SERVICES
From study design to verifiable delivery
Five service modules form one data chain: design and collection decide whether the data is usable, processing and annotation decide how finely it can answer a question, and validation plus compliance decide whether the conclusion can be cited.
- Service modules
- 5 modules
- Verified corpus
- 12,400 hours
- Median sync error
- 1.4 ms
- Annotator agreement
- 0.91 κ
01 / Scope
One team across the data lifecycle
From devices and venues to annotation systems and licensing agreements, one team owns the whole path, which removes the loss of standards that comes from handing work between vendors.
Lab-grade multi-stream synchronized capture
Field and real-world collection
Remote and crowdsourced collection
Multilingual participant recruitment
Transcription and forced alignment
Annotation scheme design and management
De-identification and privacy engineering
Verification reports and reproducibility audits
02 / Service modules
Five modules, one complete data chain
Choose a single module for one stage of the study, or hand the whole chain to one team. Each module works as a standalone delivery and can also plug into an existing data workflow.
Multimodal data collection
LAB · FIELD · REMOTE
From lab-grade multi-stream capture to field, home and remote settings, we design the collection plan around the research question, then handle recruitment, consent and on-site quality control.
- Collection plan, variable list and equipment roster
- Participant recruitment materials and consent documents
- Synchronized multi-stream raw data and backups
Processing and alignment
SYNC · CLEAN · ALIGN
Map every stream onto one time base, then run denoising, segmentation and forced alignment to produce standardized data that can be re-run and verified.
- Time-base alignment results and synchronization report
- Denoised, channel-corrected and converted media
- Forced alignment, segmentation and annotation intermediates
Annotation management
SCHEME · TRAIN · REVIEW
Start from the label set and annotation manual, train annotators, run double-blind annotation with adjudication, and deliver annotation data with an agreement report.
- Annotation manual and label set
- Annotator training and certification records
- Double-blind annotation and adjudication results
Validation and QA
SYNC · CONSISTENCY · AUDIT
Independent verification of data and annotations: sync measurement, content sampling, agreement assessment and reproducibility audits, delivered as a traceable quality report.
- Sync and signal quality report
- Content verification and sampling records
- Agreement assessment results (κ / α)
Privacy and ethics
CONSENT · DE-IDENTIFY · LICENSE
Build consent, de-identification, access control and licensing scope into the workflow, so sensitive corpora are delivered with an auditable compliance trail.
- Ethics review materials and consent drafts
- De-identification procedure and operation records
- Data licensing and usage notes
03 / Delivery flow
A shared delivery flow with a predictable rhythm
A typical 12-week project, stage by stage. Every stage has defined outputs and acceptance points; delays usually come from unclear criteria, not from execution speed.
- 01W1
Study design
Settle the question, variables, coding manual and statistical plan with the research team.
- 02W2–W4
Collection
Capture synchronized multi-stream data or connect existing data, with per-session QC.
- 03W5
Processing and alignment
Unify the time base, denoise and segment, then output standardized intermediates.
- 04W6–W9
Annotation and review
Train annotators on the manual, run double-blind annotation and adjudicate disagreements.
- 05W10
Validation and QA
Check sync, content and agreement; failed batches are reworked, never shipped.
- 06W11–W12
Delivery and audit
Hand over data, scripts, data dictionary and the verification report as one package.
Timeline varies with data volume, setting and ethics approval · Typical schedule above, final dates confirmed in the plan
04 / Why TOMCUBE
Verification built into the flow, not left to acceptance
The risk in multimodal data is rarely volume. It sits in alignment, annotation criteria and the licensing chain, and those three decide whether the data can be cited.
Design before collection
Variables, coding manual and statistical plan are fixed before collection, so the data points at the research conclusion from day one instead of being explained afterwards.
Verification in every stage
Sync, content and agreement checks are distributed across design, collection, processing and annotation, rather than saved for a final sampling pass.
Deliveries that reproduce
Data, metadata, processing scripts and the verification report ship together. A different person on a different machine can follow the documentation and get the same result.
An auditable compliance chain
Consent, de-identification, access control and licensing scope are each logged, so ethics and licensing status can be checked by the institution or an auditor at any time.
ISO 27001 Information Security Management · ISO 20252 Research & Data Services · IRB / research ethics collaboration
Start a project
Bring us the research question, from design to verification
Tell us what you need to answer, what data already exists and your timeline. We will come back within one business day with an executable plan outline and a budget range.
Reply within one business day · NDA and data licensing supported · Dedicated academic track