Hey everyone,

I'm introducing my very last SaaS, or maybe I should say more set of tools, b10.studio ! When reposting videos or photos on social medias, you often get limited on reach, ban, shadow-ban or flagged as a reused activity. Most of the content spoofer you'll see online propose you to change some pixels and clear the metadata. But that often create poor quality videos, with abnormal exif metadata that gets instantly flagged on social medias such as Instagram, X, TikTok or Threads.

Where b10 excels, where others fail

The goal of b10 is to create very high quality spoofs, visually as well as technically, spoofing and repacking into a coherent file with USA (or your preferred location) recent iPhone fingerprint.

It has been tested on multiple detection algorithms: perceptual hashing like PDQ/vPDQ, content recognition, audio ACR. About the quality also, changing pixel is kind of dangerous for your reach because the platform will see your content as a low quality visual and never make you reach your engagement and views goal. We developped a very powerful content analyzer using the same technics as the big social medias, in order to never pass the line of the degraded media. Bypassing duplicate-detection by keeping the quality of the content, that is the real miracle where b10 excels.

Go deep, choose your customized preset

For people that really need deep content spoofing, we also have higher tiers including adversarial ML that runs on GPUs. You can use that to spoof very high impact content, sur as your creatives on Meta Ads. For other presets, b10 applies a randomized combination of transforms like visual filters, audio, metadata output, timing change, speed change etc. Each photo/video that get out of the spoofer is a completely unique content.

The technical details

Before any spoof, we'll apply a cropping of the dead zones, in order to work on a clean file. We are gonna then apply the spoof.

1. Visual filters

2. Geometric transforms: Zoom, rotation, lens correction, perspective, aspect ratio, pixel shift, horizontal flip, wave warp — then recrop/rescale to hide the black edges and restore the original dimensions

3. Audio We'll apply another layer of spoofing, in order to break the audio fingerprint (anti-ACR): - Pitch shift that preserves duration (so it can't desync from the video) - Subtle volume gain - Spectral perturbation: highpass/lowpass, EQ nudge at a random band, low-shelf bass - all below the audible floor Even shazam will not recognize your Dua Lipa song!

4. Temporal clipping That is also where b10 crashes all the competitors: taken frame by frame, the input and the output video will be completely different. Because we apply an edit, clipping filter in order to modify the speed, trim the head/tail and audio.

5. Authentic iPhone re-encode This is also a very powerful feature: b10 doesn't scream like "IT'S A SPOOFED FFMPEG CONTENT". The output looks like a real iPhone capture. For that, we apply things like: - .mov output with the qt (QuickTime) brand (or jpeg for images) - CRF, audio bitrate, FPS (30/29.97) randomized per variant - Correct bt709 colorspace metadata - +bitexact to strip the Lavf60

6. Metadata (spoof + post-process) To finish, we strip all the original metadata, for our iPhone fingerprint to stamp again your content with a coherent identity (device/iOS/GPS/date/name) created via exiftool.

It's not only about spoofing

At b10, we propose a complete and global experience. You'll find other tools on the suite, such as: - Image and Video File Converter - Risk and Quality Analyzer - File Reducer with 3 different level of shriking - Frame extractor, in order to work on your hooks or AI content - Audio extractor, to reuse and edit videos with brand new songs - Share a file, to be able to work smoothly with your team and virtual assistants.

See you soon on b10.studio