WEC 24-hour races generate 10,000+ photos across continuous day-night-dawn shooting, multiple driver changes, and three distinct car classes. Here's the workflow that keeps pace with a 24-hour marathon.
WEC and Le Mans 24h photography is endurance shooting at its most extreme. You're shooting continuously for 6, 12, or 24 hours, watching cars return 80-100+ times each, managing extreme lighting transitions from sunset to pitch black to dawn, and tagging drivers who switch seats mid-race. The cars are grouped in three classes (Hypercar, LMDh, LMGT3) with distinct appearance but similar numbering schemes. Battery, storage, and mental fatigue management are as critical as the tagging workflow itself.
Typical Event
24 hours continuous (Fri 3 PM → Sat 3 PM), with pre-race Fri morning (6-8 hours of additional photos from practice/qualifying)
Photo Volume
5,000-10,000+ photos across the full 24-hour race, with additional 1,000-2,000 from Friday practice/qualifying
Delivery
Continuous during race (batch deliveries every 2-4 hours), full archive within 4 hours of checkered
Key Challenge
Lighting transitions and driver changes. A Hypercar passes the camera 100+ times under different lighting conditions (day, twilight, night with headlights, dawn). Different drivers sit in the same car at different times. You must tag each pass individually and account for the driver currently piloting.
Import the official WEC entry list (car numbers, teams, classes). More critically, get the team-provided driver schedules — who drives which stint, which driver #1/#2/#3 sits in car #63 at what time. Tag photos with 'driver stint' metadata so you can filter deliveries by 'morning shift' or 'night shift' later.
Pro tip
Create a simple spreadsheet with columns: Car#, Driver1, Driver2, Driver3, Stint1-Start, Stint1-End, Stint2-Start, etc. At Le Mans especially, drive changes happen every 4 hours. You'll reference this constantly.
Don't try to tag everything at the end. Every 2-4 hours (end of a driver stint), offload your cards to a laptop in the paddock. Cull the batch quickly, back up to both local and cloud drives (24h race = can't lose footage to card failure), then stage for tagging.
Pro tip
Use dual card bodies in overflow mode. Every 2 hours one card is full — extract that card while the other fills. Never run out of storage mid-race. Bring 3-4 extra cards to Le Mans; you'll use them.
Run RaceTagger on each 2-4 hour batch. The AI detects car numbers and classes. You manually assign the driver stint — RaceTagger records which driver was piloting based on your stint guide. This way, a photo of car #63 at 11 PM is tagged to 'Driver 2', not 'Driver 1' (who finished at 8 PM).
Pro tip
Don't skip the driver assignment. Yes, it adds 2-3 minutes per batch, but your deliveries to teams become 'all photos of Driver #2' instead of 'all photos of car #63'. Teams value driver-specific packages for media relations.
RaceTagger flags low-confidence detections. In WEC, expect higher flagging at transition moments: sunset (mixed color temp, dramatic shadows), 2-3 hours before dawn (darkest part of night), and right at dawn (rapidly changing contrast). Also review driver-change photos manually — confirm the stint assignment is correct.
Pro tip
Night shots under headlights will have a higher flag rate (8-12%). Budget for it. The photos are often the most visually striking, so the manual review is worth the effort.
Every 4 hours, export tagged photos from that stint as a separate delivery: 'Car #63 - Driver 2 - 8 PM-midnight'. Teams and sponsors want to see their content continuously, not in one dump at the end. Export to a shared folder, notify the team, move on to the next stint.
Pro tip
Set up a shared Dropbox or OneDrive folder with the team before the race. Agree on a delivery schedule: 'We'll deliver every 4 hours starting at 7 PM.' They'll appreciate the predictability.
After checkered, import all 24h photos into Lightroom in one master catalog. Filter by class: Hypercar, LMDh, LMGT3. Create separate galleries for each class. Deliver the final comprehensive archive along with class-specific highlights. Store everything for long-term archiving — Le Mans photos have commercial value for years.
Pro tip
Hypercar (the prestige class) photos sell best. Create a separate 'Hypercar highlights' gallery with your best shots — this becomes portfolio material and attracts future WEC contract work.
Why it's hard: Car #63 photographed at 2 PM has bright sidepod lighting; at 2 AM under headlights it's silhouette + glare; at 6 AM pre-dawn it's gray-blue twilight with reduced contrast. The same number reads completely differently under each lighting regime.
How AI helps: The AI adapts to lighting context. It understands that 'hard to read' doesn't mean 'incorrect' — the model re-trains on WEC-specific lighting transitions and reads numbers robustly across all conditions.
Why it's hard: Hypercar numbers sit on a silver background, LMDh on gold, LMGT3 on blue. A #63 in each class looks visually distinct, but at distance or under night lighting, the color differentiation is lost. You need to ID both number AND class simultaneously.
How AI helps: The AI learns class-specific visual cues: car body shape, wheel design, spoiler profile, and number plate color. It detects the number and the class together. Photos are auto-tagged to the correct starting list based on both identifiers.
Why it's hard: Photo of car #63 taken at 11 PM shows Driver #2. Same car at 3 AM shows Driver #3 after a pit stop. You can't tell from the car alone who's driving. You must correlate your shooting time with team-provided stint schedules.
How AI helps: RaceTagger detects the car number correctly. You manually assign the driver based on the timestamp and your pre-race stint guide. The AI doesn't guess at driver identity — it records your assignment in metadata so deliveries are driver-specific, not just car-specific.
Why it's hard: At night, cars blast through the frame with headlights at full intensity. Light reflects off the door panel, creating hotspots that overexpose the number area or wash out the number into the bright reflection.
How AI helps: The AI reads numbers using structure and edge detection, not just luminosity. A number partially washed out by glare can still be identified from its geometric shape and context.
Why it's hard: This isn't a technical challenge — it's human. After 16 hours of shooting and tagging, your attention lapses. Manual reviews become sloppy, you miss flags, deliveries get delayed.
How AI helps: AI tagging removes the 8-hour manual review that would destroy your concentration. You focus on shot composition, positioning, and strategic placement. The AI handles the repetitive tagging. By hour 20, you're fresher and make better creative decisions.
Manual Tagging
16-24 hours of manual tagging spread across the 24h race (1-2 hours per stint, post-race review)
75-85% overall — drops significantly in night sequences and with driver changes
With RaceTagger AI
25 minutes processing total + 60-90 minutes distributed manual review across 24 hours (5-10 min per batch)
95%+ on clean shots, 90%+ even in night sequences with intelligent flagging
Real-world scenario
It's Saturday 7 PM at Le Mans. You've shot practice/qualifying (1,200 photos), culled to 600, tagged with RaceTagger (8 minutes), delivered to teams by 3 PM. Race started at 3 PM, you're at your first position for the parade lap. By 7 PM, you've shot 2,000 photos across the first 4 hours. You retreat to the paddock, offload cards, cull to 1,200 (15 minutes), run RaceTagger (6 minutes). All Hypercars from the first stint are tagged and assigned to drivers. You export the 'Hypercar - Stint 1' folder and drop it in the shared Dropbox. Team managers are seeing their content within 90 minutes of pit stop. At midnight, you repeat — 1,800 more photos, tagged, reviewed, delivered by 12:15 AM. Night shots have a slightly higher flag rate (11% vs 7%), but you catch and fix them quickly. By 4 AM, you're tired but not destroyed because RaceTagger handled the mechanical work. At sunrise (6 AM), the light shifts to dawn blue-hour — RaceTagger adapts, flag rate rises to 12% for the twilight zone, but the final deliver of 'Hypercar Sunrise Stint' is crisp and professional. By noon Sunday, checkered flag, you've delivered 8 interim packages and now you're importing the final 24-hour master archive.
Competitors who manually tagged everything are delivering Monday afternoon, delirious from sleeplessness. You're delivering Sunday 5 PM, every interim package already in teams' hands, and you still have mental energy for the final curation. Your Le Mans portfolio becomes the best in the field.
500 free tokens included. No credit card required. Upload 500 photos from a previous 24h race across day, night, and dawn sessions to see how it handles lighting transitions.
Start tagging for free →How do I handle the driver changes when the same car swaps drivers every 4 hours?
You create a stint reference guide before the race (team provides this). During tagging, you assign each batch of photos to the correct driver stint based on timestamp. RaceTagger records this in metadata, so deliveries can be driver-specific ('All photos of Driver #2') instead of just car-specific.
What's the expected accuracy in night shooting under headlights?
Night accuracy is 90%+ with a slightly higher flag rate (8-12% vs 5-8% in daylight). Headlight glare and darkness challenge the AI, but it still reads numbers from context and structure. You'll spend 10-15 minutes per stint manually reviewing flagged night shots.
Can I deliver interim batches to teams every 4 hours instead of waiting until the end?
Yes, and it's highly recommended. Process and deliver every stint: teams see their car and drivers immediately, not waiting for the 24-hour dump. Set up a shared Dropbox folder with agreed delivery schedule before the race.
How many photos should I expect to shoot in a 24-hour race?
Typical range is 5,000-10,000 depending on positioning, number of pit stops you cover, and whether you're shooting practice/qualifying too. Pro tip: shoot conservatively in early stints (3 PM-9 PM) and maximize in the evening/night/dawn when light is more dramatic. You'll still exceed 10,000 total.
Does RaceTagger handle the three WEC classes (Hypercar/LMDh/LMGT3) separately?
Yes. Import all three classes in a single CSV. The AI detects the class from visual features (car shape, spoiler style, number plate color) and matches to the correct entry. Your exports can be filtered by class: 'All Hypercars' vs 'All LMGT3', etc.
Related
Night Photography at 24-Hour Endurance Races: Technique and AI WorkflowHalf of Le Mans is shot in darkness — night-specific lighting strategy and AI accuracy expectations
Related
Driver Change Detection in Endurance Racing — Matching Stint Schedules to PhotosDeep dive into managing multi-driver vehicles and keeping stint assignments accurate
Related
Hypercar vs LMDh vs LMGT3: Class-Specific Detection in WEC PhotographyVisual guide to distinguishing the three WEC classes and how AI identifies each from body geometry