Strip the colour out of your reference and see what the values are really doing. Load a photo below — it never leaves your device.
A cadmium red and a viridian green can sit at almost the same lightness. Side by side they look like a strong contrast, since the hues are opposites. Photograph the same passage in black and white and it turns into one flat grey shape.
This is why paintings that felt vivid on the easel look muddy in a photo, and why "add more colour" almost never fixes it. What is missing is not saturation but value separation. Converting the reference to grayscale is how you find out before you have committed three hours of paint to it.
A plain black-and-white conversion still holds every gradation the original had. It removes hue, but it does not simplify — and the problem you are trying to solve is usually too much information, not too much colour.
So do it in two steps. Convert to grayscale to remove the hue, then reduce the level count until the image breaks into flat masses. Three or four levels is where most value problems become obvious: if the whole picture collapses into two shapes at four levels, your range is too narrow. If it stays legible, the structure is sound.
For planning a painting from those masses rather than just checking them, the image simplifier covers the same conversion with the composition workflow around it.
Check the reference itself. A photo with weak value contrast produces a weak painting no matter how well you render it. Better to find that out while you can still change the reference.
Photograph your own canvas and run it through. This is the honest one: the painting you think you are making and the one on the easel diverge quietly, and grayscale shows the gap.
A last check on focal point and read. If the area you meant to emphasise is not the area of strongest value contrast, the viewer's eye will go somewhere you did not intend.
Because colour hides value errors. A saturated red and a mid green can be almost identical in lightness, so a painting can look correct in colour and collapse into mush the moment you photograph it in black and white. Removing colour is the fastest way to see whether the value structure holds on its own.
A phone filter gives you one fixed conversion. Here you can also reduce the result to a set number of tonal levels, which is the part that actually helps: a continuous grayscale still contains every gradation you were trying to escape. Two to four levels shows you the structure; sixteen shows you the map.
Squinting cuts the light reaching your eye, which merges close values and reveals the big masses. It works, but it is imprecise and you cannot hold it while you paint. A grayscale conversion is the same test, held still and measurable. Use both: squint at your canvas, run the converter on your reference.
That is the test working. It usually means your values are bunched in the middle: plenty of mid-tones, no real darks, no real lights. The fix is to push the extremes, deepening the darkest shadow and brightening the lightest light, rather than adjusting hue. Contrast in value is what reads from across the room.
No. Conversion happens in your browser on your own device. The image never leaves your computer.