Maximum input tif size?

[quartz]

Is there a limit on the input file size?
I have just downloaded Maptiler (Windows version), and tried running
two different images - one is 1.2GB, the second is 75MB.
Both processes got through to the scrren "Details about the tile
pyramid" and then ran for over an hour, using 10 to 17% of RAM and
(for the smaller file) approx. 4MB of memory.
Am I just being impatient?

I have a decent laptop (4GB Ram, i7 processors) so I don't think that
is the limitation. Running on Windows 7 OS.

[Joey Kopera]

Echo... I've been reluctant to purchase the map tiler pay version as of yet because every file I try to load (usually 2+ GB, with really needing something to tile 6-7 GB files...) has given me the "Too Big" exception.

___________________

Joey Kopera

joeyk...@gmail.com

___________________

On Apr 17, 2013, at 9:48 PM, deepw...@winterroot.net wrote:

Right.. what causes the 'too big' exception?  I am using the pay version of maptiler (Maptiler Start).  why won't it process my file?

--
You received this message because you are subscribed to the Google Groups "MapTiler User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maptiler+u...@googlegroups.com.
To post to this group, send email to mapt...@googlegroups.com.
Visit this group at http://groups.google.com/group/maptiler?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Klokan Petr Přidal]

MapTiler Free and Start is targeting hobbyist - that is the reason for 10.000 x 10.000 pixels limit and a maximally simplified interface.

For professionals and business use we have the MapTiler Cluster (http://www.maptiler.com/) - which has a lot of additional features, full automation, no input file size limits, merging of multiple files into a seamless tile set, possibility to use all CPU cores of your computer, keep control over rescaling, render into custom map projections, soon direct support for MBTiles, etc.

With MapTiler Cluster you can rapidly render optimised tiles for whole countries and even continents - have a look for example at http://os.tileserver.com/.

We can provide you with a demo of this utility - so you can try it on your computer and on your own data for free, before you purchase a license.

If you are interested, please contact us at in...@klokantech.com.

Best regards,

Petr

--
http://blog.klokan.cz/
http://www.maptiler.org/
http://www.oldmapsonline.org/

[deepw...@winterroot.net]

So this is really annoying.. Cluster is a $1500 product if I'm not mistaken.  For a $20 app, MapTiler should not have a 'Too Big' limit - it should just run slowly.   I get the 'Too Big' limitation even on areas of relatively easy to work with size.  Cluster is great if you need to run a cluster, but sometimes you just need to build tiles one time that have a bigger area.

gdal2tiles is a free script that anyone can use to tile files of their choosing.  So far, my purchase of MapTiler Start has been a lot of frustration and limitation - I notice that often when people post questions on this forum the answer is 'buy our really expensive product'.  The additional features of MapTiler cluster look great - but forcing people to buy it just to meet a size limitation that's actually quite small is very suspicious.    

[deepw...@winterroot.net]

So here's a super useful piece of information - many of the features that are not supported in 'Start' are fully supported in the old open source version of MapTiler.  You can set zoom levels, for instance, and also time an arbitrary sized tiff.

'Maptiler Start', the version in the store, appears to be a hobbled version of the old maptiler.   What a disappointment.  Is there a way for me to get a refund of the $20 purchase of MapTiler Start?   The open source tool appears to work great..  For others interested, just make sure to install wxPython correctly.

[Robert Fisch]

I don't think that you are right saying that the store version is a hobbled version of the old maptiler. I've done some testing: Rendering with the *free* store version on the same map was a lot faster than using the old version of maptiler. The produced tiles with the new version only took 1/3 of the space the tiles rendered with the old version. I think this is because the store version is based on the new core and not on the GDAL engine.

Regarding the custom zoom levels, you are right ... I'm awaiting this option as well to be hopefully present again in the next version.

[Petr Pridal]

Hello,

I am really sorry you are disappointed from the situation and I apologize for a confusion. Let me react to the mentioned points:

1) Quality of new MapTiler

There are many new features (e.g. direct tile size optimization, improved KML, Google Maps V3 viewer with overzooming and opacity slider, new raster and georeferencing formats supported, etc) and fixed rendering problems (world maps, black border around tiles, etc) in the new MapTiler.

Some features are missing yet, because the rendering core as well as the GUI has been rewritten from scratch from Python into C/C++ and from wxWidgets into QT. Now the software is packed in an installer which you get under Mac with one click and with automatic updates too. 

 

It has been technically very problematic to upgrade and maintain the old version especially on Mac OS X. The old version is hard to install and python binding to GDAL also implies penalty in speed and features, when some things are practically not possible.

The new MapTiler core is a 4x faster on a single core then older MapTiler and the speed can multiple with every CPU core added. The produced tiles are in average at least 50% smaller.

The new app represents several hundreds of hours of work of a team of people.

We still give it for free - with very light watermark and for $20 without watermark at all, if you are rendering 100 MegaPixel data.

There are many features on our roadmap - such as native support for MBTiles, upload of rendered maps directly to iPhone and iPad via iCloud, direct upload to Amazon S3 with OGC WMTS compatibility, etc. But to get there, the software must be self sustainable - otherwise we can't work on it. This means we must have a paid version and limits on the free version.

2) "I would definitely be more interested in talking with my organization about Cluster if I could actually use Start to get the simplest of tasks done."

We can send you a free demo of the Cluster version - if you think it seriously. Then you can indeed use the software with your task and prove that even complicated tasks can be done.

3) "you state definitely that there 'is no input file size' when in fact there IS a file size limitation, albeit expressed in pixels.  I think this needs to be made more clear to paying customers"

I am sorry for this confusion. My original answer in this thread has been posted an year before new MapTiler has been released. Now the post is deleted to avoid confusion between MapTiler versions.

While purchasing upgrade to Start version you must have seen and confirmed in big letters written: "No watermark on the produced map tiles. Filesize limit 10000 x 10000 px.". You could tried rendering of your files with the free version before paying anything.

With the recent upgrade to 0.3.1 - which is now more then a week in the Mac App Store these messages are even more clear, as well as the "too big" dialog is more explanatory. But it seems you have not upgraded before you wrote into this forum.

Apple does not permit returns for purchases made in the App Store unfortunately. In case the situation really botters you and you feel cheated from our side, please let me know you PayPal address and I will send you personally the money back.

4) "gdal2tiles is a free script" and "The open source (MapTiler) tool appears to work great..".

Well, I am the author of both. Go ahead and use them - I have given them to the world for free and with open-source license. You can also fix some of the issues and provide it back to other users. This would be very welcome.

Unfortunately these tools has also technical limits, mentioned above. That's why we started with the rewrite into C and now with a release of a new family of MapTiler tools. We plan to come up with an updated website too, which will speak more clearly about what MapTiler is, what it does and what versions are available. But all of this needs work which somebody must do - so it takes time.

We have several happy customers of MapTiler Cluster, and only thanks to them all of this can happen now - and thanks to these we can also work on new exciting features for the MapTiler Free and Start and on a new website. Thank you!

Kind regards,

Petr

On Thursday, April 18, 2013 8:11:11 AM UTC+2, deepw...@winterroot.net wrote:

[Jay Osborn]

Hi, Petr--

I wonder if you would consider some sort of intermediate application
(for say $50 or $100 or whatever) which would tile very large or
unlimited-size images, albeit slowly, with the rest of the features
from the current applications?

I am definitely in the 'hobbyist' category. I volunteer (unpaid) for
two non-profit family history organizations; I assemble large
historical map images from scanned archive fragments (using simple
programs: PhotoShop Express and GIMP), then tile those images and
create viewers for web browsers. Many of those maps are plain images
and not aligned to the earth:
http://maps.geshergalicia.org/
but some are, and others have very simple OpenLayers vector features also:
http://www.ckcckc.org/m/f/horn/cadmap/cmv_rohatyn_1846a.html
http://www.rsrg.org/rohatyn/map/hnl/hnl.html

Working part-time, it takes me weeks to assemble a large map, so
tiling speed is not an issue for me; I can wait a day or more if I
need to. Many of our maps are large; we have only one gigapixel map
to date, but there are others in the queue for me which I believe will
approach 50kpx a side.

Since 2011, I have been tiling maps using the open-source version of
MapTiler on an old Mac. I like MapTiler so much I donated $50 to the
project because I can't donate code, and I credit the program on the
front page of our map site (first link above) and in the comments in
every viewer file, as well as in the reference page where I describe
how we do our work. I am not a programmer (to me, a python is a
reptile I hope never to meet); I am just a tool user.

So the image size limitation in the new App is a problem for me; I
recently replaced my old Mac and bought the Start version of the new
App on the new Mac, and yes I saw the size limitation when I bought it
so I'm not asking for my money back. What I'm hoping for is an App
with no practical limitation on image size for tiling, and I'm willing
to pay for that and to wait for large maps to tile. I can't justify
the cost of the cluster version (and couldn't manage the server for it
anyway).

For now I am hanging on to my old Mac, just to run MapTiler. When the
Mac dies, I guess I'll look for another used one somewhere that hasn't
had its OS updated. From my remarks and other comments here recently,
you can probably see that there is a market for a large-image App that
doesn't need the cluster capability. Unless there is a technical
limitation to image size in the code for the App, I'd be grateful if
you'd consider this.

Thanks,
Jay

[deepw...@winterroot.net]

Hi Petr,

Ok, I start to understand the full picture better now - as a newcomer to the scene I had some misconceptions about what was going on.  First of all thanks for you awesome open source efforts and also for putting the time in to rewrite the base code in C/C++.  I think if you work to make all of what you wrote more clear to users of the various versions of MapTiler a lot of confusion will be avoided.   I see now that you are an independent developer doing his best to provide quality software and also survive financially, so this all makes sense.  

I agree with Jay - there's probably a large class of users, like us, who work for non-profits or on small scale projects that nevertheless need a tool that doesn't have pixel size limitations.  I would easily pay not $20, but $200 for such a software personally.  With nonprofit work, acquiring a $1500 piece of software is a slower process - the org needs to really know that they need it, and that it's the right thing.  I think that in the long run, Cluster is obviously the superior product and the orgs I work with are likely to adopt it, but for orgs that are just getting into map tiling they need something to work with for a while before jumping for the mainstream commercial product.  I guess that right answer here might be the open source maptiler, but it sure would be great if there was an intermediate version that didn't have all the optimizations but ran on your newer codebase.

m.

[Jimmy Utterström]

I decided to delete my previous post die to the fact that it was maybe a bit tio harsh having in mind that the software creator has overall done a great job with this as dell as previous software. Though I still really would like to se a cheaper version of maptiler that still offers unlimited image sizes.

[Klokan Petr Přidal]

Dear Jimmy (and others),

We are quite close to a release of MapTiler Plus, which is probably going to have restrictions set on one gigapixel (33.333 px x 33.333 px or similar). It should be for a price in range 100 - 150 USD and it will also support rendering into Retina tiles for recent mobile devices.

This is a reaction to what has been discussed in this thread and I hope it will still fit to the advanced hobbyists and other not-for-profit users.

Comments and feedback on this are welcome. We also seek testers for this version, until it is officially released.

Regarding the general complains about switch to the commercial version please read my answer above in this thread. I have developed and released the free and open-source utility which you can still freely reuse, improve, study... but there is a long way and improvement in quality, features and speed which we have done since then to the todays MapTiler.com.

Regards,

Petr

On Tuesday, April 15, 2014, Jimmy Utterström <jimu...@gmail.com> wrote:

I decided to delete my previous post die to the fact that it was maybe a bit tio harsh having in mind that the software creator has overall done a great job with this as dell as previous software. Though I still really would like to se a cheaper version of maptiler that still offers unlimited image sizes.

--
You received this message because you are subscribed to the Google Groups "MapTiler User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maptiler+u...@googlegroups.com.
To post to this group, send email to mapt...@googlegroups.com.

Visit this group at http://groups.google.com/group/maptiler.
For more options, visit https://groups.google.com/d/optout.

[Jay Osborn]

Petr, let me be one of the first to applaud and thank you for this
effort and the intended result. I do have images that are a bit
larger than 1Gpx to tile, but many of my files will work within your
proposed limit and I may be able to find a workaround for the really
large images. If I could vote I would request a slightly higher Gpx
limit for a slightly higher price, but I realize you have your own
issues to balance. What your 'Plus' version means for me is I may be
able to drop more than 3kg from my suitcase (the second, older Macbook
and its peripherals) and I will probably have a forward path for the
non-profit I serve using your tools; that's quite significant for me,
and very welcome.

I'm really happy with the older software in everyday use, and I really
appreciate that you listened to the in-between people in this field.
I do understand your original focus on 'serious' users; I'm just not
in that group. I wish I could contribute code or something similar to
your projects, but I'm not in the league of most of the people here.
So a solution which considers my needs without needing special
handling is huge for me.

I might be able to serve as a tester for the 'Plus' version, if you
can use feedback from (very) ordinary users. I'll follow your future
notices to see if there's some way I can help.

Děkuji,
Jay

[Jimmy Utterström]

Thank you very much Petr. The "plus" version sounds really great and I do believe that there are more people who would also appreciate a solution like that for being able to work with larger images.