*

tperomaa Näkötutkijan näkökulma

Colors of “The Dress” (original & red-green versions): percepts of 884 readers

  • Figure 1. The original (left) and the red-green version (right) of the dress.
    Figure 1. The original (left) and the red-green version (right) of the dress.
  • Figure 2. Color ratings for the original dress.
    Figure 2. Color ratings for the original dress.
  • Figure 3. Color ratings for the red-green version of the dress.
    Figure 3. Color ratings for the red-green version of the dress.
  • Figure 4. The frequency of seeing the "most often perceived" colors, for the two conditions.
    Figure 4. The frequency of seeing the "most often perceived" colors, for the two conditions.
  • Figure 5. The comparison between the ratings under the two conditions.
    Figure 5. The comparison between the ratings under the two conditions.

[This is the technical version of my blog, for international readers.]

Purpose: The idea was to investigate, whether the color-swithing of the dress occurs also if the colors of the dress lie on the red-green axis (and not on the daylight axis, as in the original dress). Also, the original and red-green conditions were compared in terms of the disributions of the perceived colors. The main purpose of this study was to provide some pilot data to aid the future research.

Method: The two stimuli were the original dress and a red-green version (produced by swapping the a and b channels in Lab-mode -- I know, it's bad, but now we are operating in an uncalibrated environment) (fig. 1). The stimuli were presented in my personal blog, in Uusi Suomi (a finnish internet-only newspaper). Readers were asked to evaluate the perceived colors in the two versions of dress, using an e-questionnaire. The first question was: “Which colors do you see most often?” (For the original version, I asked to consider all the different versions of the original dress the person had seen.) For the original dress, the color choices were: blue-black, light blue-gold and white-gold. For the red-green version, the choices were: green-black, green-red and green-brown. Also "Another"-option was provided, where the subject could describe the color. The second question was: “How frequently do you see those colors?” Here, the choices were: always, often, varies a lot. Subjects also indicated their sex and defects of color vision.

Results: 884 persons replied to the questionnaire (425 men, 448 women, 11 not given), during the 27 hours that the questionnaire was open. For the original dress, the distribution of colors was quite even between the three main color options (fig. 2). For the red-green dress, the most typical color perceived was red-green; but also green-black and green-brown was often perceived (fig. 3). Considerable variation of the particular shades of green and red was evident from the free responses. Most subjects reported that they always saw the same colors (fig. 4); only 9/1% of subjects reported frequent variation of the perceived color. Also, it can be seen that the persons seeing black-blue in the original version tended to see black-green in the red-green dress (fig. 5).

Conclusions: The range of perceived colors varies under both conditions, but for the original (blue-yellow) version this is more evident. The particular percept that the subject has seems quite persistent. If the person is correcting illumination under one condition, there is also tendency to do it in the other condition.

Piditkö tästä kirjoituksesta? Näytä se!

1Suosittele

Yksi käyttäjä suosittelee tätä kirjoitusta. - Näytä suosittelija

Mainos

Netin kootut tarjoukset ja alennukset