For AI teams and data vendors
A Los Angeles Production Partner for Speech Data Collection
MIX Recording Studio provides a dedicated capture facility for speech and conversational data collection projects and programs at 539 S Rampart Blvd, Los Angeles. Your team supplies the requirements or protocol, the scripts or prompts, and the participant consent form or release. We provide the room, the recording chain, an engineer on every session, and recurring scheduling for the life of the project. If you need voice talent (VTs), we recruit, manage and schedule them in Los Angeles. Every project is scoped and quoted individually.
MIX records speech and voice data for AI labs, enterprise voice and AI programs at major technology companies, Fortune 500 brands, and the AI data vendors that manage collection on their behalf.
The room
MIX I, a dedicated tracking room at 539 S Rampart Blvd. Acoustically treated, with room for a multi-speaker session and the engineer.
The chain
Neumann U87 and TLM 103 into a Universal Audio Apollo 8/x8, recorded in Pro Tools.
The files
24-bit WAV at 48 kHz by default (44.1 kHz on request), one track per speaker, raw takes unless your spec says otherwise.
How a Session Runs
An engineer sits in on every session. For two-person conversation work, each speaker gets their own microphone on their own input, so a buyer receives clean separated channels rather than one mixed stereo file. The engineer sets levels before the first prompt, watches for clipping and handling noise, keeps the session log and metadata, and stops for a retake when a technical fault spoils a take. The engineer conducts each session to your written requirements and guidelines; content and prompting follow your protocol.
MIX I is acoustically treated for low room tone and a low ambient noise floor, the same standard our ACX audiobook sessions are recorded to. Measured through the full signal chain, our room tone sits at about -66 dBFS RMS, comfortably under the -60 dBFS most data specifications ask for. If your team prefers to listen in or direct live, remote monitoring over Source-Connect or ipDTL is available as an option.
Getting Started
Share your requirements or a project summary: speakers per session, required sample rate and file format, total hours and target dates. We reply with a production timeline and a quote. Before booking we can send room-tone and test files, and a pilot runs before full production. The facility has on-site parking at 539 S Rampart Blvd, with direct access from the 101 and the 10.
- ScopeProtocol, speakers per session, format, hours and dates.
- Test files and pilotRoom-tone and sample files, then a short pilot.
- Scheduled productionReserved weekly blocks, an engineer on every session.
- DeliveryRaw WAV per speaker, session log and metadata, checked and sent via your secure upload.
Reserved Capacity and Scheduling
MIX operates 24 hours a day, 7 days a week. Project hours are reserved before they are released to other clients, so each session starts on schedule in a room already configured to your protocol. Blocks recur weekly for the duration of the project, and evening and weekend capacity can be added to meet quota.
Participants, Consent and Paperwork
Participant and voice talent recruiting in Los Angeles is available on request, or we work from your roster. At check-in we collect signatures on your consent form or release, verify each participant's ID against it, and record attendance in the session log. Vendor onboarding is straightforward: we provide a W-9, execute your NDA, and complete your vendor questionnaire. Project details are kept confidential.
Questions Program and Project Managers Ask
Can a data vendor book MIX Recording Studio as a Los Angeles capture facility?
Yes. Data collection projects book MIX I as a dedicated capture room with an engineer on every session. You supply the requirements and the participant consent form; we supply the room, the recording chain and the schedule. Each project is quoted individually.
What audio format does MIX deliver for speech data?
24-bit WAV at 48 kHz by default or 44.1 kHz on request, with one track per speaker for multi-speaker sessions. Takes are delivered raw unless your specification asks for processing.
Can the data be exclusive to our team?
Yes. Projects can be collected on an exclusive basis, with every recording and its rights delivered to you alone, or on a non-exclusive basis, where the same recordings may also be licensed to other buyers and the project is typically priced lower. Exclusivity is set in the project scope and reflected in the participant consent form.
Can we hear the room before we commit to a project?
Yes, without travelling. We send our room-tone and impulse measurement files as 48 kHz / 24-bit WAV, with room tone (noise floor) at about -66 dBFS RMS against the -60 dBFS most specifications require, and can record a short sample to your specification. A pilot can run before full production. Site visits are welcome but not required.
Can our researchers direct sessions without travelling to Los Angeles?
They do not need to. Most projects run entirely from your written requirements and guidelines, with our engineer conducting each session. If your team wants to listen in or direct live, remote monitoring over Source-Connect or ipDTL is available as an option.
Does MIX recruit the participants?
On request. We recruit participants and professional voice talent in Los Angeles, or work from your existing roster; in either case every participant signs your consent form at check-in.
Scope Your Project
Share the project basics: your organization, the data type, speakers per session, estimated studio hours and target dates. We reply with a production timeline and a quote.
- A production timeline and quote
- Room-tone and test files on request
- Your NDA executed and vendor questionnaire completed
Recording voiceover rather than data? See our voiceover recording studio.