It looks like you are accessing from Korea🇰🇷
Do you want to move to Korean page?
Today we reconnect with Jeonju Choi, the 2023 KCRC Roasting Champion and 7th place finisher in the World Championship, for a deep-dive into his roasting philosophies and technical approach. Having previously explored his journey with Your Home Coffee Roasters, this conversation focuses on the nuances of his craft—from maintaining consistency across production batches to finding the perfect balance between digital precision and sensory evaluation.
I've used Cropster since I first learned roasting, so I continued using it when I opened my own café. One frustrating limitation with Cropster was the inability to change reference curves during an active roast. Firescope supports this feature, which I really appreciate.
Initially, the different UI was a bit confusing, but after adapting, I've been using it comfortably. I've tried both Artisan and Cropster before settling on Firescope. In my opinion, Artisan has no advantages except being free - it's particularly slow to respond and comparing roasts is cumbersome. I stuck with Cropster mainly for its cleaner UI.
Firescope offers much better value than Cropster while maintaining many of the conveniences. The ability to instantly switch reference curves is extremely useful when roasting new coffees and wanting to compare with similar past roasts. I also appreciate how quickly the Firescope team responds to feedback and implements improvements.
Roasting definitely becomes easier with digital profiling. Some people claim to roast by feel, but that's difficult for me. However, I don't blindly trust the data - I always verify through QC cupping to ensure the sensory experience matches what the data suggests.
I recently encountered this issue when relocating my roasting space. Even though I followed the same roasting profiles and achieved similar curves, the sensory evaluation revealed less sweetness than expected. These discrepancies are impossible to detect without cupping.
In a new environment, creating fresh profiles can help produce similar coffees. While data-based roasting helps maintain consistency, it can't replace sensory evaluation. I'd say roasting is 80% data-driven, but the remaining 20% sensory assessment is absolutely essential.
For me, sensory perception is what allows a roaster to determine if they've roasted correctly. Since becoming a champion, many roasters send me samples requesting evaluations. While it would be nice to only give positive feedback, I believe roasters grow by identifying and improving defects.
Sometimes roasters are very satisfied with their samples, but I might not completely share their enthusiasm. Providing such feedback relies on sensory abilities. Ultimately, I think a roaster's satisfaction comes from how well they can identify defects through sensory evaluation.
I strive to view my own coffee objectively and find any defects. Sensory skills are essential for continuously exploring and advancing coffee roasting.
I design my roasting based on the bean's altitude, processing method, and screen size. Previously, I didn't pay much attention to temperature and humidity data, but ignoring these variables has sometimes led to differences in actual roasting outcomes.
Particularly when bean temperature rises rapidly after the turning point, I quickly reference past roasting records to determine how to adjust the heat. With continued roasting experience, I've built up responses to most scenarios, allowing me to roast more confidently.
The most important roasting indicator I watch is the turning point temperature. However, I don't focus exclusively on that. The turning point sets the foundation for the overall roasting profile. Rather than fixating on specific metrics, I take a comprehensive view of bean ROR (Rate of Rise), the speed at which bean temperature increases, and the overall roasting curve. This holistic approach helps me determine if the roast is progressing too quickly or too slowly and adjust accordingly.
Comparing the peak bean ROR values can provide insight into the flavor profile. While other roasting defects may appear simultaneously, I believe that peak bean ROR generally correlates with flavor intensity when those defects are controlled for.
However, a higher bean ROR typically leads to faster overall roasting time, which might not translate to sufficient sweetness or balanced coffee. Therefore, setting targets for the maximum ROR point and adjusting heat accordingly is where I focus most of my attention.
Unlike some other roasters, I don't frequently reduce the heat level. I prefer to set up conditions that allow stable roasting without constant heat adjustments. I only modify the heat when I determine it's too strong at first crack as I finish the roast. The Diedrich roaster I use has excellent heat retention, so slightly reducing heat has minimal effect - I typically just turn off the heat before dropping the beans.
The biggest improvement from using roasting software is being able to roast without expending as much mental energy as when I first started. Relying solely on sensory perception means depending on memory, which introduces uncertainty. With data, I can precisely track how long each stage lasted.
For example, by monitoring data from the drying and development phases, I can identify when a batch is taking longer than previous ones and make adjustments for subsequent batches.
From my personal experience with the Diedrich roaster, which doesn't allow for very precise heat control adjustments, new roasters can sometimes panic. I'd want them to know that Firescope's ability to reference past roasting records means they can achieve quality roasts simply by following historical data closely.
Additionally, recording thoughts in Firescope's notes feature after each roast means you can immediately understand your intentions when referencing that roast later. Rather than teaching new roasters various methods, I'd show them a few past records and help them learn independently through the notes and data.
Many people send me samples for evaluation and when I ask about their roasting software, most use Artisan. I understand using it because it's free, but I sometimes wonder why they persist with it despite its inconveniences.
Occasionally, roasters working with very expensive coffees come to me with roasts that could use improvement. When I try to provide feedback, it's difficult to pinpoint exactly where variables occurred if they're using Artisan. In those cases, I recommend Firescope, telling them "using Firescope would make our discussions more productive."
I especially recommend Firescope to new roasters or those who use roasting software but don't analyze it thoroughly.
Paying for something means it has corresponding value. I believe roasting without investing in a good software like Firescope wastes both the time spent roasting and the green coffee itself.
For those thinking, "I'll adopt Firescope after I improve my skills with more practice," I'd suggest abandoning that mindset. Using quality tools from the beginning to accelerate development is the correct approach to building roasting skills.
Jeonju’s approach exemplifies how modern roasters can leverage technology to enhance their craft while maintaining their unique roasting philosophy, showing that consistency and creativity aren’t mutually exclusive. Whether you’re just beginning your roasting journey or looking to refine your process, we invite you to explore how Firescope can help you develop your own distinctive roasting style while ensuring excellence.