Category: Aerial Landscape Photography

  • DJI Mavic 4 Pro RAW Image Quality for Landscape Photography

    DJI Mavic 4 Pro RAW Image Quality for Landscape Photography

    The DJI Mavic 4 Pro (image courtesy of DJI.com)

    DJI Mavic 4 Pro Image Quality for Landscape Photographers

    Aerial landscape photographers using drones have come to accept that achieving optimal image quality is an art in itself. DJI make the best drones for photographers, but as recently as a few years ago, the image quality was suitable for web use but would not have been deemed suitable for larger prints.

    Improving DJI drone image quality

    My first drone was a DJI Air 2 which took 12 megapixel RAW files with a small sensor and a wide angle lens. I then progressed to an Air 2s which had a 20 megapixel 1 inch sensor. My next upgrade was to a DJI Mavic 3 Pro which had a wide angle camera with a 4/3 sensor, and a 3x (12/48 megapixel) and 7x camera (12 megapixel).

    The difference in image quality between the Air 2 and Air 2s was huge. The DJI Air 2 felt like a flying cell phone camera, with noisy, soft images and distinctly average colours. As someone who was used to shooting on a full frame Canon R camera I found myself excited by the aerial perspective, but firmly grounded in terms image quality.

    However, the Air 2s with its larger sensor was somewhat of a revelation. No longer did I feel like I was flying a cell phone camera, but rather I was flying a very capable imaging system. I estimated to be similar in quality to an older generation DLR with a basic kit lens. By that I mean it captured nice colours and clear images, but there was a lot of digital noise and the edges and corners of the images were a little soft.

    My current drone is a DJI Mavic 3 Pro. This has a larger 4/3 sensor than my Air 2s. Whilst the image quality was significantly better than the Air 2s, it wasn’t a night and day difference in the way the jump from the Air2 > Airs 2s had been.

    Aerial landscape photograph taken with the DJI Mavic 4 Pro, capturing intricate ice sculptures and a giant Ferris wheel at sunrise in a snow-covered winter theme park — showcasing exceptional dynamic range and colour depth.
    DJI Mavic 4 Pro Sample 28mm (1x) RAW Photograph Credit: DJI.com

    DJI Drone Megapixels: Advertised vs Native Resolution

    When I bought my DJI Mavic 3 Pro it was advertised as having a 48MP 3x camera and a 12MP 7x camera. It should be noted when buying a DJI drone that quite often the advertised resolution is often as a result of some technical wizardry. For example, in the case of my 48 megapixel 3x camera, this is actually a 12 megapixel native-resolution sensor which has the ability to produce 48 megapixel files.

    Under perfect conditions in a landscape evenly lit by strong sun, the 48 megapixel files are quite good. However, if you have any areas of shadow at all or anything but the bright direct light, you are better to capture native resolution images at 12 megapixels. The larger images are full of digital noise and the colours and dynamic range are not as good. If you try and brighten shadow areas you will have a message image.

    With DJI drones, you will be capturing RAW files at the native resolution in 90% of circumstances.

    A good way to think about it is you are able to choose between a high quality native resolution, or a lower quality higher resolution file.

    Mavic 4 Pro Image Quality

    The newly released DJI Mavic 4 Pro makes some big claims. The first is that the wide angle camera captures 100 megapixel images. However, this is actually a 25 megapixel sensor with the ability to produce 100 megapixel files. On the DJI website they have made RAW file samples available and these are all 25 megapixels.

    I think it is safe to assume that as I mention above, the larger files will only be suitable for use under extremely optimal conditions. This is a drone with a 25 megapixel main sensor – still an improvement on the Mavic 3 Pro’s 20 megapixel sensor.

    Similarly, the 3x camera is a 12 megapixel native size with 48 megapixel advertised, and the 7x camera is a 12 megapixel native size with a 50 megapixel advertised resolution.

    Mavic 4 Pro CameraAdvertised ResolutionNative Resolution
    1x / 28mm / wide angle100 megapixels25 megapixels
    3x / 70mm / medium tele48 megapixels12.1 megapixels
    7x / 168mm / tele camera50 megapixels12.5 megapixels

    As you can see, the advertised resolutions are 4x bigger than the native file size, because each sensor is using

  • Using ChatGPT to critically evaluate your aerial landscape photography

    Using ChatGPT to critically evaluate your aerial landscape photography

    DJI Mavic 3 Pro aerial landscape photograph of Chapel Vitaleta in Val d’Orcia, Tuscany, surrounded by cypress trees and golden hour fields

    The problem with landscape photography critique

    One of the most challenging aspects of modern landscape photography is obtaining honest, constructive feedback on your work. Unless you are regularly submitting images to competitions, paying for portfolio reviews, or have a trusted network of peers, developing through unbiased feedback is not an avenue open to all.

    Most photographers primarily share their work on social media. This compounds the problem as most of the critique found there, whilst well meaning, is somewhat superficial and limited to positive feedback only.

    Why use ChatGPT for landscape photography critique?

    ChatGPT has been trained on vast amounts of visual and artistic knowledge, including photography principles. When used carefully and strategically, it can simulate the feedback of an experienced photographer or competition judge.

    Whilst you should not allow ChatGPT to take precedence over your own judgement and artistic vision, I think it is a useful, cutting edge way to improve your photography.

    ChatGPT prompts for landscape photography critique

    For those of you not familiar with the use of tools like ChatGPT, you must instruct it in the form of a ‘prompt’ to obtain your desired output.

    You can give very basic prompts, but you will have more useful results if you take a little bit of time and give a more detailed prompt:

    Basic prompt: ‘Identify 3 main areas of improvement for this landscape photograph.’

    Advanced prompt: ‘I want direct, unbiased critique of this landscape photograph which was captured using a DJI Mavic 3 Pro drone with its 7x lens in Val d’Orcia, Italy. Identify 3 main areas of improvement for the photograph.’

    You can see that in the advanced prompt I have entered both conditions (I want direct, unbiased critique of this landscape photograph which was captured using a DJI Mavic 3 Pro drone with its 7x lens in Val d’Orcia, Italy) and instructions (identify 3 main areas of improvement for the photograph).

    Prompt Conditions + Prompt Instructions = Best ChatGPT Results

    An example of the ChatGPT critique

    I uploaded this aerial landscape photograph, taken in Tuscany, Italy, to ChatGPT along with the advanced prompt example given above.

    ChatGPT critique:

    1. Haze Management & Atmospheric Clarity

    • Observation: The image has strong atmospheric haze, especially in the background, which reduces contrast and detail in the distant hills.
    • Improvement: Apply selective dehazing in post-processing (e.g., Lightroom’s Dehaze or using luminosity masks in Photoshop) to bring out subtle textures in the background while preserving the dreamy quality in the foreground mist. A mild clarity boost in specific midtones could help without ruining the soft morning mood.

    2. Color Temperature & White Balance Uniformity

    • Observation: The warm morning light is effective, but the yellow/orange glow in the mist may be too saturated, creating a slightly unnatural look—especially around the chapel and house area. The rest of the scene has a cooler green palette, creating a slight color imbalance.
    • Improvement: Consider local adjustments to harmonize the warmth between the foreground mist and the cooler greens. Toning down the orange hue in the fog area can lead to a more balanced, realistic tonality.

    3. Subject Emphasis and Framing Tightness

    • Observation: The small chapel and farmhouse are compelling subjects but are positioned quite low in the frame. The expansive upper half of the image, while beautiful, lacks visual anchors and feels a bit top-heavy compositionally.
    • Improvement: Either crop a portion of the sky/horizon or adjust future drone captures to elevate the subject’s vertical position slightly. This helps balance the composition and places more emphasis on the focal point.
    DJI Mavic 3 Pro aerial landscape photograph of Chapel Vitaleta in Val d’Orcia, Tuscany, surrounded by cypress trees and golden hour fields

    This photograph of Chapel Vitaleta, Tuscany, Italy is one of my favourites from a recent trip to the region. Whilst I was happy with the photograph, no image is perfect and I had identified that the composition didn’t seem perfectly balanced to me and that the image was a little ‘yellow heavy’ in its colour palette.

    Feedback Verdict:

    1. The background of the image is low contrast, although I think that helps to emphasise the main subject of the image. In that sense ChatGPT is factually correct, but I’m not sure that increasing the contrast in the background will help make this image stronger. However, it wasn’t something I had considered and it does give me pause for thought.
    2. Upon checking the histogram of the image in Photoshop I did discover that the mist is marginally clipped in the red channel. I must have missed this in the excitement of editing – well spotted ChatGPT! Something for me to fix before printing this photograph.
    3. ChatGPT has identified a problem with the composition, but in the vertical axis. I actually think that the problem is in the horizontal axis. This feedback hasn’t told me anything I wasn’t aware of, but it is food for thought.