It was suggested that a test be done wherein a camera would be used to photograph and measure light output from factory fluorescents and LED replacements. I finally got the chance to do so.
First, the test - both rigs are midbaths. One has darker wood than the other. That only mattered for one of three shots - the one I took at floor level. I placed a white sheet of paper in three spots - directly under the light above the table, then at the end of the table and finally on the floor below the oven. The darker wood was taken out of the equation by placing aluminum foil against the wood paneling between the floor and the bottom of the oven in both rigs. The camera was set so that it "saw" only the white paper. No light from the table or floor actually reached the camera sensor. I set the camera in manual mode, f5.6 and ISO 200, then used the manual adjustment to change the shutter speed. When the camera's light level reported correct exposure settings I was able to use the shutter speed numbers to compare light levels. Taking a pic of the white paper allowed me to measure the color of the light produced.
The LED replacements used are from Jirah (www.jirahled.com, item JC-72-FL-T) and consist of a double strip of 36 LED's per side, two per fixture, or 144 LED's in the original housing. The 144 LED's draw about 1.4 amps, compared to 2.4 for the fluorescents, if I remember correctly.
The fluorescents read at 3600 K and the LED's at 5150 K. This means that the fluorescents were a bit yellowish when compared to the more white light of the LED's. Some people prefer that warmer look rather than the white of the LED's. Me? I like the white light.
Directly under the light, the fluorescent reading was 25, or 1/25 of a second and the LED reading was 20, or 1/20 of a second. The light produced by the LED's was 80% of the light produced by the fluorescents. At the end of the table, the readings were 13 and 10, meaning that the LED's had about 77% of the intensity of the fluorescents. On the floor against the stove, the fluorescents read 3 and the LED's read 2.5, or 83% of the fluorescent brightness. The variation from 77% to 83% is not a significant difference, given the degree of precision of the test. Essentially, the light intensity of the LED strips is about 80% of the intensity of the fluorescents, and is no more directional than the fluorescents, radiating the same amount of light to the side.
The LED's used in this test draw 58% of the power of fluorescents, produce 80% of the light of fluorescents, have an equivalent lateral diffusion with fluorescents, and have a significantly whiter color than the fluorescents. That being said, both LED's and fluorescents are available in different degrees of whiteness.