Vol. 1 · No. 338Blender 3.6 / 4.2 / 4.5 / 5.0 / 5.2 / 5.3Cycles · EEVEE
JANCTION
Render · Blender on RTX PRO 6000 · Tokyo
0.1 JPYper GPU-secondFirst 100 JPYfree, no card

Case No. 2  Turntable

A 24-frame turntable of headphones

orbit=True circles the camera around the object and returns an MP4. No keyframes in the script.

● Turntable · CyclesRTX PRO 6000 · Tokyo
Cycles1280 × 720Samples 64MP4
Fig. — Turntable: headphones (24 frames, MP4). A real render from the production GPU.
The short answer

This 24-frame headphone turntable was rendered as an MP4 on a JANCTION GPU with orbit=True. The orbit camera circles every non-flat object and ignores the thin pedestal when it frames the shot, so the script needs no camera animation. It took 58.3 GPU seconds on the production GPU, 5.8 JPY at 0.1 JPY per GPU-second.

What is in the frame

A stylised pair of headphones on a low pedestal, seen from every side in one loop. Turntables are the standard way to show a product or a model in a portfolio, and they are the first thing people ask an AI to make once a model exists.

How the scene is built

The script contains the headphones and the pedestal, nothing else. orbit=True adds a camera that circles the scene; orbit_frames sets how many frames make up the turn. The pedestal is flat, so the framing logic ignores it and keeps the headphones filling the shot.

Adapt it

Swap in your own model: a glTF, FBX or USD file is imported into an empty scene, so the same call works without writing any Blender code. This one took 58.3 GPU seconds, about 5.8 JPY.

5.8 JPY for this render. The free welcome credit of a new key covers about 17 of them, with no card.

Render this example with your AI

ChatGPT · Claude · Grok · Perplexity

Use JANCTION Render (MCP server https://render.janction.jp/mcp) to render this example: render_preview(scene_url="https://render.janction.jp/samples/ex02_turntable_headphones.py"). Then suggest how to change the colors, shapes and camera for my own subject. Ask me before the final render.

Each link opens the app with the text below filled in (if it is not, copy and paste it). If JANCTION Render is not connected in that app yet, add https://render.janction.jp/mcp as a connector first.

Blender warnings

  • orbit: an orbit camera circles the scene once over 24 frames (elevation 18 deg); the scene's own camera is not used
  • environment preset 'studio' replaced the scene's world (HDRI lighting)

The script (bpy, runs as is)

ex02_turntable_headphones.py. It builds the scene, the camera and the frame range; resolution, samples, denoising and the GPU are set by the service.

"""JANCTION Render example 02: turntable of a stylized pair of headphones (orbit=true, MP4).

Reproduce: render_final(scene_url="https://render.janction.jp/samples/ex02_turntable_headphones.py",
                        orbit=True, orbit_frames=24, width=1280, height=720, samples=64,
                        environment="studio", output="mp4")
The orbit camera circles the non-flat objects; the thin pedestal is flat and is ignored for framing.
"""
import math

import bpy
from mathutils import Vector

for ob in list(bpy.data.objects):
    bpy.data.objects.remove(ob, do_unlink=True)
scene = bpy.context.scene
scene.frame_start = 1
scene.frame_end = 1


def material(name, color, roughness=0.5, metallic=0.0, **inputs):
    mat = bpy.data.materials.new(name)
    mat.use_nodes = True
    bsdf = mat.node_tree.nodes["Principled BSDF"]
    bsdf.inputs["Base Color"].default_value = (*color, 1.0)
    bsdf.inputs["Roughness"].default_value = roughness
    bsdf.inputs["Metallic"].default_value = metallic
    for key, value in inputs.items():
        bsdf.inputs[key].default_value = value
    return mat


def smooth(ob):
    for p in ob.data.polygons:
        p.use_smooth = True
    return ob


def add(op, name, mat=None, **kw):
    getattr(bpy.ops.mesh, op)(**kw)
    ob = bpy.context.object
    ob.name = name
    if mat is not None:
        ob.data.materials.append(mat)
    return ob


def look_at(ob, target):
    ob.rotation_euler = (Vector(target) - ob.location).to_track_quat("-Z", "Y").to_euler()


def camera(location, target, lens=50.0):
    data = bpy.data.cameras.new("Camera")
    data.lens = lens
    cam = bpy.data.objects.new("Camera", data)
    scene.collection.objects.link(cam)
    cam.location = location
    look_at(cam, target)
    scene.camera = cam
    return cam


def light(name, kind, location, energy, target=None, size=1.0, color=(1.0, 1.0, 1.0)):
    data = bpy.data.lights.new(name, kind)
    data.energy = energy
    data.color = color
    if kind == "AREA":
        data.size = size
    ob = bpy.data.objects.new(name, data)
    scene.collection.objects.link(ob)
    ob.location = location
    if target is not None:
        look_at(ob, target)
    return ob


plastic = material("Plastic", (0.07, 0.07, 0.08), roughness=0.45)
accent = material("Accent", (0.85, 0.62, 0.3), roughness=0.25, metallic=1.0)
cushion = material("Cushion", (0.11, 0.11, 0.12), roughness=0.95)
base = material("Pedestal", (0.75, 0.75, 0.73), roughness=0.5)

band = add("primitive_torus_add", "Headband", plastic, major_radius=1.0, minor_radius=0.075, major_segments=96,
           minor_segments=24, location=(0, 0, 0.9), rotation=(math.pi / 2, 0, 0))
smooth(band)
for sign, side in ((-1, "L"), (1, "R")):
    cup = add("primitive_cylinder_add", "Cup" + side, accent, radius=0.42, depth=0.26, vertices=96,
              location=(sign * 1.05, 0, 0.9), rotation=(0, math.pi / 2, 0))
    smooth(cup)
    pad = add("primitive_torus_add", "Pad" + side, cushion, major_radius=0.34, minor_radius=0.1, major_segments=96,
              minor_segments=24, location=(sign * 0.9, 0, 0.9), rotation=(0, math.pi / 2, 0))
    smooth(pad)
    arm = add("primitive_cylinder_add", "Arm" + side, plastic, radius=0.05, depth=0.5, vertices=32,
              location=(sign * 1.0, 0, 1.25), rotation=(0, sign * 0.35, 0))
    smooth(arm)
pedestal = add("primitive_cylinder_add", "Pedestal", base, radius=1.7, depth=0.04, vertices=128, location=(0, 0, 0.02))
smooth(pedestal)

light("Key", "AREA", (2.5, -3.0, 3.5), 300, target=(0, 0, 0.9), size=2.0)
camera((0.0, -6.0, 1.7), (0.0, 0.0, 0.9), lens=50)

Questions about this case

How do I make a 360-degree turntable video of a 3D model?

Call render_final with orbit=True, a frame count in orbit_frames and output="mp4". The camera circles the model automatically; the script or file needs no animation.

How much does a turntable cost to render?

A 24-frame 720p MP4 at 64 samples took about 59 GPU seconds on a simple scene, about 5.9 JPY at 0.1 JPY per GPU-second.

More from the Journal