FLIR One Developer / Blog / Topics / flagship, glitch, incompatible, s7, s7 edge, samsung, slow, work around / Samsung S7 – SLOW, and glitchy stock app.. Why?! (also a kind-of work around)

Samsung S7 – SLOW, and glitchy stock app.. Why?! (also a kind-of work around)

Home Forums FLIR developer community General discussion Samsung S7 – SLOW, and glitchy stock app.. Why?! (also a kind-of work around)

This topic contains 7 replies, has 4 voices, and was last updated by Profile photo of Sarah Sarah 4 years, 1 month ago.

Viewing 8 posts - 1 through 8 (of 8 total)

  • Author


  • #3816

    Profile photo of Sarah


    Does anyone know WHY the flagship Samsung S7 doesn’t get 8 FPS? Is there some requirement of the USB bus that’s not met, it can’t be processing speed.

    Also – the images are glitchy – bars of heat appear both horizontally and vertically.

    It’s a USB interface streaming bytes!….. why is there so much incompatibility with Android phones? Webcams, scanners, speakers, TV’s all do this – why can’t FLIR One?

    What’s nearly as bad is that I’ve emailed, FLIR and had no response all week.


    I’ve made a tweak to the FLIR One example program.

    The reason I’ve tweaked it is because ALL the phones I have (including the Samsung new S7) don’t update the display very fast at all.

    I discovered not why they’re slow, but I DID find out that the “Thermal Linear Flux 14bit” (Thermal grey-scale) in the demo updates the screen MUCH faster than the official Google Play program does.

    You’re going to say “But in that mode, it doesn’t take lovely MSX Pictures!”
    Well actually it DOES!

    My “big idea”, is to preview the FLIR One heatmap using the “Thermal Linear Flux 14bit” image mode, but take the MSX image when you press “Take photo!”

    In detail – when you hit “Take photo”, while you are in “Thermal Linear Flux 14bit” mode, it takes THREE images:

    1: FULL resolution grey-scale image in PNG.
    2: Cropped MSX image. (Fully compatible with FLIR Tools temperature analysis tool)
    3: RAW thermal output. (You can’t view this unless you process the data first)

    Cropped image? What am I on about!?

    Well, to align the normal web-cam image with the thermal image, the FLIR-One internally crops 10 or 20 pixels around the edges of the image, it does this because the engineers in the factory position the two cameras VERTICALLY by cropping parts.

    Obviously the horizontal offset of the visual image and heat image can be altered while you’re using the FLIR-One app.

    Full resolution kitty and her food!

    Full resolution toilet!

    Check these comparisons out, look how much is cropped!:
    MSX photo, and FULL RESOLUTION greyscale images:

    Download here:
    (It doesn’t need any access to your phone details.)

    One little glitch I haven’t been able to fix… You need to press “Change view” if you want the MSX images to show a proper preview in image programs. If you forget, the image will just look grey. But! It’s not wasted, you just have to load the image in FLIR Tools, and assign a correct heat map range.

    Press “Change view” (bottom left), and then press “Thermal Image” to get the SLOW MSX image update for a few seconds. You need to do this so that it calibrates the heat map correctly.

    Then go back to “Change view”, and put it back to “Thermal Linear Flux 14bit” so it takes the three images at once. (Big greyscale, cropped MSX image, and RAW image)


    Profile photo of Thomas


    there is not only the awful cropfactor in non MSX images,
    Flir add also a fake noise to the images.
    This shifts the NEDT from 50mK (Lepton 3) to about 200mK (FlirOneApp)

    compare this images, and you see what I mean

    As you see above the “Thermal Linear Flux 14bit” loaded in FlirTools for PC
    You can convert a saved 16 Bit(!) PNG with this php script to a real Flir thermal image:

    If you have images with larger temperatur spans, you don’t see the fake noise artefacts.
    compare with this image galery (Flir One G1 80×60, Flir One G2 160×120, Flir E4, Flir E40)


    Profile photo of Cal Loo
    Cal Loo


    Hi Sarah and Thomas,

    Apologies for the delayed response on this, I’ve forwarded your post to our technical support team and I expect you’ll receive a reply very soon.


    Cal Loo
    FLIR Developer Relations Manager


    Profile photo of Sarah


    Thank you Cal.


    Profile photo of Cal Loo
    Cal Loo


    Hi Sarah,

    Our engineers are aware of and working on a list of performance issues with the S7. We expect to have new Android app releases over the next couple of weeks that address all of the known issues.

    Thanks again for bringing this issue to us,

    Cal Loo
    FLIR Developer Relations


    Profile photo of Benjamin Hutchinson
    Benjamin Hutchinson


    Maybe it’s just the subject which I was testing it on, but the 14bit raw image (which contrast stretched to fill the black-to-white range 8bit range for display) actually looks noisier than the proprietary processed image that’s taken by the normal FLIR app. I tend to think that rather than adding noise, FLIR’s processing actually uses some kind of noise-reduction. Though this noise-reduction algorithm may introduce other artifacts (which I’ve seen under certain conditions that resemble JPEG artifacts even in the semi-processed radiometric image embedded in the JPEG file, not just the JPEG preview image), I don’t think any noise has been added by FLIR’s processing.


    Profile photo of Cal Loo
    Cal Loo


    Hi Sarah,

    Tying into our offline convo, I’m glad to hear the new Android app is working with your S7! We’re aware of the other issue you raised and are working on it, stay tuned 🙂



    Profile photo of Sarah


    Yes, it’s working so smoothly now. 100% better. Thanks Cal!

Viewing 8 posts - 1 through 8 (of 8 total)

You must be logged in to reply to this topic.