Consider a commercial troffer that clears internal testing and the photometric report goes out to the utility program for rebate qualification. Two weeks later the submission comes back rejected. The measured total luminous flux in the file disagrees with the driver input wattage keyword by more than the program allows, the test date field is formatted in a way the parser will not accept, and the vertical angle grid stops at 90 degrees on a fixture that puts measurable light above horizontal. None of these are optical problems. They are photometric-file problems rather than optical-design failures, yet each can delay qualification and force avoidable rework.
The photometric file is the deliverable that leaves the optics group and enters the commercial world. Specifiers load it into lighting layout tools, utility programs parse it for qualification, and regulators evaluate it against distribution templates. It is also the artifact most often produced as an afterthought, generated once at the end of a program from a single measurement and never reconciled against the design model that produced the fixture. This article covers how to produce luminous intensity distributions from a ray trace in TracePro, what the IESNA LM-63 and EULUMDAT formats actually encode, how to configure candela-plot orientation and angular sampling that survive scrutiny, and how to correlate a simulated distribution against goniophotometer data so that the file you ship matches the hardware you built.
What a Photometric File Actually Encodes
An IESNA LM-63 file is a text description of luminous intensity as a function of direction, plus a header block of metadata. The intensity table is organized on a grid of vertical angles, conventionally called gamma, and horizontal angles, conventionally called C planes. A Type C photometry file for an interior downlight might use gamma from 0 to 180 degrees in 2.5 or 5 degree steps and C planes from 0 to 360 degrees in 22.5 degree steps, exploiting rotational symmetry where it exists. An automotive or roadway fixture with a strongly asymmetric distribution will need a finer grid and will often use Type B photometry with a different angular convention.
Two facts about the format drive most downstream problems. The first is that the file encodes intensity in candela, a far-field quantity, not illuminance or flux. Any tool that reads the file reconstructs illuminance at a target by applying the inverse square law and the cosine of the incidence angle, which is only valid at distances where the luminaire behaves as a point source. The second is that the header keywords are load bearing. Input watts, ballast or driver factor, lamp catalog number, test laboratory, test date, and luminous opening dimensions are all parsed by qualification programs and layout software. Downstream systems may use these fields to derive efficacy or validate a submission, so the header must be reconciled with the simulated or measured distribution before release.
EULUMDAT, carried in LDT files, encodes the same physical content with a different structure and is widely used in European lighting workflows. TracePro writes both, so the choice is a matter of which market and which downstream tool consumes the file rather than a modeling decision.
From Ray Trace to Candela Distribution
TracePro builds a luminous intensity distribution by sorting the rays that escape the model into angular bins. Every ray that leaves the fixture carries a flux value and a direction, and the candela distribution analysis accumulates flux into solid angle bins, then divides by the bin solid angle to produce intensity. This is a direct consequence of Monte Carlo ray tracing rather than a separate calculation mode, which means the quality of the distribution is governed by the same considerations that govern any non-sequential trace. Ray tracing method and its role across the design workflow are covered in more depth on the TracePro ray tracing hub.
Choosing the Angular Grid
Bin size is a trade between angular resolution and statistical noise. Narrow bins resolve sharp cutoffs and beam edges but collect fewer rays each, so the intensity estimate in every bin carries more variance. A useful starting rule is to sample the narrowest feature with several angular points, then repeat the trace with a finer grid and a higher ray count. The distribution is sufficiently converged when the values that drive the requirement no longer move materially between runs. Wide distributions can tolerate coarser sampling; narrow beams, sharp cutoffs, and low-intensity wings require finer sampling and more rays. For a cutoff luminaire being evaluated against an angular template, the bins that matter most sit at the cutoff itself, and resolution there governs whether the file will show a compliant or a non-compliant fixture.
Far-Field Interpretation and the Point-Source Assumption
An IES or LDT distribution is a far-field representation: downstream software treats luminous intensity as a function of direction and applies distance and incidence geometry when calculating illuminance. A real luminaire has a finite luminous opening, so near-field measurements may not obey that approximation.
TracePro candela plots collect angular data from missed rays, rays exiting a selected surface, or rays incident on a selected surface. They do not require a virtual goniophotometer at an arbitrary radius. The critical simulation choices are the ray set, plot orientation, symmetry assumption, and angular sampling. When correlating to hardware, use the measurement distance and setup required by the applicable laboratory method, and document any limits on using the exported distribution close to an extended luminaire.
Source Models Set the Ceiling on File Accuracy
No amount of care in the export step recovers information the source model never contained. A photometric file generated from a Lambertian disk standing in for a phosphor converted LED will produce a plausible looking distribution that disagrees with hardware in exactly the places that matter, because the substitute source does not carry the real angular emission profile, the real emitting area, or the real spectral power distribution.
TracePro accepts measured ray data files from LED manufacturers, which encode position, direction, and flux for large numbers of individual rays sampled from a goniometric measurement of the physical part. For a fixture where the optic sits close to the emitter, and that includes most total internal reflection lenses, reflector cups, and edge lit light guides, the ray file is the difference between a usable model and a decorative one. Where a ray file is not available, a surface source with a measured angular distribution and the correct emitting geometry is the next best option, and the resulting file should be labeled internally as provisional.
Spectral content matters as much as angular content when the deliverable includes color metrics. TracePro traces wavelength resolved rays, so the same trace that produces the candela table can produce the spectral power distribution at the exit aperture. That spectral output supports color-coordinate and spectral analysis, and it is also what reveals color over angle variation, the mechanism behind the yellow ring or blue center artifacts that show up on a wall wash long before anyone measures them.
Materials, Surfaces, and Where Flux Actually Goes
A luminaire model that traces to the correct total flux for the wrong reasons will produce a photometric file that is right in aggregate and wrong in shape. The common failure is treating internal surfaces as ideal. A white painted reflector is not a perfect Lambertian diffuser at 96 percent reflectance across the visible band; it has a wavelength-dependent reflectance curve and a scatter distribution with a specular lobe. A specular aluminum reflector is not a perfect mirror; it has an angle-dependent reflectance and a surface roughness that broadens every reflected beam by a measurable amount.
TracePro applies bidirectional scatter distribution function models to surfaces, using measured data where it exists and parametric models such as ABg or Harvey Shack where it does not. The practical significance for photometric work is that scatter redistributes flux into the wings of the distribution. A fixture that meets a cutoff requirement in a model with ideal surfaces can fail the same requirement in hardware purely because scatter from a textured reflector pushes a fraction of a percent of total flux above the cutoff angle. Since cutoff and uplight criteria are evaluated against small absolute intensity values, a fraction of a percent is often the whole margin.
Property assignment should be audited before any file is generated. Flux accounting in TracePro reports how much flux is absorbed, how much escapes through each surface, and how much is lost to ray termination thresholds. A model with unexplained lost flux from ray termination is not ready for low-intensity wing or cutoff decisions, regardless of how good the central intensity looks.
Correlating Simulation Against Goniophotometer Measurement
Simulation and measurement should be treated as two instruments observing the same object, and the correlation exercise is what converts a design model into a model that can generate shippable files. Run the comparison on three quantities in order.
Start with total luminous flux. A simulated flux that is high by 8 percent against a sphere measurement usually points at an optimistic source flux, a missing absorbing surface, or an internal reflectance that was assumed rather than measured. Resolve this first, because every other comparison inherits the error.
Move to peak intensity and beam angle. If total flux agrees but peak intensity is high and the beam is narrow relative to measurement, the model is under scattering. Increase surface roughness or apply a measured BSDF to the reflector and retrace. If the peak is low and the beam is broad, the source emitting area in the model is likely larger than the physical emitter.
Finish with the wings and the far off axis region, comparing on a logarithmic intensity scale where a factor of two error is visible. This is where scatter models, baffle geometry, and housing leakage show up. Path sorting in TracePro answers the question the plot raises, because it separates the rays arriving at a given angular region by the sequence of surfaces they interacted with, which identifies whether unwanted intensity at 85 degrees came from the reflector edge, a gap at the lens seat, or a second bounce off the driver housing.
Once the three comparisons close, record the correlated model and the version of every material property it used. Subsequent variants of the same platform can then generate photometric files directly, with hardware measurement reserved for confirmation rather than discovery.
Photometric Quantities That Are Routinely Confused
Distribution files are misused often enough that the underlying quantities are worth stating precisely, because the errors show up in specifications and in customer complaints.
Luminous intensity, in candela, is luminous flux per unit solid angle in a direction. It is a property of the source and does not fall off with distance. It is what an IES file contains.
Illuminance, in lux, is luminous flux per unit area arriving at a surface. It falls off with the square of distance and with the cosine of incidence angle. It is what a light meter on a work plane reads and what a lighting layout tool computes from the IES file.
Luminance, in candela per square meter, is intensity per unit projected area of the emitting or reflecting surface. It is the quantity that correlates with perceived brightness and glare, and it is the quantity that governs discomfort glare ratings and display readability.
TracePro produces all three directly. Candela plots give intensity, illuminance maps on a receiving surface give lux, and luminance maps on the fixture itself give the brightness distribution a viewer sees. Confusing intensity with illuminance is the single most common cause of a specifier reporting that a fixture does not deliver its rated performance, when in fact the fixture is fine and the calculation distance was inside the far-field limit.
Using Photometric Files Before Hardware Exists
The commercial value of generating distribution files in simulation is that they arrive early enough to change decisions. A candela file exported from a design model at the concept stage can be loaded into a lighting layout tool and evaluated against the actual application, which answers the question that matters to a specifier rather than the question that matters to an optical engineer. A warehouse aisle layout will show immediately whether a proposed distribution delivers required average illuminance at the target spacing to mounting height ratio, and whether the uniformity ratio holds at the aisle ends.
That feedback loop is short enough to run several times before tooling is committed. A reflector profile change, a diffuser haze change, or a shift in LED pitch can each be evaluated as a photometric file rather than as an optical plot, which puts the trade study in the language the sales and specification side already speaks.
The same workflow supports regulatory pre screening. Roadway distributions can be checked against classification templates, and outdoor fixtures can be evaluated for uplight and backlight behavior, all from simulated candela data. None of this substitutes for accredited testing on a production article, and the files should be marked as simulated until measurement confirms them. It does mean that the accredited test becomes a confirmation step rather than the first time anyone finds out what the fixture does.
Conclusion
A photometric file is an engineering deliverable with a commercial audience, and treating it as a byproduct of testing gives up most of its value. Generating candela distributions from a correlated ray trace lets the optics group control the angular grid, understand where the wings come from, catch metadata problems before a qualification program does, and put a usable distribution in front of specifiers while the design can still change. The prerequisites are ordinary and specific: a source model built from measured ray data, surface properties with real scatter behavior, a far-field distance that respects the point source assumption, and a documented correlation against goniophotometer measurement.
Request a TracePro demonstration or start a free trial to build a validated photometric model and generate specification-grade distribution files before your first goniophotometer session.