[D'rek]

ArchieVist, on 16 September 2016 - 07:36 PM, said:

D'rek/Werthead,

I love this Stratem map. Would you mind if I added it to the Wiki? With a credit of "Created by D'rek, modified by Werthead", or something similar?

I have no objections.

[Werthead]

Sure, go for it.

It occurs to me that there is no reasonable expectation that we'll get any more of the Malazan world revealed in future: the Toblakai Trilogy will presumably cover a lot of lands we've already seen, presumably Walk in Shadow will have the existing Kharkhanas Trilogy maps and the future ICE books will remain predominantly set on Quon Tali and Malaz Island. Maybe if they go a lot further we'll eventually see Stratem when the Empire takes on the Crimson Guard there, but that's a long way down the pipe and the chances of us seeing the rest of Seven Cities, central Lether, the southern tip of Assail, Genstoel or Umryg seem fairly non-existent. Maybe the RPG or companion book will happen, but after 10-15 odd years of false starts on that front, it seems unlikely.

I wonder if on that basis we could get Steve and Ian to show us their rough working maps in full, and look to create a proper canon map?

[D'rek]

Werthead, on 17 September 2016 - 02:51 PM, said:

Sure, go for it.

It occurs to me that there is no reasonable expectation that we'll get any more of the Malazan world revealed in future: the Toblakai Trilogy will presumably cover a lot of lands we've already seen, presumably Walk in Shadow will have the existing Kharkhanas Trilogy maps and the future ICE books will remain predominantly set on Quon Tali and Malaz Island. Maybe if they go a lot further we'll eventually see Stratem when the Empire takes on the Crimson Guard there, but that's a long way down the pipe and the chances of us seeing the rest of Seven Cities, central Lether, the southern tip of Assail, Genstoel or Umryg seem fairly non-existent. Maybe the RPG or companion book will happen, but after 10-15 odd years of false starts on that front, it seems unlikely.

I wonder if on that basis we could get Steve and Ian to show us their rough working maps in full, and look to create a proper canon map?

Given how they write/worldbuild, I don't think their draft maps (if they even have any for the remaining areas) are going to have any actual detail, though. From what we've heard and how they write, it sounds more like their working maps are just blobs of areas, and the specific details only get hashed out when they actually write about the area in-depth.

It's possible though, and nothing to lose by asking.

[PLUGO]

Now sure who's into podcasts and such, but This episode of IMAGINARY WORLDS features some prominent fantasy map makers.

Quote

J.R.R. Tolkien spent decades mapping out Middle-earth on graph paper because everything had to be invented from scratch (and given a name.) Many of his maps weren't even published until after he died, but today's fantasy cartographers owe a great debt to his work. They also have a post-modern understanding that to create a believable fantasy map, they have to sow doubt in the minds of readers whether we should trust the mapmakers. With Isaac Stewart, Priscilla Spencer, Ethan Gilsdorf and Stefan Ekman.

[Werthead]

D'Rek's map is now Spherified for everyone's pleasure.

I'm impressed at how well this works. It also makes it a bit more obvious how Lether can be isolated given it location. Stick it on a sphere and western Lether ends up quite a way from everyone else. So that works.

Also, staying away from the poles is extremely helpful. Trying to make this work for Westeros and it doesn't, really, because the continent connects to the north pole and that goes weird very quickly.

[Werthead]

Also, Malazan Spherified with Names

(go to Options and reduce the "Light Intensity", the default setting is a bit bright!)

This post has been edited by Werthead: 13 December 2016 - 06:35 PM

[Aptorian]

Okay that thing is seriously cool. Somebody should send Erikson a message and hear what he thinks about the thing.

You should make a separate thread for it and have a mod sticky it everywhere.

I'm curious about the globe though. How does the proportions actually fit? Is there actually that big of a sea between Lether and Assail or is it just a case of that's the way it looks on a globe unless you try and stretch and break everything?

How would ice caps look compared to everything else if you tried to apply them to the Malazan globe?

EDIT: I went ahead and sent Erikson a mail via his website. I'm curious to see if he has a response.

This post has been edited by Apt: 13 December 2016 - 06:54 PM

[Werthead]

The original map proportions were from 2004, someone worked out what the dimensions would be for a flat map if the world was the same size as Earth. Then over the course of 12 years (!) various people have fiddled around with the continents on that flat image, until here we are.

So yes, the Domain Ocean is pretty huge, the closest to Erikson's version of the Pacific.

[Egwene]

Hadn't really noticed the continental drift between Assil and Korelri before!

Amazing - we want it for the Wiki!!!!!!!!

[Gorefest]

Okay, this is just seriously cool. i will spend the next three hours aimlessly spinning the globe and grinning.

[MTS]

That is amazing (I'm also going to use this tool for my teaching!). Seeing the continents in situ (as it were) on a sphere adds so much clarity.

I also like how you can make some sort of sense of it climatically. The only areas around the tropics that have jungle areas for instance are Dal Hon and Himatan (which are themselves quite close to each other). The other areas around the equator have all been affected in some way by external events (Lether by Jaghut ritual; northern Korel by fall of CG; Assail by Jhag inhabitants).

[D'rek]

This is really cool!

Apt, on 13 December 2016 - 06:41 PM, said:

I'm curious about the globe though. How does the proportions actually fit? Is there actually that big of a sea between Lether and Assail or is it just a case of that's the way it looks on a globe unless you try and stretch and break everything?

How would ice caps look compared to everything else if you tried to apply them to the Malazan globe?

Werthead, on 13 December 2016 - 07:41 PM, said:

The original map proportions were from 2004, someone worked out what the dimensions would be for a flat map if the world was the same size as Earth. Then over the course of 12 years (!) various people have fiddled around with the continents on that flat image, until here we are.

So yes, the Domain Ocean is pretty huge, the closest to Erikson's version of the Pacific.

I think Apt meant more how does the software deal with the proportions of converting a flat map to globe.

Looks like the program is doing a reverse plate carrée/equirectangular transformation. It assumes the image you upload has equally spaced latitude and longitude, and that the vertical and horizontal center lines are the equator and central meridian of the map. Then it goes through the uploaded image pixel by pixel and draws the image onto the globe using a relationship like:

For pixel {X,Y}, transpose this to GlobeLocation{Longitude = X/cos(V)+H, Latitude = Y+V}

X and Y being pixel coordinates, V and H being the vertical and horizontal distances from 0,0 to the the equator and central meridian (in an equirectangular projection these are distances to the "standard parallels" where the projection is to-scale (i.e. where the projection crosses the earth's surface), but I'm assuming the tool just uses plate carrée where the standard parellels are the equator).

An equirectangular/plate carrée projection keeps the vertical spacing of a globe, but stretches out the poles to the corners of the map, so as you go north or south away from the equator things get progressively more horizontally stretched.

So with the reverse transformation from map to globe, you can see we're getting the opposite effect of a lot of squishing at the top and bottom, like in Umryg and Genebackis. This might even be exacerbated because my map isn't actually the right dimensions to be an equirectangular map for a planet that is roughly spherical, and is probably a bit too short. (It also has a ton of extra space on the sides, so the Domain Ocean is huuuuge).

I can give a try to making a properly dimensioned map, and maybe even one that over-compensates for the reverse equirectangular transformation in order to make a better globe from it. I can also toss some ice sheets in to the far north and south so the poles don't look so empty!

[Werthead]

D, on 14 December 2016 - 03:08 AM, said:

I think Apt meant more how does the software deal with the proportions of converting a flat map to globe.

Looks like the program is doing a reverse plate carrée/equirectangular transformation. It assumes the image you upload has equally spaced latitude and longitude, and that the vertical and horizontal center lines are the equator and central meridian of the map. Then it goes through the uploaded image pixel by pixel and draws the image onto the globe using a relationship like:

For pixel {X,Y}, transpose this to GlobeLocation{Longitude = X/cos(V)+H, Latitude = Y+V}

X and Y being pixel coordinates, V and H being the vertical and horizontal distances from 0,0 to the the equator and central meridian (in an equirectangular projection these are distances to the "standard parallels" where the projection is to-scale (i.e. where the projection crosses the earth's surface), but I'm assuming the tool just uses plate carrée where the standard parellels are the equator).

An equirectangular/plate carrée projection keeps the vertical spacing of a globe, but stretches out the poles to the corners of the map, so as you go north or south away from the equator things get progressively more horizontally stretched.

So with the reverse transformation from map to globe, you can see we're getting the opposite effect of a lot of squishing at the top and bottom, like in Umryg and Genebackis. This might even be exacerbated because my map isn't actually the right dimensions to be an equirectangular map for a planet that is roughly spherical, and is probably a bit too short. (It also has a ton of extra space on the sides, so the Domain Ocean is huuuuge).

I can give a try to making a properly dimensioned map, and maybe even one that over-compensates for the reverse equirectangular transformation in order to make a better globe from it. I can also toss some ice sheets in to the far north and south so the poles don't look so empty!

Do you know a good way of reversing the process? On the ASoIaF maps, the maps are all accurate at a local level, but have been combined into a world map, so on the flat maps they look like they should from space, rather than they should on a flat map. On the flat base map, the North and Beyond the Wall should expand outwards significantly so, when you put them on the globe, they retain the same propotions they do on the pre-existing flat maps.

I've been trying to manually work out a way of doing it but I'm not getting anywhere.

[Egwene]

Sounds all totally impressive although I don't understand a word of it

[D'rek]

Werthead, on 14 December 2016 - 04:04 PM, said:

D, on 14 December 2016 - 03:08 AM, said:

I think Apt meant more how does the software deal with the proportions of converting a flat map to globe.

Looks like the program is doing a reverse plate carrée/equirectangular transformation. It assumes the image you upload has equally spaced latitude and longitude, and that the vertical and horizontal center lines are the equator and central meridian of the map. Then it goes through the uploaded image pixel by pixel and draws the image onto the globe using a relationship like:

For pixel {X,Y}, transpose this to GlobeLocation{Longitude = X/cos(V)+H, Latitude = Y+V}

X and Y being pixel coordinates, V and H being the vertical and horizontal distances from 0,0 to the the equator and central meridian (in an equirectangular projection these are distances to the "standard parallels" where the projection is to-scale (i.e. where the projection crosses the earth's surface), but I'm assuming the tool just uses plate carrée where the standard parellels are the equator).

An equirectangular/plate carrée projection keeps the vertical spacing of a globe, but stretches out the poles to the corners of the map, so as you go north or south away from the equator things get progressively more horizontally stretched.

So with the reverse transformation from map to globe, you can see we're getting the opposite effect of a lot of squishing at the top and bottom, like in Umryg and Genebackis. This might even be exacerbated because my map isn't actually the right dimensions to be an equirectangular map for a planet that is roughly spherical, and is probably a bit too short. (It also has a ton of extra space on the sides, so the Domain Ocean is huuuuge).

I can give a try to making a properly dimensioned map, and maybe even one that over-compensates for the reverse equirectangular transformation in order to make a better globe from it. I can also toss some ice sheets in to the far north and south so the poles don't look so empty!

Do you know a good way of reversing the process? On the ASoIaF maps, the maps are all accurate at a local level, but have been combined into a world map, so on the flat maps they look like they should from space, rather than they should on a flat map. On the flat base map, the North and Beyond the Wall should expand outwards significantly so, when you put them on the globe, they retain the same propotions they do on the pre-existing flat maps.

I've been trying to manually work out a way of doing it but I'm not getting anywhere.

Can you show me an example?

[Werthead]

D, on 14 December 2016 - 05:53 PM, said:

Can you show me an example?

Right, so the current world map looks like this. We've worked out that the northernmost edge of the world is only about 200-300 miles south of the north pole itself.

When you put it on the globe, it stretches the North and Beyond the Wall like so.

But, because the maps are all accurate on a local level, on the globe it should look like it does on the flat map. Which, I believe, means we have to deliberately distort the North and Beyond the Wall on a 2D map, so that when applied to the globe, it makes them look right. I have no idea how to do that or make sure it ends up looking correct.

Cheers.

[D'rek]

Werthead, on 14 December 2016 - 06:44 PM, said:

D, on 14 December 2016 - 05:53 PM, said:

Can you show me an example?

Right, so the current world map looks like this. We've worked out that the northernmost edge of the world is only about 200-300 miles south of the north pole itself.

When you put it on the globe, it stretches the North and Beyond the Wall like so.

But, because the maps are all accurate on a local level, on the globe it should look like it does on the flat map. Which, I believe, means we have to deliberately distort the North and Beyond the Wall on a 2D map, so that when applied to the globe, it makes them look right. I have no idea how to do that or make sure it ends up looking correct.

Cheers.

Ok, yeah, since the only "looks right" you have is the flat map (not the globe), I don't think there's any "reverse process" you can use to automatically create a deliberately-distorted map that will transfer to the globe correctly. Have to do it by hand (same as I'm intending to do).

Photoshop might have some sort of variable-stretch tool where you could specify zero stretch in the middle, how much at the top and it transitions smoothly from one to the other. But I'm not sure. I might just do mine in a series of "strips" of increasing stretch, and manually fixing the lines that get broken.

[Gorefest]

Speaking of adding ice sheets to the poles, do we have any indication that there would be ice at the poles? Isnt that a rather unfounded Earth extrapolation in light of the different existing climate zones on the current map? Or does it make sense in-world?

[D'rek]

Gorefest, on 14 December 2016 - 10:11 PM, said:

Speaking of adding ice sheets to the poles, do we have any indication that there would be ice at the poles? Isnt that a rather unfounded Earth extrapolation in light of the different existing climate zones on the current map? Or does it make sense in-world?

Stonewielder *may* have a reference to southerly ice. As Ereko, Kyle and co. are departing Stratem they see ice sheets/icebergs to the south. To me, that passage reads as if it is before they reach the actual Jacuruku ice sheets, but it may be the intent was just that these were more of the Jacuruku ice sheets.

Genebackis has tundra in the north.

Other than that, no I don't really recall anything specific in that regard.

[MTS]

Umryg in the far north of the world has a connection with the Jaghut, and Envy mentions that the Pannion Domin was originally a Genostelian colony (also in the far north), perhaps founded from refugees fleeing T'lan Imass persecution (thus the whole story with Pannion).

Dont have access to the books atm though, so I cant confirm that.