You're nearly halfway through the 8-week NIR Fundamentals course. Pick up where you left off — Lesson 12, slide 8 — or jump to your week-4 quiz to unlock Week 5.
The player keeps the narration video and slide deck in sync. The slide navigator on the right lets you jump to any slide — completed slides are checked, the current one is highlighted, and the rest are one click away.
Short knowledge checks appear inside lessons and are graded the moment you answer. End-of-week quizzes gate the next week — pass to unlock.
PDS builds local regression windows that translate host-instrument spectra into the master's response space, preserving calibration accuracy. MSC and SNV are scatter corrections; Savitzky–Golay is a smoothing/derivative filter.
Take notes pinned to the exact video timestamp and slide. Flag priority — normal, high, or done — and export the whole notebook to Markdown for review.
PDS window size is a tradeoff — too wide overfits, too narrow misses scatter. Test 5/7/9 channel windows on validation set.
Direct standardization = single global transform matrix. Simpler than PDS but weaker on non-linear instrument differences.
Remember: slope & bias correction handles only linear offset between instruments.
Stored per slide in lesson_notes · markdown export includes timestamps + priority flags.
Slide decks, reference sheets, and worked spreadsheets attached to each lesson — download as PDF, PPTX, or XLS.
Finish a week's lessons, pass the week quiz, and the next week unlocks. You've cleared 6 of 8 quizzes — Week 5 is live now.
Spectra is an AI assistant trained on the course material. Ask it to clarify a concept, explain a slide, or work through a chemometrics problem — it answers with the lesson context in mind.
Why does NIR use overtone bands instead of fundamental vibrations like MIR?
Fundamental vibrations live in the mid-infrared and absorb very strongly — so strongly that samples must be diluted or thinly sliced. In the NIR region (780–2500 nm) you see overtones and combination bands of C–H, O–H, and N–H bonds.
Those are 10–1000× weaker, which is actually the advantage: light penetrates deeper, so you can scan intact grain, flour, or whole kernels with little prep.
See Lesson 5 · Slide 4Every lesson has a discussion thread. Post questions, answer peers, and upvote the most helpful replies. Instructor answers are highlighted.
When transferring a calibration, how many standardization samples do you realistically need? My host instrument is the same model — can I get away with fewer?
Same model usually transfers well with 15–30 standardization samples spanning your constituent range. Always validate on an independent set — if RMSEP creeps up, add samples at the range extremes first.
Is PDS better than slope/bias correction for moisture in flour, or does it depend on the spectrometer drift?
Week-quiz results and assignment grades with instructor feedback. Two assignments are still pending submission.
Direct messages from your instructor — nudges, feedback, and announcements land in your inbox.
You're crushing Week 3. Your week-4 quiz is open now — clear it and Week 5 (Calibration) unlocks. Don't forget the L16 industry brief is due in 4 days.
This week covers real-world deployments in grain, dairy, and feed. The L14 case study is my favorite — watch for the moisture drift example.
Finish all 32 lessons and pass all 8 week quizzes to generate a verifiable PDF certificate. You're 47% of the way there.
Everything above is a labeled sample. Enroll free to get your own dashboard, player, notes, quizzes, and certificate path.
No credit card · First 2 lessons always free