Guide 02
Build the output first. Then choose how each frame is captured.
FPS controls delivery, shutter controls exposure and samples control motion-blur reconstruction. DepthOfField is a separate aperture-sampling workflow, so the mode must be chosen before the quality controls make sense.
The three controls that matter first
RenderFpsHow many finished frames exist each second
24 creates a familiar cinema cadence. It is independent of the frame rate GTA is running at while RE+ works.
RenderShutterHow much time is exposed inside each finished frame
0.5 means half the frame interval, which is the 180-degree convention. At 24 fps that equals 1/48 second.
RenderSamplesHow many real instants are averaged across that exposure
More samples make the trail continuous. They do not make the shutter longer or add more blur.
Shutter at 24 fps
| RE+ value | Angle | Exposure | Character |
|---|---|---|---|
0.25 | 90° | 1/96 sec | Crisp and urgent |
0.50 | 180° | 1/48 sec | Natural starting point |
0.65 | 234° | about 1/37 sec | Soft and expressive |
1.00 | 360° | 1/24 sec | Heavy trails |
A 180-degree shutter is 0.5. Typing 180 asks for an exposure many times longer than the frame interval.
Sliding or Walking
Let the clip keep living
The replay plays in controlled slow motion. Particles, TAA, SSR and temporal history continue between samples. Shutter placement is approximate.
Use it for most shots and shutters above 0.5.Seek every sample exactly
The replay pauses and jumps to exact sample times. It is deterministic, but temporal systems reset or arrive cold after each seek.
Use it for short shutters, exact timing or a broken Sliding result.Sliding frames/output ≈ samples ÷ shutterWalking = settle + samples × (1 + sub-settle)Need the beginner version? Read Sliding, Walking and Samples for a visual breakdown, worked examples and source links beside each claim.
The third mode: physical depth of field
RenderCaptureMode=DepthOfField was added at source commit d6c0a318. It uses Walking timing but asks the modified IGCS Connector to accumulate every output frame across a real aperture. RenderSamples is ignored; bokeh quality controls the aperture sampling, while RenderShutter still controls temporal exposure.
Every output frame becomes a full aperture pass that can take tens of seconds. Use it because the shot needs geometric defocus, not as a general higher-quality switch.
The lens controls, requirements, autofocus behavior and first-test workflow live in the dedicated Physical depth of field chapter.
Pick an editing master
| Output | Use it for | Trade-off |
|---|---|---|
prores_hq | Premiere editing and grading | Large MOV files, easy to decode |
Frames + PNG | Recovery, compositing and maximum fidelity | Thousands of files and high storage use |
h264 | Ready-to-watch delivery | 4:2:0 compression and less pleasant editing |
lossless | Archival or exact intermediate storage | Very large MKV output |
The included ProRes preset writes 10-bit 4:2:2 output. It does not create dynamic range that was absent from the captured game frame, but it avoids delivery-style chroma subsampling while you edit.
A dependable final baseline
RenderCaptureMode=Sliding
RenderFps=24
RenderSamples=48
RenderShutter=0.5
RenderHighlightBoost=0.9
RenderJpeg=0
RenderMode=Video
RenderVideoPreset=prores_hq
RenderAudio=0This is a starting point, not an instruction to make every shot identical. A drift impact may benefit from a shorter shutter. A slow memory sequence may benefit from 0.65.