Microfading Tester International Discussion Group

Dear Christel,

Apologies for the delayed response.

The measurement of dark colors has always been an issue due to low signal to the spectrometer which can then be more easily impacted by electrical noise or stray light. For this reason, I've generally steered towards the assessment of samples that are medium to light density when available. For dark samples, you might consider frequent dark and light calibrations to increase accuracy and analysis in a dark room to reduce ancillary light.

Regarding your two issues, I think it would be helpful to share spectra and the resulting colorimetry data. For the two situations you described, what constitutes a "high DE" (issue 1: consistent results that do not reflect change in %R) and an "abnormal super high DE" (issue 2: only observed in some runs)?

It is an interesting idea to determine a lower L* limit below which we need to be cautious in interpreting our MFT results. However, your suggestion of a lower L* limit of 25 seems high to me, as I would expect we should still be able to get repeatable results at that level. (Note that when I solve your equation for L*, I get a value of 8.) I must admit that I'm unsure about the use of the equations for this purpose, and might need you to walk me through your rationale a bit more. We might also consult our color science colleagues!

Alternatively, we could approach the issue experimentally. For example, we can identify a set of gray scale ceramic tiles with decreasing reflectance percentages (i.e., decreasing L* values) which could then be tested by different MFT users to assess the measurement consistency of each given setup and protocol. Of course, we can add additional samples, such as BW standards, to expand our insights into the performance of our MFTs.

Happy to discuss further, and I promise I'll respond faster next time!

Dear all,

I am writing as I would appreciate input/confirmation from the MFT users community regarding some of the limitations I have been facing when testing very dark colors. It happened a few times in the past to obtain puzzling recurrent results that I can divide into 2 categories listed below. Considering the fact that dark colors can be rather sensitive depending on the media, I have tried to make sense of these 2 types of recurring results to further consolidate my interpretation. I have formulated my hypotheses and reasoning below and would highly appreciate your feedback, critical perspective, comments and additional suggestions. I wish such discussion can be be useful for the MFT user community 

I thank you in advance for your feedback and wish all of you to have a good weekend ahead, especially to our US colleagues to celebrate a peaceful, inclusive and joyful 4th of July !

With my very best,

Christel

Issues, interpretation and reasoning: 

1- Issue 1: Consistent results obtained between a couple of runs showing a rather high final DE, that however does not seem to accurately reflect the change in the %R induced by the test

I went back to re-explore some of the fundamental basics in color science, especially how CIE La*b* coordinates are calculated from the tristimulus value to better understand the potential limit of application of the proposed computing. 

a- For reference, I used the equations provided by the website indicated below:

To ease the reading I copied the equations alongside the respective conditions, understanding that I have focused working on the one related to L* considering my interest in testing dark colors.

b- I have then calculated the limit of initial L* under which traditional equations used in color sciences to calculate La*b* and successive DE.

Solving the equation [(L*+ 16)/116]3 < 0.008856 , I have deduced that caution should particularly be applied when interpreting MFT results obtained on a color with an initial L* < 25 

Do you think that it would make sense and be useful to consider using such a threshold for data interpretation ?

c- For those values, it may be beneficial to apply the set of equation provided below ( from the same site):

d- Furthermore, this may be worth integrating into the spectral viewer, check and compare the results obtained to assess if it helps achieve a stronger alignment between DE obtained and induced change in %R. 

2- Issue 2: Successive runs among which some show abnormal super high Delta E are found. 

I interpret such a situation as highlighting the impact a measurement error, even though small, can have on a signal as low as the ones characterising a very dark color. 

I am not sure much more can be done so I usually dismiss the run and not use the obtained result.

 

Dear Christel,

Apologies for the delayed response.

The measurement of dark colors has always been an issue due to low signal to the spectrometer which can then be more easily impacted by electrical noise or stray light. For this reason, I've generally steered towards the assessment of samples that are medium to light density when available. For dark samples, you might consider frequent dark and light calibrations to increase accuracy and analysis in a dark room to reduce ancillary light.

Regarding your two issues, I think it would be helpful to share spectra and the resulting colorimetry data. For the two situations you described, what constitutes a "high DE" (issue 1: consistent results that do not reflect change in %R) and an "abnormal super high DE" (issue 2: only observed in some runs)?

It is an interesting idea to determine a lower L* limit below which we need to be cautious in interpreting our MFT results. However, your suggestion of a lower L* limit of 25 seems high to me, as I would expect we should still be able to get repeatable results at that level. (Note that when I solve your equation for L*, I get a value of 8.) I must admit that I'm unsure about the use of the equations for this purpose, and might need you to walk me through your rationale a bit more. We might also consult our color science colleagues!

Alternatively, we could approach the issue experimentally. For example, we can identify a set of gray scale ceramic tiles with decreasing reflectance percentages (i.e., decreasing L* values) which could then be tested by different MFT users to assess the measurement consistency of each given setup and protocol. Of course, we can add additional samples, such as BW standards, to expand our insights into the performance of our MFTs.

Happy to discuss further, and I promise I'll respond faster next time!

Dear all,

I am writing as I would appreciate input/confirmation from the MFT users community regarding some of the limitations I have been facing when testing very dark colors. It happened a few times in the past to obtain puzzling recurrent results that I can divide into 2 categories listed below. Considering the fact that dark colors can be rather sensitive depending on the media, I have tried to make sense of these 2 types of recurring results to further consolidate my interpretation. I have formulated my hypotheses and reasoning below and would highly appreciate your feedback, critical perspective, comments and additional suggestions. I wish such discussion can be be useful for the MFT user community 

I thank you in advance for your feedback and wish all of you to have a good weekend ahead, especially to our US colleagues to celebrate a peaceful, inclusive and joyful 4th of July !

With my very best,

Christel

Issues, interpretation and reasoning: 

1- Issue 1: Consistent results obtained between a couple of runs showing a rather high final DE, that however does not seem to accurately reflect the change in the %R induced by the test

I went back to re-explore some of the fundamental basics in color science, especially how CIE La*b* coordinates are calculated from the tristimulus value to better understand the potential limit of application of the proposed computing. 

a- For reference, I used the equations provided by the website indicated below:

To ease the reading I copied the equations alongside the respective conditions, understanding that I have focused working on the one related to L* considering my interest in testing dark colors.

b- I have then calculated the limit of initial L* under which traditional equations used in color sciences to calculate La*b* and successive DE.

Solving the equation [(L*+ 16)/116]3 < 0.008856 , I have deduced that caution should particularly be applied when interpreting MFT results obtained on a color with an initial L* < 25 

Do you think that it would make sense and be useful to consider using such a threshold for data interpretation ?

c- For those values, it may be beneficial to apply the set of equation provided below ( from the same site):

d- Furthermore, this may be worth integrating into the spectral viewer, check and compare the results obtained to assess if it helps achieve a stronger alignment between DE obtained and induced change in %R. 

2- Issue 2: Successive runs among which some show abnormal super high Delta E are found. 

I interpret such a situation as highlighting the impact a measurement error, even though small, can have on a signal as low as the ones characterising a very dark color. 

I am not sure much more can be done so I usually dismiss the run and not use the obtained result.

 

Dear Christel and Vincent,

Thank you for the interesting discussion. Happy to contribute some data and spectra for further discussion.

A bit of background. I have been testing the black on three Fuji Instant dye diffusion transfer photographs (Fuji FP100c). For each of the photographs, I tested three areas of black and between the tests, ran dark and white calibration. For the first two photographs, the L values of the blacks ranged above 10 to about 13, the DE values were quite consistent and the reflectance curves didn't register any big change. Please see attached data and fading/ colour shift curve of example 1. I think this example would belong to issue 1 that Christel highlighted. I'm interested to know whether the results are valid, given that there is consistency, DE values are high and yet the reflectance curves don't show a significant change. 

For the third photograph, please see attached example 2. The L values of the three black areas tested were below 10 and the three tests yielded very different results. Surely the run of BLK 4.3 shows the limitation of the spectrophotometer and belongs to issue 2 that Christel highlighted? 

Looking forward to hear your opinions, to see whether I missed out something and can improve my mft practice and interpretation in some way. Many thanks.

Regards,

Esther

Dear Christel,

Apologies for the delayed response.

The measurement of dark colors has always been an issue due to low signal to the spectrometer which can then be more easily impacted by electrical noise or stray light. For this reason, I've generally steered towards the assessment of samples that are medium to light density when available. For dark samples, you might consider frequent dark and light calibrations to increase accuracy and analysis in a dark room to reduce ancillary light.

Regarding your two issues, I think it would be helpful to share spectra and the resulting colorimetry data. For the two situations you described, what constitutes a "high DE" (issue 1: consistent results that do not reflect change in %R) and an "abnormal super high DE" (issue 2: only observed in some runs)?

It is an interesting idea to determine a lower L* limit below which we need to be cautious in interpreting our MFT results. However, your suggestion of a lower L* limit of 25 seems high to me, as I would expect we should still be able to get repeatable results at that level. (Note that when I solve your equation for L*, I get a value of 8.) I must admit that I'm unsure about the use of the equations for this purpose, and might need you to walk me through your rationale a bit more. We might also consult our color science colleagues!

Alternatively, we could approach the issue experimentally. For example, we can identify a set of gray scale ceramic tiles with decreasing reflectance percentages (i.e., decreasing L* values) which could then be tested by different MFT users to assess the measurement consistency of each given setup and protocol. Of course, we can add additional samples, such as BW standards, to expand our insights into the performance of our MFTs.

Happy to discuss further, and I promise I'll respond faster next time!

Dear all,

I am writing as I would appreciate input/confirmation from the MFT users community regarding some of the limitations I have been facing when testing very dark colors. It happened a few times in the past to obtain puzzling recurrent results that I can divide into 2 categories listed below. Considering the fact that dark colors can be rather sensitive depending on the media, I have tried to make sense of these 2 types of recurring results to further consolidate my interpretation. I have formulated my hypotheses and reasoning below and would highly appreciate your feedback, critical perspective, comments and additional suggestions. I wish such discussion can be be useful for the MFT user community 

I thank you in advance for your feedback and wish all of you to have a good weekend ahead, especially to our US colleagues to celebrate a peaceful, inclusive and joyful 4th of July !

With my very best,

Christel

Issues, interpretation and reasoning: 

1- Issue 1: Consistent results obtained between a couple of runs showing a rather high final DE, that however does not seem to accurately reflect the change in the %R induced by the test

I went back to re-explore some of the fundamental basics in color science, especially how CIE La*b* coordinates are calculated from the tristimulus value to better understand the potential limit of application of the proposed computing. 

a- For reference, I used the equations provided by the website indicated below:

To ease the reading I copied the equations alongside the respective conditions, understanding that I have focused working on the one related to L* considering my interest in testing dark colors.

b- I have then calculated the limit of initial L* under which traditional equations used in color sciences to calculate La*b* and successive DE.

Solving the equation [(L*+ 16)/116]3 < 0.008856 , I have deduced that caution should particularly be applied when interpreting MFT results obtained on a color with an initial L* < 25 

Do you think that it would make sense and be useful to consider using such a threshold for data interpretation ?

c- For those values, it may be beneficial to apply the set of equation provided below ( from the same site):

d- Furthermore, this may be worth integrating into the spectral viewer, check and compare the results obtained to assess if it helps achieve a stronger alignment between DE obtained and induced change in %R. 

2- Issue 2: Successive runs among which some show abnormal super high Delta E are found. 

I interpret such a situation as highlighting the impact a measurement error, even though small, can have on a signal as low as the ones characterising a very dark color. 

I am not sure much more can be done so I usually dismiss the run and not use the obtained result.