VEED Fabric audio must be under 10 MB: stereo wav tops out near 56 s

Sume's catalog says VEED Fabric 1.0 accepts a maximum of 300 seconds of audio, and the audio must be Sume-hosted and under 10 MB. A 44.1 kHz stereo 16-bit wav reaches 10 MB at about 56 seconds, so it fails the size limit long before the duration limit. An mp3 at 128 kbps reaches 10 MB only at about 625 seconds, so the 300-second duration cap is the one that binds.

The catalog was read on 2026-10-09. Fabric is priced at $0.1875 per audio second at 720p and $0.10 at 480p.

Bytes per second by format

Uncompressed wav size is sample rate x bytes per sample x channels. At 16 bits, each sample is 2 bytes. That gives 88,200 bytes per second for 44.1 kHz mono and 176,400 for stereo. A 16 kHz mono wav is 32,000 bytes per second. An mp3 at 128 kbps is 16,000 bytes per second. I use 10 MB as 10,000,000 bytes; if the limit counts 10 MiB, add about 5 percent to each figure.

How to get a smaller file inside Sume

Audio detach lets you choose format, channels and sample rate when you pull audio out of a Sume-hosted video: wav (the default) or mp3 at 128 kbps, mono or source channels, and 16,000, 44,100 or 48,000 Hz. A 16 kHz mono wav or an mp3 keeps a full 300 seconds well under 10 MB. Detach costs a flat $0.01 per job.

If the audio comes from Sume TTS, choose the output_format when you generate it. TTS defaults to mp3 at 44.1 kHz and 128 kbps, which is small enough for the limit at any length Fabric allows. The wav format is only needed when you want sentence slices.

What this costs at 300 seconds

Fabric reserves ceil of the audio seconds times the per-second rate. For a 300-second clip at 720p that is 300 x $0.1875 = $56.25, and at 480p it is 300 x $0.10 = $30.00. Convert a too-large wav before you submit instead of finding out at admission.

For a 20-second reply, a stereo wav at 44.1 kHz is about 3.5 MB, and is fine. The limit matters for long takes and for batch work where nobody checks each file.

• Under 56 seconds: any format fits.
• 56 to 300 seconds: mono wav at 44.1 kHz fits to about 113 seconds, so use mp3 or 16 kHz beyond that.
• Always check the file size before submitting a long clip.

A quick check before you submit

Compute the expected size from the audio you are about to send: duration in seconds times the bytes per second in the table. If the number is near 10,000,000, convert. Your own storage tool will also show the size of the file on disk, which is a more reliable check than arithmetic if the source has an odd bit depth.

Fabric also has a minimum to respect in the sibling MiniMax H3 Max Lip Sync route: its audio must be between 5 and 14.8 seconds. For clips that short, file size is never a problem, since even a stereo 44.1 kHz wav of 14.8 seconds is about 2.6 MB.