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.
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:
- 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.
- 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.
- 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.