Beiträge von perspixe

    i am looking for a way to let the viewer add information to a tour.
    on each pano, i see that as the viewer double clicking anywhere, for an input field to appear where information can be written and saved for all other viewers to see. infos stay attached to their pano, h and v
    before i start coding this i was curious of existing methods.
    thanks

    This is working great. on a 650M 1080p fullscreen.
    Android S7 with Opera jumpy frame rate. Will try lower demanding settings.
    Is it possible to define a mask to apply the blur?
    - only on certain areas of the pano (hv coordinates)
    - only on certain areas of the screen (circle feathered mask for example).
    That's a demanding... *wink*
    Very impressed.

    4 passes still too much for S7, 2 ok and not too visible artifacts on small screen, but on a desktop, 2 has a lot of artifacts.

    Is it possible to break immediately a loadscene with merge and keepmoving.
    Example: i have a tour with 10 scenes that load automatically every 30s but they have a merge blend of 10s.
    If i want the user to be able to load a specific scene at will i need to stop the current merge blend.

    Thank you!
    My panos are sphere images (equirectangular), not multi res, so yes I understand the reasons. A bit high res also.
    I need to keep it that way though, just uploading equirectangulars to the server with no further process.
    If i add a preview image, would this help by loading and decoding more quickly maybe, waiting for the full res to display?

    KEEPMOVING does exactly what needed but I detect a slight stop, very short, when loading next scene, in iOs an Android. On Safari Mac OS it is beautifully smooth.
    Is it to be expected or something I can optimize to have a really continuous autorotation on mobile devices?
    Thank you.

    I need to have the rotation on the pitch do the same as on yawn, continuous -inf to +inf, therefore ending half of the time upside down.
    This is not possible because positions upside down simply cannot be used (-90 to +90 only) without changing roll.
    I have the impression that computing pitch and roll in the limited range using calculations from a tweening -inf+inf variable is possible, to cover all positions, refreshing at each frame, using asyncloop.

    Anyone having some alternate thinking ideas?

    I
    Not sure if I use the right words for what things are.
    I have an object constructed that way:
    myvar[i].h
    myvar[i].v
    myvar[i].fov

    Then I want to save a temporary version of all myvar to be able to recall it later.

    To save in temp:
    set (myvarTEMP, get(myvar));

    then myvar is modified but at some point I want to recall to original myvar so I do:
    set (myvar, get(myvarTEMP));

    The saving works.
    The recall doesn't. In fact it looks like myvarTEMP is in fact a reference to myvar so all the modifications done to myvar were updated for myvarTEMP (bad way to say it as it is a reference I suppose). So when I call back the old values they were actually the mod ones.
    I need the temp not to be a reference but a completely independent variable/object.

    I tried search and find ways around this.
    It feels like some basic thing I don't understand.

    Thanks for help anyone.

    Code
    var xlml = "<krpano><include url='photospheresDisplay/presets.xml' /><view hlookat='0' vlookat='0' fov='110.0' fovmin='1.0' fovmax='179.0' fisheye='1.3' maxpixelzoom='1.0'/><image><sphere url='"+baseMediaDirectory+newfile+"'/></image></krpano>"; var krpanoThetaGD = document.getElementById("pano_ThetaGD"); krpanoThetaGD.call("loadxml(" + xlml+ ",null,KEEPALL,BLEND(0.5))");

    This is javascript code. my presets.xml define a fisheye value of 1.3 so the 360 shows as a sphere.
    The div that contains the embed has a background-color: transparent; parameter
    The original embedpano (via javascript happens on first pano image, then keeps it for other ones, until it is removed), has a wmode:"transparent" parameter. html5 only, no flash.

    So the space surrounding the sphere is transparent.
    But when I load a new sphere via the code above, I get a black flash in the background before or while the new pano image appears.
    And as the load of the new image takes about 1s, the BLEND often happens from previous image to nothing (transparent) then new image comes suddenly, unless images are already cached.

    As a side note I wonder if it possible via javascript to preload the next(s) image in advance?

    Sorry I can't give a link. This is a work in progress that I can show.

    I am trying to load in a pano a new url for the source image, from javascript.

    var krpano01 = document.getElementById("myPanoID");
    krpano01.set("image.sphere.url", new file);

    Can't get it to work... Maybe an alternative way?

    FI, The pano is created just before loading the url as
    embedpano({xml:"thepano.xml", id:"myPanoID", target:"pano_div", html5:"only", mobilescale:1.0, passQueryParameters:true, wmode:"transparent"});
    the xml file contains important things but the url of the image is "", then I try to set an url with the command above.

    A potential issue with what I am doing, embedding and removing panos often, is described in
    https://krpano.com/forum/wbb/inde…&threadID=12300
    I'll post there to see if something had been found

    I know this is out of the scope of krpano really, but maybe someone could direct me to some solution.
    I am working a lot with krpano at the moment for an art installation.
    I use 360 videos (but not only 360), with their own audio, other audio.
    I would have been interested in using audio characteristics to automatically incur movements and actions in what is happening.
    I am using html5 only, no Flash.
    What I mean by audio characteristics can be volume, frequencies etc.
    I can do this beforehand using audio software, and get a table or text list file of the information I need, and trigger krpano actions from this, reading the data from files. But a complex interactivity with potentially real time input would make an auto analysis much simpler.

    Hello

    I have some onmousedown events set up in krpano tag but from time to time I need specific actions for the next mousedown, only the next, not any other ones.
    I thought of using

    Code
    callwhen(event.mousedown(), myAcion());


    It doesn't seem to work.
    Am I missing something?

    Thanks

    Hello

    I have an array of points of view settings this way:

    Code
    <pov name="pov1" h="181" v="-20 " fovmul="1.17" fish="0.995" act="callwhen(((plugin[video].time) GT 50), setPOV('pov2'));" />
    <pov name="pov2" h="181" v="-73 " fovmul="0.83" fish="0.995" act="callwhen(((plugin[video].time) GT 60), setPOV('pov3'));" />


    Then I have an action that set the pov using the variables with

    Code
    <action name="setPOV">
         tween(view.fisheye, get(pov[%1].fish), smooth(%2,%2,%2));
         mul (thisfov, fovST, get(pov[%1].fovmul));
         lookto (get(pov[%1].h), get(pov[%1].v), get(thisfov), smooth(%2,%2,%2));
         get(pov[%1].act);
     </action>

    It all works fine but for the act value. I want to put there as string a serie of actions to execute; the example in the code is a single callwhen but potentially more things, for an automated virtual "tour".

    How do I execute as krpano xml code the content of a variable as string?

    oh well, I feel very stupid. Let's try to make it work...

    I declare in a startup action:

    Code
    set (thisPOVe, 0);


    then in an action dealing with onkeydown events I have:

    Code
    if (keycode == 16, switch (thisPOVe,1,2,3,4,5,6,7); trace (get(thisPOVe)););

    So every time the shift key is pressed, a new value from the switch action list is traced.
    In the log I get nothing else than original declared value 0.

    Missing something I guess...