Stereo Consumer Goods Assembly Data

It is first-person video of a person doing light assembly work on a bench: assembling consumer products from components, fitting plastic parts in a jig and fastening them with an electric screwdriver. It is recorded from their own point of view with two calibrated cameras, with hand pose and sensor streams alongside. Stereo describes the capture, and MCAP describes the container. They are separate choices.

The site does not state that stereo is already running on these rows. The catalog lists a head-mounted phone with a multiview rig as their capture setup, so a stereo pair is a capture choice agreed in the skill spec and confirmed on a Spec Pilot. The data is human capture by paid, consented contributors, not robot or teleoperation data. The site does not state product or brand names, cycle times, defect or yield rates, throughput, factory or customer names or tool brands, so this page promises none. Firsthand is not affiliated with Foxglove or the MCAP maintainers.

Three, all in the Plastic & Light-Assembly Manufacturer environment of the Manufacturing & Workshops sector. "Consumer goods assembly" is rated Medium and described as "Assemble consumer products from components on the bench, fasten and check each unit." "Assemble and fasten plastic components using electric screwdrivers and assembly jigs" is rated Easy and described as "Fit plastic parts in the jig and fasten them with an electric screwdriver." "Light Manufacturing & Assembly" is rated Easy and described as "Assemble, inspect and pack light consumer goods on a bench or moving line."

All three are listed as Ready to deploy, which the catalog defines as contributors and site access already set up. The site says nothing about delivered hours or delivered stereo footage for these rows, so this page claims none. The environment’s example line reads "Consumer goods assembly, packaging operations, labeling, quality inspection on a production line". The only tools the rows name are an electric screwdriver and an assembly jig, and the first table below reproduces the rows as the catalog states them.

The egocentric light assembly manufacturing page covers all 15 rows of the same environment, including these three, with egocentric catalog treatment and no stereo or MCAP treatment. The stereo capture page for hardware and light-assembly plants covers four plant types, door and window hardware, electronics assembly, lighting products and furniture. The assembly task data page covers bench assembly in general, such as fastening, measuring, press-fit and wire work.

This page is the stereo and MCAP treatment of the three consumer goods and jig fastening rows, with its own stream map and failure modes. The comparison table below sets the pages side by side.

Bench assembly is about where a part sits relative to the jig, what the hands do to seat it and how the tool meets the fastener. A stereo pair recovers depth from the calibrated frame pair rather than from a standalone depth sensor, so the part, the jig, the driver and the hands share a frame. The site gives no depth accuracy figure for these rows and no label for whether a part seated or a unit passed, so what a buyer does with that depth is their pipeline question.

A stereo pair adds depth, not a verdict on the unit. The site states 1080p per eye for dataset 06 and gives no figure for how well small parts such as screws resolve, so that is Not stated and a spec should check it on a Spec Pilot. The site also states that Firsthand does not capture at a customer’s own facility.

The site states 3D hand pose of 21 joints per hand with a per-joint visibility flag, and contact and grasp events that mark when fingers meet a tool or an item. It describes the two-handed structure of bench work as one hand steadying the work while the other places, presses or fastens, and the visibility flag is what carries it when a hand closes around a part or a tool.

The site does not say which hand holds the part and which holds the driver on these rows, how finely hand pose resolves small parts, or whether the part, the jig or the driver counts as the object for contact events. Settle those in the spec. Wrist cameras give a closer view of the hands, and a multiview rig adds synchronized third-person angles when the spec asks for them.

The site describes two stereo setups and does not say which one a consumer goods assembly order uses. The hard-environments stack is a global-shutter 4K/60 head camera with a 120° lens, two 1080p/60 wrist cameras, a depth sensor at 848×480 and 30 fps, and 200 Hz 6-DoF IMUs. The off-the-shelf stereo head rig, dataset 06, is dual fisheye cameras at 1080p per eye, a 62 mm baseline, per-unit calibration, a 9-axis IMU and hand-tracking output.

Which rig, resolution and streams an order uses is Not stated on the site, and the third table below sets what is stated beside what is not. Intrinsics and extrinsics are re-solved at the start and end of every session, and streams are hardware-triggered onto one clock with a 2.0 ms cross-stream ceiling. Bench layout, lighting, jig design, tool model and camera clearance are Not stated, so any requirement is settled in the spec.

MCAP is an open source container file format for multimodal log data, supporting multiple channels of timestamped pre-serialized data. It is serialization-agnostic, so messages can be Protobuf, DDS (CDR), ROS, JSON and more. Its design is row-oriented and append-only, it stores schemas alongside the data, and it offers an optional index plus optional LZ4 or Zstandard chunk compression.

Reader and writer libraries exist for C++, Go, Python, Rust, Swift and TypeScript, and the mcap CLI can inspect, merge and split files. The foxglove/mcap repository is MIT-licensed; that is the MCAP project licence, not the licence of a Firsthand delivery. The MCAP hub page and the glossary entry cover the format in full.

Firsthand offers MCAP delivery, and dataset 06 is stated as shipping as MCAP. For these rows the site states hand pose, a stereo-derived depth stream with confidence maps, IMU, contact events, action segments, per-frame instance masks and one task instruction per episode elsewhere. What it does not state is which of those are channels in an MCAP file, how they are named, what encoding they use or at what rate.

No topic name, message schema or rate is stated on this page, and none should be assumed. Name the streams that matter in the skill spec and confirm them on a Spec Pilot batch. The fourth table below sets each stream beside what the site states about it.

Small parts and screws are the first risk. The site gives no figure for how finely hand pose resolves very small parts, and it does not describe how depth behaves on small screw heads, so those are Not stated and should be checked on a Spec Pilot. The site describes small parts held close to the lens as a place where a single camera loses information.

Jig alignment is the second. The row names an assembly jig, but the site states no jig geometry, alignment tolerance or label for whether a part seated, and the site lists no label for the outcome of a step. Action segments carry start, end, verb and noun, so a spec that names each step gives a sequence that can be sliced by step.

Tool-in-hand occlusion and two-handed part holding are the third. A driver, a part or the other hand can hide the fingers that hold a component. The visibility flag on each joint is the answer the site gives, contact events mark when a hand takes hold of a part or a tool, and a spec can ask for wrist views or a second angle. An episode is rejected when the manipulated object is out of frame for more than 15 percent of it.

Glossy plastic surfaces and close working distance close the list. The site describes glossy or dark surfaces as places where a single camera loses information, and it states that depth degrades on wet and reflective surfaces and ships per-frame confidence maps so a buyer can mask rather than guess. How a stereo pair behaves on glossy plastic is Not stated. Distance-invalid pixels beyond the depth sensor reliable range are flagged rather than interpolated, but the site does not state how depth behaves at the close range of a bench. An episode is rejected when more than 35 percent of depth pixels in the manipulation region are invalid.

Calibration is measured at the start and end of every session under a 0.28 px RMS gate, an episode is rejected when an undocumented bystander appears, and a reject log ships with every batch. Product or brand names, cycle times, defect or yield rates, throughput, factory or customer names, tool brands and safety certifications are Not stated and are not claimed here.

For every delivered episode the site states that faces, plates and screens are blurred irreversibly, and that footage that cannot be de-identified is discarded. The anonymization record ships with the data, and an undocumented bystander is a stated reject criterion.

The site does not state a rule specific to a production bench, product labels or packaging printed with branding, so this page makes no such claim. Settle it in the spec.

The October 2026 off-the-shelf catalog lists ~324,800 hours across 18 datasets, of which stereo is ~64,800 hours across 9 datasets. Its Assembly tag covers appliance, LED, carton and electronics assembly, plus fitting and fastening. One stereo dataset, S5 (~2,000 hours, head stereo with two wrist cameras), lists "Manufacturing across 50+ factories: appliance and LED assembly, converter lines, packaging, fastening, tool use". One mono dataset, M3 (~50,000 hours), lists "Electronics and components: handling, assembly and repair".

No listed dataset names consumer goods assembly, and the site does not state that any of them contains the jig and electric screwdriver task in these rows. Stereo consumer goods assembly in MCAP is therefore custom light manufacturing data collection: custom collection at scale through a network of thousands of capture providers, for a consistent, recurring data flow to labs. Where coverage is deep, the median time from signed spec to first validated batch is a median of 48 hours; where it is thin, turnaround runs three to four weeks. The site does not say which speed applies to this order, so the timeline is quoted against your spec. Collections are billed per validated hour, and rejected episodes and recaptures are never billed.

Describe the steps and where each starts and ends: fitting a part in the jig, fastening it with the electric screwdriver, checking the unit. Name the product categories you expect to see, whether an episode covers one unit or a longer stretch of work, what counts as a reject and the variation your pipeline is missing. Then name the streams you want in the file, the topic naming and message encoding your reader expects, whether calibration ships inside the file or beside it, and whether MCAP is the only delivery or sits beside RLDS, LeRobot or WebDataset.

Most programs begin with a Spec Pilot, a small first batch that proves the spec and accept criteria on real footage before you scale. The fifth table below shows what each spec element decides.