Posts by nicoclark

    I have a site with video alpha channels and with the most recent version of chrome and edge it will not load if alphachannel is set to SBS. A similar problem happens with the example on the KRPano website but reloading it sometimes fixes the problem and sometimes not. There's no issue with Firefox

    This is the error that I see in the console. Does anybody have any idea what might be going wrong?

    Uncaught TypeError: Failed to execute 'uniform4fv' on 'WebGLRenderingContext': The provided float value is non-finite.
    at ac (eval at r (krpano1.21.2.js:6:368), <anonymous>:3:285179)
    at U (eval at r (krpano1.21.2.js:6:368), <anonymous>:3:268397)
    at ia.renderpano (eval at r (krpano1.21.2.js:6:368), <anonymous>:3:319099)
    at Q.renderview (eval at r (krpano1.21.2.js:6:368), <anonymous>:3:190180)
    at of (eval at r (krpano1.21.2.js:6:368), <anonymous>:3:41822)
    at b (eval at r (krpano1.21.2.js:6:368), <anonymous>:3:433975)

    This browser comes with the Pico Neo 3 and 4. The problem is that KRPano does not recognise the Pico browser as mobileVR. The Pico 4 enterprise is the most viable B2B headset solution for KRPano especially over here in Germany. Oculus is not even sold here because the government has an argument with Meta because it requires every owner to set up a Meta account.

    The workaround on the Pico Browser is to wait for the WebVR_onavailable and check if the webVR.device name is webXR. The shame about that is that information is not available when the XML is loaded so you have to delay until the event happens.

    In theory one could do some JavaScript to check the user agent but it would be cool if the KRPano device settings could recognise this browser straight away. I don't know if anybody else works with this browser - it would be nice to compare notes.

    There may be an issue with KRPano recognising WebXR capability in Wolvic. I've been in touch with Igalia directly but they believe it's a problem on the KRPano side.

    As an example this Krpano example VR site has reversed head tracking.
    https://krpano.com/releases/1.20.…tml5=webgl+only


    https://www.tourdeforce360.com/test/webvr/index.html
    When you run this test site in Wolvic on the Oculus Quest 2 it indicates that Wolvic has neither WebXR or WebVR capability but is a VR Browser. (in Oculus browser it shows both as available)

    It is possible to force webVR in Wolvic and then it works but by default Wolvic want to allow only webXR.
    Perhaps the device detection in KRpano needs to be adapted for Wolvic.

    That makes sense. I wouldn't have guessed the result would be a string.
    Very useful function in the end. The amount of times I'm creating a variable just so it can be rounded. Now I can just round it within the calc function. Woohoo.

    I guess I've often wondered if roundval actually speeds anything up. I do it because the original values have 10 or so decimal places so I thought the calculations were creating really long numbers which would slow things down. Maybe the rondval is even slower than just working with the decimals.

    Hi,

    Trying to use the new inline functions but they seem to be buggy.

    This returns -75.00
    calc(FOVnow,roundval(view.fov,2));
    calc(var.Result,0 - FOVnow);

    This Returns 075.00
    calc(var.Result,0+roundval(view.fov,2));

    This Returns NaN
    calc(var.Result,0 - roundval(view.fov,2));

    If you go onto the headset and click on the viveport app you can sign up (don't need to pay) and then download Firefox Reality which is the only proper WebVR Browser for Pico. It automatically goes into VR Mode.
    It's important to be aware that the thumb pad of the controller acts like a sort of scroll. When you touch it during the depth map VR tours it's seemingly out of control and you can get nauseous pretty quickly.
    Anyway - have fun.

    If you check this link on various browsers you will see that on Google Chrome on Windows the star disappears. It seems, if you make a layer the child of distorted hotspot, Windows Chrome cannot handle an alpha between 1 and 0. All the other browsers (including OSX Chrome) on all the platforms I have access to, work fine.

    Does anybody know a work around which doesn't involve using an alpha of 1?

    here's the code

    addhotspot(Test);
    set(hotspot[Test].url,%SWFPATH%/Dot.jpg);
    set(hotspot[Test].distorted,1);
    set(hotspot[Test].ath,0);
    set(hotspot[Test].atv,0);
    set(hotspot[Test].ry,45);
    set(hotspot[Test].rx,0);
    set(hotspot[Test].rz,0);
    set(hotspot[Test].alpha,.5);
    addlayer(Test);
    set(layer[Test].url,%SWFPATH%/Star_.png);
    set(layer[Test].parent,hotspot[Test]);

    I'm trying to block the log window appearing in Safari. I've used showerrors(false) but still the log window appears although no error is listed.
    A workaround would be not check if the log is open and then close it. Is there a readable variable for showlog(true)?

    Sorry Klaus!
    I really am not knowledgeable about what is possible but with HTML5 but one would expect to recognise a link in text in any browser. The flash version of Krapno recognised a:hover but that doesn't function in the HTML 5 version of Krpano.

    I asked about this back in March you said "krpano doesn't include a full featured css/html parser."

    Am I right in understanding from your last post that this lack of CSS function like a:hover has nothing to do with with the HTML 5 version of KRpano and that this will never be possible and there is no work around?

    Hi Klaus,

    Many congratulations on the new version of the software. I'm so glad that I started working with KrPano because you have never stopped developing and improving it.

    I noticed the webVR cursor that you created and realised that that sort of cursor is what I've been really looking for since going into mobile devices. The ability to hover over hotspots without the use of a mouse and communicate something about their function before actually using them gives back a lot of functionality that has been lost in the transition to mobile devices.

    What you suggested here is perfect to use on a smart phone unfortunately the code doesn't seem to be transfer between platforms. If I test it on the desktop the trace comes back but if on the iPhone simulator there is no trace coming back. I'm sure there is something straightforward to address here but it would be so helpful if you could doublecheck it and see where the problem lies.

    Many thanks,

    Nico