Signal Veil
AI generated blind A B X listening bench with aligned loops, optional RMS matching and a transparent trial report.
Screenshots
Screenshots supplied by the seller.
Description
AI generated. Signal Veil compares two versions of a sound without exposing the identity of X until an answer is committed. Load two local files or explore three original generated pairs.
Set a shared region and B alignment offset, optionally match whole region RMS, and compare A, B and X with phase aligned looping sources and short switching fades. Both versions receive shared peak protection.
A trial log records guesses and revealed identities. The score includes a descriptive binomial chance tail under independent random guessing, with its assumptions clearly stated. Export a session report for review. No microphone, external analysis service or commercial audio is required.
Condition
Development stage: Mostly built (seller-declared)
Seller-declared: provided by the seller, not independently verified by SellMyVibe.
- Solo creator
- Mainly AI-generated
- Under one month in development
- Tests included
- Setup guide included
- Demo available
What works
- Two local stereo sources or original generated pairs
- Shared region selection and signed B alignment
- Optional whole region RMS matching
- Shared peak protection
- Phase aligned looping A and B sources
- Short fade A B X switching without transport restart
- Random hidden X identity per trial
- Committed answers with reveal and trial log
- Descriptive binomial chance tail calculation
- Session JSON export
- Alignment, matching, random choice and PCM tests
What’s unfinished
- No calibrated hearing or equipment quality evaluation
- No native plugin, live input or hardware audio routing
- Completed downloads and local file picker unverified
Included assets
Editable source code
Setup and usage documentation
Original sound definitions
Audio verification checks and notes
Real running screenshot
Technologies
HTML, CSS, JavaScript, Web Audio API
Setup requirements
Open index.html or serve the folder with any static HTTP server. Use a deliberate play action to enable audio.
Dependencies and ongoing costs
No third party libraries, paid APIs or required services. A current Web Audio capable browser is required. Optional static hosting can be free.
Known limitations
Standalone browser comparison tool. Two files up to 20 MB and 30 seconds each, decoded at 44100 Hz with the first two channels retained and mono duplicated. Region length is 1 to 12 seconds. B offset is limited to plus or minus 50 ms. RMS matching is not perceptual loudness normalization. Crossfades are 12 ms and can briefly combine correlated signals. Random identities use browser crypto where available. The binomial statistic assumes independent chance guesses and does not prove hearing or equipment performance. No automatic storage.
Listings are reviewed for completeness and clarity before publication. Approval is not a guarantee of quality, ownership, security or commercial success. Read the details and ask questions before buying.