AZMAN42 posted these photos in a thread on Look at the top photo carefully. Do you notice something odd about it? The glow of the white phosphor tube from the NODs is shining on the shooter’s face. The people in the background have a soft orange glow on one side which I assume is coming from a fire or other source of light. And yet the image is bright and you can clearly see everything. Now what if I told you this photo was shot at night? Yes this was photographed at night. Given the amount of light, it looks like the moon was out and the sky was cloudless.

Here is another photo taken during the same time.


What sorcery is this? You could call it night vision, but it is not image intensifying, rather it is the ridiculously high ISO sensitivity of the Sony A7SII camera. As far as I know this was not your stereotypical long exposure shot.

Here is a photo I took over a decade ago with a Sony point and shoot camera that had some limited manual controls. The photo was taken at midnight on a full moon. It was a very long exposure but it almost looks like daylight.

If the Sony A7SII has such great night vision capabilities, why dont we see that technology in night vision goggles?


  • ActionPhysicalMan

    Uh, I don’t get the point. I love the pictures though. Thanks:-)

    • JD

      I think what the author is trying to get across is that all the pictures in the article were taken in darkness, but yet it looks like daylight. So why are night vision devices not producing imagery like this? Instead of the green or grey that is typically associated with nvg.

  • M C

    Because night vision goggles aren’t much use if they need to take a couple of hundred shots over the course of several hours and then spend a few more hours processing the raw images before showing you the one percent of images that look good.

    • Johnny G

      Now I have something else to put on my “buy before I get married” list. Thanks TFB…

      • Roper1911

        NVG’s and Thermal scopes have really dropped in price recently. you can pickup simple monoculars for $150-300. no need to shell out $1500

        • Twilight sparkle

          They’ve been that price or cheaper for a long time, you just aren’t going to get anything very good at that price. May as well have a flashlight that cost the same amount of money

        • Mike

          Maybe he meant the Christmas sweater

  • Johnny G

    I would say it’s because of the exposure time. For static surveillance, a good SLR and lens might make a good area coverage device. But for mounting on the head? No way a long exposure would work. Cheaper solutions, like in camcorders with night mode, add unacceptable motion blur at high light amplification without a strong source of illumination.

  • Jeff

    As a photographer/videographer. The Sony A7s2 has taken the market for high ISO capability by storm… be amazed… puts our 3500 dollar camera and thousands of dollars worth of lenses to shame lol

  • Paul Rain

    You just to put your head on a tripod for a few seconds. No big deal.

  • Uniform223

    Looking “operator” and wearing a Christmas sweater… AWESOME! 😀

    • Roper1911


    • DangerousClown

      Christmas sweater? I hope you’re not referring to the guy on the end. If so, you’ve lived a pretty sheltered life. That’s clearly a Senor Lopez.

      • Twilight sparkle

        He’s talking about the guy just next to him, look at the top pic you can see snowmen and trees

        • DangerousClown

          Thanks. I had to view it on my computer to see that. Originally, I read the article on my phone.

    • KidCorporate

      When Christmas is under attack, it fights back.

  • Noah

    I work in the film industry with sensors on par with that every day, and that technology would get operators killed due to latency. When you look through the viewfinder you don’t see that image, you see dark, after the picture is taken and processed then appears as the image you see. Even if you had a super computer strapped to your head you couldn’t see that image in real time, not because of exposure time but processing time. Also gen 3 NV can tolerate flashes of bright light. Super high ISO or ASA would over expose with something like gunfire because the dynamic range is very limited.

    • digitalindie

      That’s not true about the latency at all. I’ve got the first generation A7s, and this is exactly what you see on the LCD or viewfinder when you shoot at extremely high ISO. There is some lag, but it’s definitely under 100 milliseconds – I’ve read the latest generation is down into the 20-30ms range. You’re correct though that the dynamic range issue is the real problem, any bright light will white out the image.

      • Steve W

        Latency is always an issue, how much of an issue depends. Dynamic range carries it’s own latency issues, and is also an issue. Then, you have the size of the lens gathering what light there is. Chances are it’s at least a 3-4 pound F2.8 with a 77mm (3 inches) or an equally or heavier F1.0-F1.4 lens with a larger 82mm objective. And then there’s the power requirements. Combine the latency of the electronics, the latency as the dynamic range responds to flashes and flares, size, weight, power requirements, and it becomes obvious you won’t be using them with the same utility as any currently available decent NVD’s.

        That doesn’t mean such technology doesn’t have it’s uses. A pod mounted zoom can make certain observation duties almost luxurious and at a cost way below an equally magnified NVD.

        The trick is in understanding the differences in technologies and their limitations and applying them to your needs.

        • digitalindie

          It’s not necessarily that heavy though. My A7s with battery, lens mount adapter, and a 35 year old 50mm f/1.4 lens weighs 2lbs. That lens is a brick though – modern equivalents weigh significantly less. Eliminate the rear LCD, streamline the body, move the battery to an external pack, and it should be entirely possible to produce a sub-1lb night vision system from the same basic components. Latency should be getting close with the latest generation systems, and there are various solutions for dynamic range issues available or in development with other camera systems. My guess would be that this kind of tech is currently at the point where a workable, high-res true color night vision system would be entirely feasible from a technology standpoint – it’s just likely not competitive from a cost standpoint for general usage. With the pace at which the tech has improved over the past few years though I’d guess it’s not too many years away from becoming practical.

      • M C

        The designers of the Oculus Rift VR headset considered a latency of 20ms to be the maximum acceptable – any more and the user feels disconnected and reaction times suffer. That’s one of the reasons why they used OLED screens instead of LCD – the latency between the top and bottom of high res LCD screens can be as much as 15ms.

        Getting shot with virtual bullets because of lag is embarrassing enough. I’m told that real bullets also hurt a lot more.

        BTW, if anyone has Sony a7s2, gen 3 or newer NVG and an Oculus Rift headset they can send me, I’d be glad to evaluate their respective lagginess.

    • crwth

      Overexposing? How would overexposing a few pixels on the sensor affect the rest of the image? (Hint: it doesn’t). Portrait photography with the sun in the background is common these days using CMOS sensors, and the dynamic range in these situations is orders of magnitude greater than gunfire flash. Post-processing is unnecessary- auto-metering exposure to median intensity is tech that’s been around forever. Image intensifier tubes on the other hand can get physically damaged by bright lights- even flashlights, hence the necessity for protective technology (e.g. auto-shutoff). CMOS sensors do not have this problem, short of being melted by long-exposure magnified sunlight, as some unwitting eclipse photographers found out recently.

      Also processing time, as mentioned by digitalindie is in the millisecond range, and could be easily improved with a faster GPU with better heatsinking- technology that exists today, in form factors smaller than credit cards. (It just isn’t done because that kind of latency isn’t at all important in a consumer video camera, esp. at the given price point.) I’m usually not into ad hominem attacks, but please don’t use your “work in the film industry” as a pretense for knowledge when you don’t understand the technology. For all we know you could be a director, prop master, or guy holding the boom-mic. Or maybe you are a camera guy, but have no experience with modern low-light technology.

      • digitalindie

        Yeah, I was a little surprised at that comment – what he said about the cameras producing a dark image is true as far as the recorded data is concerned. But that’s not what you see in the viewfinder – most professional cameras, viewfinders and external monitors have the ability to apply a LUT that converts the flat, dark, low contrast image into something much closer to the finished image in real time.

        As far as the overexposure issue though, it’s not a concern in terms of damage, more just in terms of utility. If you boost the exposure enough to see in the dark then even small amounts of light can wash out large portions of the image. Light a match and the flame will become a giant white ball; any reflective surface within 20-30 feet will wash out to nearly white. Turning on even a small flashlight would be even worse. Something like a muzzle flash would likely make the entire image go white briefly. It doesn’t damage anything, just makes it much harder to see what’s going on. There’s technology in the pipeline that can deal with that, but as I mentioned in my other comment above it’s probably still at the point where it’s not yet cost effective.

  • DangerousClown

    Forget night vision. I’d be happy if Canon offered better native high ISO performance like this.

  • Peter Nissen

    I’ve picked up from Aliexpress the P1-0540 5×40 night vision monocular and it has night brightness enhancement (passive night vision) and IR emitter – all for $115 delivered to Aust and whilst not “Operator” grade, for my uses – just amazing!

  • Mark Horning

    Because a Sony camera can’t take that kind of image at 250 times a second, unlike an image intensifier tube.

  • Tom

    Is there latency issues? As others have stated, yes. Is there dynamic range issues as others have mentioned? Yes. Instead of immediately shooting down a new technology as non-viable maybe we should be looking at this as a technological step in the right direction.

    Look at every piece of technology we have now and understand that it is just the better, leaner, more streamlined version of technology that came before. It has benefited from millions of user hours to help determine the best configuration. What works and what doesn’t. What needs to be improved. Back in Vietnam the only body armor available to the soldier was a flak jacket incapable of stopping bullets and a helmet. Now we have full out body armor capable of stopping most modern calibers and lightweight ballistic helmet technology. Things improve. The technology improves but more importantly the lessons learned from the past are incorporated into the newer tech.

    It’s the same thing with these cameras. Right now it might be only available in a larger platform. The latency is greater and maybe the dynamic range is an issue. However, they produced a camera. Maybe the tech isn’t viable as an NVG in it’s current form. So they didn’t produce an NVG but only the specs needed for it to be used on a camera. That camera will get picked up by artists, photographers, and film makers and used in unique and new ways the creators wouldn’t think about. Those usages will show where the tech excels and where it lacks. The next generation of the tech will improve upon said disadvantages.

    The thing is, the capability to make this a viable NVG is likely there. However things like this require a lot of user testing and viability testing. The cost to make the latency less might be incredibly expensive to miniaturize into the platform when it’s shown to not be necessary for a camera. The dynamic range programming and capabilities to handle bright flashes might have been found to be unneeded in it’s current form and if included could increase the cost of the platform given it’s target market.

    Give the technology some time. This is why I refuse to be an early adopter of new technology. You are the test monkey. You get a basic platform devoid of any additional features. You are using a platform laid out by designers in their best guess of how they feel it will be used by the masses. They gain feedback from you to improve the second generation of their product. Then that version 2 comes out and you are stuck with a choice. Keep the gen 1 which costs a lot of money. Or upgrade to the gen 2 which has all of the annoying little things your gen 1 had figured out.

    Once this camera is used in applications where the latency or dynamic range prove to be a real hindrance you will start to see it fixed in the newer versions. Those guys strapping this camera to a UAV and flying around the beach at night or following extreme sports, the TV stations picking up this tech for use at professional games or productions to get low light situations, amateur shooters using it in amateur comps, and much more. The market opens up with the producer seeing what the consumer demands. Once the demand is there and the market is there it becomes more justifiable to invest in improving the tech.

    Your first car didn’t have all the bells and whistles you have in your car now. The companies have improved, introduced new tech, improved upon that, and innovated. Excited to see what comes from this in the next few years.