D-I(1)

📦 version 7.44.0, in nixpkgs unstable

💅 d-i — the manual, but make it make sense

okay bestie, so this is d-i from the BRL-CAD fam (manual section 1, she's official ✨). Think of her as the little converter that takes your fancy double-precision floats and squishes them down into cute lil' 16-bit short integers. It's giving "glow-down but on purpose" 💖

✨ NAME ✨

d-i — converts doubles into 16-bit short integers. that's it, that's the whole job 🎀

🌸 SYNOPSIS 🌸

d-i [-n | scale] < doubles > shorts

🦋 DESCRIPTION 🦋

So here's the vibe: d-i reads a stream of native, machine-format, double-precision floating point numbers from standard input, multiplies every single value by a scale factor, and writes the results out to standard output as native 16-bit signed short integers. no cap, that's the entire pipeline 💫

The optional single argument is what picks your scaling, and you've got choices:

If you pass a numeric scale, every input value gets multiplied by it. If you pass -n, the scale is set to 32767.0, so input data sitting in the range -1.0 to +1.0 maps all the way onto the full short integer range. And if you pass no argument at all, the scale just defaults to 1.0. easy 💅

Now for the spicy part: values that exceed the representable short range get clipped to 32767 or -32768, and a count of the clipped high and low samples gets reported on standard error. so you'll know if your data got a little too extra 😭

⚠️ okay important: this command is deprecated and scheduled for removal — she literally prints a deprecation warning on startup. so appreciate her while she's here 🫶

She's the exact inverse of i-d(1), and she typically shows up in pipelines alongside other BRL-CAD signal-processing tools like dmod(1) and dconv(1). iconic teamwork honestly 👑

✨ OPTIONS ✨

  • -n — Normalize: use a scale of 32767.0 so that input in the range -1.0 to +1.0 fills the full 16-bit signed range. Mutually exclusive with a numeric scale argument, so pick one bestie, not both 🙈
  • scale — A floating point multiplier applied to every input sample before conversion. A scale of 0.0 is rejected as an error. sorry, zero, you're not invited 💀

🍓 EXAMPLES 🍓

Example 1. Normalize unit-range doubles to full-scale audio samples 🎧

d-i -n < signal.d > signal.pcm

This treats the values in signal.d as samples in the range -1.0 to +1.0 and rescales them to fill the 16-bit signed integer range, producing raw PCM-style short data in signal.pcm. literally her main character moment 🌟

Example 2. Apply an explicit gain in a processing pipeline 🔥

dmod -m 0.5 < input.d | d-i 100 > scaled.shorts

This multiplies the double-precision stream by 0.5 using dmod(1), then scales by 100 and converts to 16-bit shorts. Any samples that exceed the short range get clipped and a clip count is printed to standard error. pipeline queen behavior fr fr 💜

🌈 SEE ALSO 🌈

brlcad(1), i-d(1), d-a(1), d-u(1), dconv(1), dmod(1)

go read those too, they're all part of the squad 🫶

👑 AUTHOR 👑

BRL-CAD Team 💖

This software is Copyright (c) 1989-2026 by the United States Government as represented by U.S. Army Research Laboratory. yes the government owns this one, slay 🏛️✨

💌 BUG REPORTS 💌

Reports of bugs or problems should be submitted via electronic mail to devs@brlcad.org. tell them everything, they wanna know 🐛💅

and that's the whole page, bestie — brought to you by BRL-CAD, the D-I(1) manual page, no notes, periodt 🌸💖

📦 where this page comes from