Any phone apps that will analyze how dark coffee beans are and give you a number value?

February 17th, 2021, 3:50 pm

Just wondering if there's an app that will give you a number corresponding to avg coffee bean color that you can use as an objective measure of roast darkness. Obviously not as good as spectrometer...

#2: Post by **daustin777** » February 18th, 2021, 12:01 pm

If there was an app, it would not work. Different lighting used to capture the photo will alter the color in the photo a significant amount.

The cheapest solution to get Agtron-ish numbers is this card from Roastrite:

https://www.amazon.com/dp/B0727VPGQK/

#3: Post by **rmongiovi** » February 18th, 2021, 12:57 pm

My bet would be that even with that card you're at the mercy of the CRI of your light source. I don't know much about the cards. Do they require a D65 illuminant?

#4: Post by **Jeff** » February 18th, 2021, 1:04 pm

I think it all depends on your needs. If you want to answer "Is this light or medium?" the cards are sufficient with a light source that isn't red-deficient (most non-premium LED bulbs are poor with red). If you're looking to control your roasts, you need something with a more controlled light source and detector. They seem to start with Roast Vision - a new and affordable roast level sensor for $299 and go up from there with build quality and industry acceptance.

February 19th, 2021, 3:04 pm

Jeff wrote:with build quality and industry acceptance.

Is build quality truly in question? I might have thought a real issue is the small sample size to the overall volume of a given roast. The science seems reasonable, sensory drift minimal & reproducible. How to quantify roast level to the larger volume to then make it acceptable to the community as a reference is a question to be discussed. Presentation regarding repeatable quality calibration will help to offer confidence. Then the question to resolve the Roast Vision proprietary scale must follow, for general acceptance as a functional resource and reference.

February 19th, 2021, 5:42 pm

Roast Vision now has a correlation with Agtron values. I use mine frequently and enjoy it. It's all relative and probably a different discussion for commercial operations vs. home roasters/consumers.

February 20th, 2021, 10:12 am

Yes, it seems that lighting makes a big difference. I tried some phone apps that give you RGB values (and then converting these to 1 total value, according an equation i found).... there was too much variability depending on lighting though.

February 20th, 2021, 12:48 pm

MerleApAmber wrote:... I might have thought a real issue is the small sample size to the overall volume of a given roast. ... How to quantify roast level to the larger volume to then make it acceptable to the community as a reference is a question to be discussed. ...

It's not a discussion; it's statistics. There is no right answer.

For example, people who use Roast Vision to quantify the roasted coffee that they buy are usually buying fairly-to-very expensive coffee. This coffee is often specialty-grade, well-sorted, carefully-prepped and -processed, and free of green coffee defects. The beans may even be from a single lot/farm/co-operative/etc. Assuming that the roaster (machine and person) does a decent job, then this coffee will be very uniform in colour within a specific roast. Relatively few samples/small sample sizes will be needed to say something about the average roast level for that specific roast for a given level of statistical significance.

On the other hand, I recently looked at a popular espresso blend from Italy. There were multiple shapes/types of beans indicating multiple origins. There was also a wide variety in screen size. There were quakers and semi-quakers, and some beans had coffee borer damage. One would have to take more samples/use larger sample sizes to say something about the average roast level for that specific roast at the same level of statistical significance.

February 23rd, 2021, 7:10 pm

baldheadracing wrote:It's not a discussion; it's statistics. There is no right answer.

I can see the points regards the retail purchasing end of the chain of coffee transactions. And yes, both make absolute sense. I could easily see either happening in front of me.

Where my focus is is that of a novice roaster who is quantifying the quality of roasts I am performing. Then taking that "optic" to the degree a professional roaster might have working more of the 'blend' mix out of a quality bulk bag due to (perhaps rare...) padding. (oh, heck, the real world here, even a good bag of beans is hobs choice unless you've paid ten times the price - may be.) Yes, boutique roasters may ensure every bean is consistent and perfect AA grade, but I'd have a hard time imagining the same performance out of a huge roast volume per load.

Cheers!

#10: Post by **nsuster** » February 23rd, 2021, 7:41 pm

For blends, I typically roast each individual coffee and then blend post roast. I'm assuming this is the best way to go so you can roast each individual coffee based on size, density, moisture content, etc. I'll pay attention to each individual value on the Roast Vision and then blend them and compare. I've not actually done this yet but will give it a go in the future.

I don't have much experience with the lower quality coffee issues but there is some variability with light vs. dark beans in my roasts so you tend to need to average across a large enough sample after the coffee is ground/stirred.