Limit zoom out on Raster charts

[wizard-merlin]

I've created some Raster charts from paper charts (using Mapcal), scale 1:30000. When I use these with CM93 Vector charts I am not able to zoom out to a sufficient level to get an overview. The max true scale zoom out showing is 16100. Against this, the zoom in is too much getting down to pixel level which is not readable.

I have tried changing the scale in the creation of the KAP file but this seems to make no difference.

How can I make the zoom out level go to a true scale of (say) 100000, whereby the zoom in would be getting into sensible levels?

Is this something to do with pixel setup in the header in Mapcal ?

Not sure what to do next, so would appreciate any help here.

[idpnd]

I've had that testing marco's NZ charts, you get "stuck" in a chart. Could be worht a bug report.

[wizard-merlin]

I'm wondering if its due to the relationship between physical scan size and the area covered by the map? It seems the scale set up in the header is merely a text descriptor (?) and has little or no effect in OPENCPN.

I've tried scales ranging from 2 to 1000000000 in the .hdr file, and this has no effect at all in the way the file is displayed on the screen.

When I use the "same" (original) file (WCI - created from ChartDir) in SeaClear I have no problem zooming out.

Even if I take the Vector CM93 charts out of the quilt I can't zoom out.

[HappySeagull]

WIZ:..you bring up a good point...I can't zoom out to see the whole chart either,however the settings are tweaked(Zoom level same,quilting,mousezoom enabled etc.)I also have scanned papercharts(hrbr plans) and on some of these ,I can see three borders but never four.That would seem a clue as to the inner workings? ...If you make a resized companion version of your map,equal to the active window size,(per your idea above"smaller scan size")I think it'll work to see the whole chart(3 sides!)...kind of like making your own thumbnails,but calibrated.

I didn't notice this before because it hasn't been a nuisance in navigating but then, I have a pile of charts in various scales covering my puddle.

[HappySeagull]

Like this?the real chart is only about 595 pixel square on my 1024 wide laptop..

[cagney]

wizard-merlin

I can't repeat your findings. With charts that I have calibrated, I can zoom out to a scale factor of ~0,45, with Marcos NZ charts and some commercial charts tested the factor is ~0.64. For a chart in 1:30,000, this means about 1:65,000 and 1:45,000 respectively. This is quite reasonable, as any smaller scale makes the chart useless for practical navigation.

I'm wondering if its due to the relationship between physical scan size and the area covered by the map? It seems the scale set up in the header is merely a text descriptor (?) and has little or no effect in OPENCPN.

I've tried scales ranging from 2 to 1000000000 in the .hdr file, and this has no effect at all in the way the file is displayed on the screen.

While the scale stated in the kap files text header doesn't change the chart itself nor it's natural scale, it does play a certain role in how OpenCPN displays the chart when first opened. OpenCPN opens the chart in about double the scale stated, or in the minimum scale, if the scale given is way to small.
Remember to run "bsbfix" and "Force Full Database Rebuild", to be sure that a change of scale in the header will work.

To see what is going on with your chart, I need to have a look at it.

Finally, I have described what I found, I don't think this is a bug, more like a feature,....but perhaps these settings needs tweaking.

Thomas

[HappySeagull]

It is interesting.This one(same as above) goes from scale .47x to 60x!


but although a large chart can do the same,(.47 to pixels)you can never see it as a whole....Except in the thumbnails seen by hoveringalong the bottom...is that what you mean,WIZ?

[wizard-merlin]

but although a large chart can do the same,(.47 to pixels)you can never see it as a whole....Except in the thumbnails seen by hoveringalong the bottom...is that what you mean,WIZ?[/QUOTE]

That's exactly it. I can't really give a screen print, as its the absence of being able to see the whole chart that I'm missing. See also my note to Thomas

[wizard-merlin]

I tried reducing the pixel size to 200 per inch, which on zoom seemed reasonable. When I redid this (I went through the full build procedure) I had the same problem as Happy-Seagull - one side with borders showing, and the other half not visible. I had excluded the borders from the chart area, so this is strange. However, the zoom is now at ,45 which is around what Thomas / Cagney is getting.

I checked the .hdr file versus the chartcal.dir file, and the long lat seem to be the same, and it works well in SeaClear.

Taking Thomas' point though.. I navigate using a small netpad (still bigger than the std "proper" charts). The area I want to look at is an estuary of around 6 miles before getting to the full sea. Having failed to get tides working in OpenCpn I have added a picture I have of the tide change diretions over 12 hrs relative to high water. I want to be able to see these all at once, which the zoom out is not giving me.

The true scale is showing as 16000, with a magnitude of 1,1, but the scale is supposed to be 30000. The most I am seeing on the screen is 2 Nautical Miles, which is not enough for planning?

Am I doing something wrong in the creation of the files?

[wizard-merlin]

Just to add, I managed to get this working.

1. Reduce the pixels per inch to 200
2. Changed the scale on the header to 60000, not the 30000 which is the charts real scale.

Does this imply there is a relationship between pixels per inch (DU in the header file) and the nominated scale.

This cleared the problem I had with the borders and only part of the chart showing, and now gives me a much bigger area to look at - 8 miles, which is the max for my display for close sailing.

[HappySeagull]

Wiz-Mmm....I'd approach this as simply OPenCPN's chart window is 1020x610 on my laptop...(but not so on a netbook),so I won't see the whole chart unless it "fits"inside those parameters (and one side at least will be anchored...).The screen also has a native resolution too(96dpi?) so detail at that size is limited if you wanted to zoom in ...so I'd be tempted to make two kaps,different sizes.

I found this in my infos from a description of BSB headers by Stuart Cunningham
DU=Drawing Units in pixels/inch (same as DPI resolution) e.g. 50, 150, 175, 254, 300

[HappySeagull]

...and i reconverted a homemade chart where DU=342.Changed dpi from 96 to 48 and the new kap has DU=48 but overall size stays the same= no diff per screen size and zoomability.It has to be actual size in pixels...?
Oh,and you noticed that the SCALE just seems to line em up along the bottom chart selection bar?- it just sets the order of charts as scale-large-to-small.

[wizard-merlin]

Happy Seagull, two thoughts.

With regard to zoom (out) being limited to screen size: if I look at the "same" chart (but WCI) in Seaclear I can zoom out until the map is quarter the size of the screen. I know this is no use for navigating, but it seems to indicate there is an opencpn or BSB conversion issue.

On the scaling: I used an old version of PS to edit and downsize the pixels. There is a selection whether to change size as well, or to keep it linked. Might this be the cause of your problem? Have you tried changing the scale?

I've done a few now, and find that I need to go to 200pxi, AND change the scale to 2x the actual scale (i.e 30000 becomes 60000). If I don't change the scale I have problems.

[wizard-merlin]

..and I can't really get my head around "scale" as an absolute on digitalised charts, unless it is to do with pxi -

I suppose the true scale should be calculable? variables would be: Width of Chart, Height, known height in miles, and pixels per linear inch, i.e sqrrt(pxi)??

will play with this.

Could you give me the link for the BSB advice, in case its one I don't have.

[HappySeagull]

Originally Posted by wizard-merlin

..and I can't really get my head around "scale" as an absolute on digitalised charts, unless it is to do with pxi -

I suppose the true scale should be calculable? variables would be: Width of Chart, Height, known height in miles, and pixels per linear inch, i.e sqrrt(pxi)??

will play with this.

Could you give me the link for the BSB advice, in case its one I don't have.

libbsb - BSB File format