Posts by LMB

    Hi Klaus,
    Has anything changed with regard to the set expiry date (with the krpanotools) with version 1.20? Can we read it or do we still need to use the workaround as described in the thread?
    Kind regards


    thanks for responding, but it's not the size or traffic which causes the problem. It's the path, it doesn't allow either 'file:////<DRIVE>:\\path\path..' or 'http:....' I think....

    But I solved it for the moment by creating a Symbolic Link (MKLINK under Windows) to the external HDD. That way it looks like the XML and images are located locally and you can use a relative link. The disadvantage here is that you need to instruct the customer how to create the Symbolic Link, but that's all.
    Then again, I am using a path as variable (queryline) and would like to be able to point to external drives and http, instead of to this local link. I might have to look into CORS and CrossDomain for that I reckon?

    My customer wants to load the images/tiles from a NAS in his domain, which is not connected to the internet.

    I am testing with localserver, with a generic HTML on one drive (C:\) and loading the XMLs with all design stuff and panos from another drive (e.g. E:\). I can manage to load all BUT the pano-tiles, whatever path I throw at it. I have tried directly with file:///E:/panos/preview.jpg (short example),
    I have also tried using 2 localservers, making the path to the tiles like: (which IS the correct path), etc, but no luck.
    The viewer (showlog) mentions
    Error: loading of '....<path>...preview.jpg' failed! (and same error for the other tiles of course)

    I had this before with the paths to the hotspots etc, but I sorted those with relative paths (skin and js files are available on both drives).

    so, it loads correctly the XML from the other drive, and the content from that XML, APART from the tiles.

    just to make the story complete, I have added <security cors="use-credentials" /> to the XML.

    any suggestions?


    Just noticed that the latest Chrome (54.0.2840.71 m) is not supporting my viewer in WebGL anymore. I noticed this on one desktop which already had the latest version, checked another desktop, still on Chrome 53 and there it worked. Then it automatically updated to Chrome 54 and it stopped working.

    Tried the latest PR8 krpano version, but no luck.

    CSS3D is working OK, but I need WebGL for some plugins. What to do?

    Hi Klaus,
    First of all a BIG thank you as usual!
    And even more so, I just had the sharpening working this afternoon! I already new how to apply all kinds of additions to a copied new canvas, but now with your example it is totally complete. Thank you very much!

    I added a slider to adjust values for sharpening, now I will add sliders to adjust Gamma, Brightness and Contrast. It's gonna be nice!

    I was just researching how to properly refresh the pano dynamically after a change from the slider. With the sharpening demo, it refreshes dynamically to about 25% (so sharpening value increase from 0 to about 25%), after that it only refreshes the minute I rotate/move the pano. Could that be the size of the bufffer or something? (in that case I can just create smaller steps of course).



    I have searched quite a long time, but feel myself getting more stupid by the minute.
    I am trying to adjust HTML5-WebGL panos for brightness/contrast/gamma. I am perfectly able to achieve this on a new canvas (appendchild, or <canvas />), copying current pano, applying filters and showing them in the new canvas, but would like to apply those to the current pano of course.
    Every time I think I am on the right track, with terms like 'pingpong rendering', 'swapping', even using three.js, I get lost again.
    There are loads of examples out there, applying filters to a stored image or video, but how would I do this dynamically to the actual (pano)canvas?

    I found the thread about WebGL createPostProcessingShader where Klaus mentions (and gives hints about) the upcoming sharpening filter, would that be the way to go? And is there a working example available?
    WebGL createPostProcessingShader returns null


    Tested again, still not working. You were right about the domain restriction though (which was very weird of course).

    No problem in the following cases:
    * If I use PR3
    * If I use Firefox or Chrome on PR4
    * If I watch Edge on a Linux Server on PR4 (and PR3 of course)
    * If I set HTML5 to 'always'

    Not working in Edge in the following case:
    * Windows Server 2012 R2, with Edge, with PR4, with HTML5=prefer or anything else than always

    So for some reason it detects Edge as non-compliant with WebGL, and if you force it to be (=always) it DOES work.
    But PR3 does not have this problem.

    I can live with setting it to always, but you said somewhere to only use it in development setup.
    If you need to investigate, I can send you the link, but privately please as it is still a testing environment.

    BTW, found another thread, where you ask to have a look at:…xture-npot.html…-good-npot.html

    The first one passes, the second one doesn't play. So it must be my GPU. Fine enough, I have seen other problems, but why would the same situation work on a Linux server?



    Up till now we were faithfully using the beautiful multires tool, but with a large amount of panos, we are trying to lower the amount of storage. So, I started testing using Cube and Sphere, and much to my surprise got very good results with the Sphere. It seems to be better quality and working quite well on desktops.
    I was expecting a result like this:…ts/basecube.xml
    But for us the sphere is actually better if you zoom in deep enough ...

    We are only working with WebGL, so would there be any reason not to use sphere apart from the longer time you see the black background (you will only see the pano once it is totally loaded)?
    We are using panosizes of 8000x4000, (which would be a cubesize of 2546 as far as I can tell), and they seem to be loading reasonably fast...

    I would still be working with Cubes, if I would not have seen a (slightly) sharper and better picture from the Sphere... How come?


    Hi Klaus,

    With the new PR4, I find that the textfield is not selectable anymore.
    With PR3, and variable selectable="true", I can select the text

    Without changing the XML:
    * if I use the PR4 http://krpano.JS/SWF, and leave the old plugins: text not selectable (handcursor)
    * if I use the PR3 http://krpano.JS/SWF, and change the textfield.swf plugin to the one of PR4: text not selectable

    are there any new settings I should be aware of?

    <layer name="xmask" type="container" align="lefttop" x="10" y="30" width="300" height="284" bgcolor="0xFFFFFF" bgalpha="0.8" maskchildren="true">
    <plugin name="xmasktext" keep="true" url="%SWFPATH%/plugins/textfield.swf" html="tmptxt" css="font-family:Tahoma; font-size:13px;" border="false" background="false" selectable="true" align="lefttop" y="2" width="100%" height="100%"/>

    Hi, I am back again:

    Only on a Windows Server 2012 R2 host, viewing with Windows 10 Edge, I get the "ERROR: Adobe Flashplayer or HTML5 Browser with WebGL support required!" in the following cases:
    - domain limited krpano (if I remove domain limit, it works, but WITH the domain limit it works on all other browsers but this Edge)
    - html5= anything other than 'always' (so forcing html5 does work)

    And if I watch this same project on a Debian/Linux server host, it also works, even in the cases mentioned above.

    Is this a known problem? What can I do about it, other than forcing HTML5?

    BTW, I see quite a lot of similar problems with HTML5 and Edge, but I haven't seen anything with the domainlimit restriction and I also find it very strange that it DOES work if the Host server is a Linux server, even with domainlimits and html5=prefer...



    I would like to show the imagesize of the original equirectangular in a textfield.
    I know how to calculate it from the tilesizes in a multires setup, but how would I calculate it from the cubesize, if I use 'NORMAL' panos?

    This works fine for multires panos:

    sub(largestlevelindex, image.level.count, 1);
    copy(cubewidth, image.level[get(largestlevelindex)].tiledimagewidth);
    mul(spherewidth, cubewidth, Math.PI);

    Can it be done for non-multires as well?
    (if not, I will just add a variable to the <scene...>)


    Like measuring, overlaying, etc.
    I create at least 15 hotspots per pano, and a lot more on the maps. Repositioning 15 x 20.000+ hotspots is of course not very enjoyable *smile*

    In any case, whichever way you calculate, I started this thread, because I wanted to know how the PREALIGN works and if it could make life easier. I think I know now, although it is still strange, so if Klaus could confirm it would set my mind at ease..