmers (like me) who have spent their careers learning how to develop software. I have never yet met a cartographer who was one of the programmers of any of these products. They write code to render the data programmatically. It will never look as good as a raster chart drawn by a pro and it will never convey as much information. Vector charts can only hope to layer the missing information through interactivity.

Bottom line...both are good, different, and complement each other - we always have both showing to help make decisions when confusing situations come up.
Jeffrey S is offline   Reply With Quote
Old 12-20-2014, 02:10 PM   #45
Senior Member
 
Richard W's Avatar
 
City: Florida USA and Ontario Canada
Country: The 3rd Rock from the Sun
Vessel Name: anytime
Vessel Model: 2007 Chaparral 270 Signature LOA 29'
Join Date: Jan 2014
Posts: 303
The way I see it ...

Technology:

Vector based chart can be zoomed in and out without loosing visual quality as it is rendered from data using formulas and not from pixels. Vector chart can layer multiple sets of data coming from various sources, the layers can be manipulated and/or switched on and off as desired.

Raster chart is a flat single layer rendering of a traditional paper chart of fixed scale. Trying to zoom it beyond intended scale only enlarges the exiting image by creating "bigger pixels" and not more detailed rendering.

Data source and quality:

Either chart type, raster or vector, is only as good as the data used to create it.

The vector method allows for more data and data layers to be "hidden" behind it to be used as/when needed. Some chart publishers use NOAA or official government charts to create their vector cartography (like Navimatics). Some (like Navionics and perhaps others) use multiple sources of data to enhance the base NOAA charts.

The vector method allows for this ... does not mean the vector chart is inherently better, the various data sets that can be put behind it can make it better.
__________________
Best ... Richard
Richard W is offline   Reply With Quote
Old 12-20-2014, 03:28 PM   #46
Scraping Paint
 
City: -
Country: -
Join Date: Oct 2007
Posts: 13,748
We use only vector charts in our plotters for the reasons I've stated. However, we also have the relevant paper charts at the helm at all times. When entering waters we are not very familiar with and that pose some challenges we use both. The paper charts give us the "real world" view to compare the plotter displays to.

So far we have never experienced the position drift Jeffrey mentioned as we've zoomed in on our C-map vector charts but that's not to say it can't be there in other locations or with other suppliers' plotters or charts. It's certainly something to be aware of and guard against lest one end up where they don't want to be.
Marin is offline   Reply With Quote
Old 12-20-2014, 03:40 PM   #47
Guru
 
JDCAVE's Avatar
 
City: Lions Bay, BC
Country: Canada
Vessel Name: Phoenix Hunter
Vessel Model: Kadey Krogen 42 (1985)
Join Date: Apr 2011
Posts: 1,602
Quote:
Originally Posted by Richard W View Post
This exemplifies the confusion and explains some misunderstandings ...

The NOAA charts come in two versions ... you can have NOAA raster and NOAA vector chart. They both look similar and are equally accurate as are based on the same data.

One cannot compare chart technology (raster/vector) with chart source/content as provided by creator/publisher. These are two different qualities/criteria ... one is the technology or method used to render a chart, another is content, and specifically quality and accuracy of the data used to create a chart.

I'm not confused at all. The NOAA charts are raster. The CE charts are Vector. I can switch between them within the CE plotting software.


Jim
Sent from my iPad using Trawler Forum
JDCAVE is offline   Reply With Quote
Old 12-20-2014, 03:50 PM   #48
Guru
 
tpbrady's Avatar
 
City: Anchorage/Wrangell
Country: USA
Vessel Name: Silver Bay
Vessel Model: Nordic Tug 42-002
Join Date: Sep 2012
Posts: 505
Mbevins

I found the CMAP charts scaled better than the NOAA vector charts in some areas when zooming in. CMAP included detail that the NOAA vector chart didn't include that you could see on the raster charts and kept that detail as you zoomed in. Alaska charts are all over the "map" so to speak. When I have steamed through land they generally all agree I have steamed through land. I'm glad the boat didn't notice any lack of water.

Tom
tpbrady is online now   Reply With Quote
Old 12-20-2014, 04:05 PM   #49
Guru
 
JDCAVE's Avatar
 
City: Lions Bay, BC
Country: Canada
Vessel Name: Phoenix Hunter
Vessel Model: Kadey Krogen 42 (1985)
Join Date: Apr 2011
Posts: 1,602
Charts on tablet, pc, laptop or LCD off chartplotter

I agree with Jeffery's comments wrt raster vs vector, except for the comments about accuracy vs precision...
""Both raster and vector have identical accuracy. They both come from the exact same source. But since the vector charts are comprised of data points and objects, they can be zoomed to a much greater level and still look clean and crisp - that's providing more precision."

That's not precision. As a statistical term, precision refers how the repeated results of an experiment or sampling of data matches the other observations. Accuracy is a term to quantify bias, i.e. how close you are to the bullseye or systematic error.

I borrowed this slide from http://www.sophia.org/tutorials/accu...d-precision--3

Name:  ImageUploadedByTrawler Forum1419109205.862137.jpg
Views: 160
Size:  72.7 KB

It was a constant debate that went on all throughout my career!

Precision with your plotter probably has more to do with how precise your GPS puts your location on the chart rather than the chart itself which has to do primarily with accuracy.


Jim
Sent from my iPad using Trawler Forum
JDCAVE is offline   Reply With Quote
Old 12-20-2014, 05:22 PM   #50
Senior Member
 
Richard W's Avatar
 
City: Florida USA and Ontario Canada
Country: The 3rd Rock from the Sun
Vessel Name: anytime
Vessel Model: 2007 Chaparral 270 Signature LOA 29'
Join Date: Jan 2014
Posts: 303
Quote:
Originally Posted by JDCAVE View Post
I'm not confused at all. The NOAA charts are raster. The CE charts are Vector. I can switch between them within the CE plotting software.
Sorry, I was ... you meant your NOAA charts are raster.
The NOAA charts come in both, vector and raster, formats.
__________________
Best ... Richard
Richard W is offline   Reply With Quote
Old 12-20-2014, 05:56 PM   #51
Guru


 
City: Full-time onboard
Country: USA
Vessel Model: Trawler
Join Date: Oct 2007
Posts: 937
Quote:
Originally Posted by JDCAVE View Post
Precision with your plotter probably has more to do with how precise your GPS puts your location on the chart rather than the chart itself which has to do primarily with accuracy.
Precision and accuracy come into play with GPS plotting too but that's not what this topic was about.

I still think this is a precision issue.

The first image attached (actually, the first small image which is the second - what's with the attachment handler here?) is a raster snippet near Ft Myers from NOAA chart 11427_1. It's a 1:40,000 chart and the snippet is drawn at 1:35,000 which is slightly overzoomed. That display represents the accuracy of the chart data. It is meant to be displayed at a zoom level no higher than this. NOAA is saying with this that the objects shown at this zoom level are in correct orientation and position to each other. Of course we know that things move and there are many chart errors, but this is the best data we've got. This is a very recent chart generated within the last month.

Note R38 and G37. R38 looks to be right on the channel and G37 looks to be a little off the marked channel. The question is, how close can you get to R38 from this image? Almost anyone looking at that would decide that R38 looks to be in pretty safe water.

The second image (really the third) is the vector chart drawn at the same scale. I'm not sure that it gives any additional information although the channel surely looks a little narrower due to the very thin lines used to draw the channel by this developer. So the first question - is the channel really more narrow than the raster chart shows? The reason it looks narrower is because the lines drawn on the vector chart are much more precise. They define a much more exact location, although, and-here's-the-important-part, the line is no more accurate than the raster chart. It's just using the error from the raster line and drawing it in any number of places that the developer could have selected (center of the raster line, inside edge, outside edge?). Most probably, the developer of this never gave it any consideration about how and where to draw the line or even how thick to make it. They kept writing the code until they thought it looked nice.

The third image (really the first) shows what happens when you zoom into the vector snippet far. It now takes a new direction where small errors in accuracy and developer decisions are multiplied and magnified. Accuracy-wise, R38 is still within the error threshold of the position from the original chart but adding a lot of decimal places to it's location (from the multiplication) has given a very poor impression to the guy at the helm. Now it looks like you better stay pretty far away from R38. And in reality, that's wrong. The channel is actually right up against R38 but the multiplication of the error of R38's position along with multiplying the error in the channel location is causing a problem. And although the pilot can zoom in this far with a vector chart, the resulting image is giving a very bad impression of reality.

There are plenty of examples where the opposite happens - it looks like there is plenty of water on one side of a channel when you zoom way in with the vector chart but actually, the aid is off the channel and it gets shallow. That usually cause hazard markers to be inserted because the boater became surprised when it showed there should be good depths when zoomed in but the real chart at the proper scale couldn't give that information. I see that about 2 times every day.
Attached Thumbnails
Click image for larger version

Name:	precision3.png
Views:	68
Size:	175.9 KB
ID:	35579  
Attached Images
  
Jeffrey S is offline   Reply With Quote
Old 12-20-2014, 08:15 PM   #52
Senior Member
 
Richard W's Avatar
 
City: Florida USA and Ontario Canada
Country: The 3rd Rock from the Sun
Vessel Name: anytime
Vessel Model: 2007 Chaparral 270 Signature LOA 29'
Join Date: Jan 2014
Posts: 303
Interesting discussion ... seems to me that both, accuracy and precision, are at play here. I cannot eloquently describe either as English is my second language but let me offer this link that defines both in context of cartography and surveying ... Error, Accuracy, and Precision

Looking at two examples provided by Jeffrey brought to the same "resolution" and superimposed ... you cannot miss the fact that they both were rendered from the same data. That data was obtained at certain level of accuracy that matched the desired scale of the original (printed) chart.



This is an example of how precision of the new rendering methods exceed the accuracy of old data. I would refer again to the point number 6 in the link already provided ... 6. Beware of False Precision and False Accuracy!

The simple solution is to render the data with the precision appropriate to the data accuracy. Simply, and some devices and apps are doing that quite well, do not allow to zoom in below what the intended "scale" of the original chart data calls for.

In all other cases, we the users have to be aware of this potential misuse/misinterpretation of chart data. We are still in the transitional period, the surveying methods and cartographic date will only get better with time.
__________________
Best ... Richard
Richard W is offline   Reply With Quote
Old 12-21-2014, 07:19 AM   #53
Guru
 
ranger42c's Avatar
 
City: Maryland
Country: USA
Vessel Model: 42' Sportfish
Join Date: Aug 2013
Posts: 3,154
Nifty superimposition, Richard.

-Chris
__________________
South River, Chesapeake Bay
ranger42c is offline   Reply With Quote
Old 12-21-2014, 11:51 AM   #54
Guru
 
JDCAVE's Avatar
 
City: Lions Bay, BC
Country: Canada
Vessel Name: Phoenix Hunter
Vessel Model: Kadey Krogen 42 (1985)
Join Date: Apr 2011
Posts: 1,602
Nice overlay Richard. But what you are showing here has nothing to do with precision. Precision has to do with sampling. Sampling information on depth. Sampling information on position. Echo sounders and gps are incredibly precise. Accuracy has to do with whether there is some sort of systematic bias going on. So if you take a bunch of soundings and use a model on the data to calculate contours in between the soundings. You then then go back and actually measure the depth 10 times on the 100 m contour at a particular location and determine that location 10 times with Gps, and you find that the depth is actually 90 m instead of 100 m. THAT is accuracy--the contour is out by 10 m or 10%. The contour is biased high at that location. Models have perfect precision--they always give the same result, given the exact same data, but they can be quite inaccurate and are therefore biased.

Another example, take chart CHS 349002, Fraser River, Sand Heads to Douglas Is. The chart is dated 2004. The soundings shown are almost certainly no longer exactly what you might encounter--silting and dredging being responsible. The original soundings were probably originally PRECISE, but they are no longer ACCURATE, because of the changes in the area. If dredging is responsible, the measured depth would be deeper than the charted depths. The charted depths are biased low. Elsewhere off the channel, there has been silting and measured depths are shallower than the charted depths, which are biased high.

Another: let's say the Datum applied to the chart is too high by 0.1 m. All depths on the chart are therefore biased low by 0.1m. This is bias and is a question of accuracy, not precision.

Another: in a hypothetical example, in July 2014 CHS returns to a location (52 degrees 47.0263' N 132 degrees 19.295'W) off the west coast of Haida Gwaii, where soundings were last taken in 1978. In 1978, the sounding was 865 fathoms, (1582m). In 2014 they record 1605 on July 6, but they return on July 9 and record 1610 and 1608 on July 10 after correction for datum or precision of +- 0.3%. However some of the issue of precision also has to do with GPS precision which is +- 5m. The difference between the 1978 and the 2014 soundings, includes elements of both precision and accuracy. The 1978 soundings were likely neither as precise or as accurate as 2014, however the degree that each type of error contributes to the difference cannot be determined from the available information. However, the 2014 researchers determined that better Salinity and Temperature information is now available, which if applied to the 1978 soundings, would result in results 0.2% higher estimates, which would represent improved ACCURACY.



Jim
Sent from my iPad using Trawler Forum
JDCAVE is offline   Reply With Quote
Old 12-21-2014, 01:10 PM   #55
Senior Member
 
Richard W's Avatar
 
City: Florida USA and Ontario Canada
Country: The 3rd Rock from the Sun
Vessel Name: anytime
Vessel Model: 2007 Chaparral 270 Signature LOA 29'
Join Date: Jan 2014
Posts: 303
When you look at the entire process from data measurement and acquisition to recording and calculations to rendering from recorded data, not just at the data acquisition step, it is about both, accuracy and precision. I will refer you to the link already provided for the explanation.

The problem of mismatch between low data accuracy and too precise calculation and rendering discussed here is not new. It goes back as far as the invention of digital calculators and computers. Engineering students in the 70's who used the new then digital calculators were failing the exams by simply using mindlessly very precise calculators.

Example, if the data is provided with the accuracy of two decimal points and one uses calculator with a precision of eight decimal points and does not mitigate the intermediary calculations, the end result although seemingly very precise contains unmitigated error introduced by the too high level of calculation precision. The error could be avoided by rounding and limiting the intermediary calculations to three decimal points, and rounding the final result back to two decimal points.

Something similar is happening with old chart data, captured with relatively low accuracy, when converted to vector format and interpreted (mindlessly) by very precise rendering devices.

The example of superimposed charts above shows this ... the vector chart locates navaids in (seemingly) very precise points while the original raster chart shows fuzzy areas where they could be located.
__________________
Best ... Richard
Richard W is offline   Reply With Quote
Old 12-21-2014, 01:54 PM   #56
Guru
 
JDCAVE's Avatar
 
City: Lions Bay, BC
Country: Canada
Vessel Name: Phoenix Hunter
Vessel Model: Kadey Krogen 42 (1985)
Join Date: Apr 2011
Posts: 1,602
Richard: there is nothing that I have said above that is inconsistent with your linked reference. However, I stand by what I have said. Precision in this instance relates almost entirely to measurement error, not post-processing of the data. In addition, it is correct procedure to NOT round until all intermediate steps are taken in a series of calculations (see floating point). Rounding errors are cumulative and can compound differences over a number of steps. I am uncertain if floating point problems are an issue to be concerned with cartography, however there may be effects in algorithms used in post processing of data. I would classify this as model error, likely resulting in "bias" rather than reproducibility error (precision).


Jim
Sent from my iPad using Trawler Forum
JDCAVE is offline   Reply With Quote
Old 12-21-2014, 02:14 PM   #57
Guru
 
psneeld's Avatar
 
City: Avalon, NJ
Country: USA
Vessel Name: Freedom
Vessel Model: Albin 40
Join Date: Oct 2011
Posts: 15,853
Except in extreme cases, I would like to see how the average nautical steering and nav skills matched up to this debate.

My guess is nada....most would nav just fine with either or neither.....and of course certain charts, certain areas, etc are always the exception.

As said.....preference at this point... hardly is either a clear winner in making nav "safer/better".

Otherwise, what wouldn't governing bodies weigh in and mandate......
psneeld is offline   Reply With Quote
Old 12-21-2014, 03:02 PM   #58
Guru


 
City: Full-time onboard
Country: USA
Vessel Model: Trawler
Join Date: Oct 2007
Posts: 937
Quote:
Originally Posted by psneeld View Post
Except in extreme cases, I would like to see how the average nautical steering and nav skills matched up to this debate.
I'm in a unique place to comment on that as it relates to charts. For 7 years, I've been the one validating every hazard marker in ActiveCaptain. We get about 15 new ones every day of the week on average. I'd imagine that 35% of them are for "shoaling" in one form or another. In a few cases each day, a skipper goes aground and while waiting for TowBoatUS, SeaTow, or other towing (SeaStart/UK, etc), they enter the hazard explaining how it wasn't their fault. I totally understand this and it's a normal reaction. I've grounded twice along the ICW (pre-ActiveCaptain!) and I know the feeling of wanting to find a reason.

So when I look at the info, there are some places where the channel is definitely spilling in. But in other cases, I can simulate the problem by zooming in with vector charts. And I think it's the charts themselves causing the problems.

Two example places where this happens a lot - Currituck Sound where you have to stay further off the greens than you think and the channel isn't well drawn when zooming in far. The second is the area just south of Haulover Canal and north of Titusville. Those 2 places are good for 4-5 hazards entered every week in spring and fall.

In some cases, a hazard is justified, not because there is shoaling happening but instead because it's a place where you need to stay alert to remain in the channel. What is happening is that the channel looks narrow. So the reaction is to zoom way in, long past the ability to get real information. It makes you feel good that it's showing you within the channel. But in reality, you're moving to the edge and outside it.

As chartplotters are all using vector charts today, it's becoming more common. Watching these channels on raster charts starts getting blurry if you zoom in so you don't accept the "fact" of being in the channel as readily.

Removing electronics altogether would force the skipper to use the aids much more and probably pay a lot more attention to their position from them. Still, electronics allow a lot more normal people to get out on the water in a safe way and surely produce fewer accidents per community mile than in the pre-electronics days where schooner sinkings were run-of-the-mill.
Jeffrey S is offline   Reply With Quote
Old 12-21-2014, 04:07 PM   #59
Scraping Paint
 
City: -
Country: -
Join Date: Oct 2007
Posts: 13,748
Quote:
Originally Posted by Jeffrey S View Post

Removing electronics altogether would force the skipper to use the aids much more and probably pay a lot more attention to their position from them.
Excellent point. And it's why we advocate that boat operators use ALL the aids available to them, including the physical navaids, their eyes, their common sense, and the healthy human trait of caution.
Marin is offline   Reply With Quote
Old 12-21-2014, 04:42 PM   #60
Guru
 
psneeld's Avatar
 
City: Avalon, NJ
Country: USA
Vessel Name: Freedom
Vessel Model: Albin 40
Join Date: Oct 2011
Posts: 15,853
Probably 95 % of the ungroundings I do (around 50-75 a year), and many times the skipper says, I've run this channel for xx+ years and never had a problem.

I nicely as I can line things up for them and show them they are well outside the channel, sometimes as much as the channel is wide.
__________________

psneeld is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off





All times are GMT -5. The time now is 11:11 PM.


Powered by vBulletin® Version 3.8.8 Beta 4
Copyright ©2000 - 2017, vBulletin Solutions, Inc.
Search Engine Optimization by vBSEO 3.6.0
Copyright 2006 - 2012