Two exports with the same preset name landed 240 MB apart on my machine. I show the CBR and VBR arithmetic behind both, and how to predict a VBR size.
Two Renders, One Preset Name
The two files sat next to each other in the project folder with the same name apart from a suffix, the same ten minute runtime, the same presenter, the same slides. One weighed 624 MB. The other weighed 384 MB. I had rendered the second one a week after the first because a colleague asked for a copy that would fit a slower connection, and in between the two renders I had changed exactly one dropdown in the export panel, from constant bitrate to variable. Nothing else in the preset had moved.
The short answer: CBR writes the target bitrate for every second of the video. VBR writes what each second needs and averages out lower on content with still stretches, so two files with the same nominal settings differ by whatever the still scenes saved. On a talking-head lesson that saving came to 240 MB, about 38 percent of the file. This article is the working behind both numbers, and the test-export habit I use now so a VBR size stops being a surprise.
What Constant And Variable Actually Promise
A constant bitrate encoder makes one promise: every second gets the same ration, 8 Mbps in my preset, whether the frame shows a presenter mid-gesture or a slide that has not changed in ninety seconds. A variable bitrate encoder makes a different promise: the average will come out near the target, but each second gets what its picture costs. HandBrake's documentation describes the same behaviour for average bitrate mode, where the encoder varies the bits given to any portion of the video while holding the average, and recommends a second pass so the encoder learns the complex scenes before spending on them.
Video bitrate by minute in my two exports (Mbps)
Data: per-minute averages I read from the two finished files. The VBR line averages 4.8 Mbps across the ten minutes.
Read the VBR line as a map of the lesson. The dips at minutes three and six are slide explanations, where the frame barely changes and the encoder spends under 3 Mbps. The spike at minute four is a screen demonstration with scrolling and a moving cursor. The flat CBR line spends the same on all of it. Neither line is a defect; they are two different answers to the question of what a bitrate is for. I compare what each codec does with rates like these in The H.264, H.265 And AV1 Comparison, since codec and rate control multiply together.
How do you tell which promise a finished file kept? Read its average bitrate in the properties panel and compare it with the preset's target. My CBR file reports a total rate of 8.32 Mbps, the target plus its audio, because it never deviated for a second. The VBR file reports 5.12 Mbps in total. The gap between a reported average and a preset's headline is the saving, stated in one number, and it takes ten seconds to find on any file you already own.
The Maths Of My Two Exports
Both files came out of the same rule, bitrate times duration divided by eight, with the 320 kbps audio track included. The only difference is which bitrate the multiplication gets: the fixed one CBR promised, or the average VBR actually delivered.
Export
Video bitrate used in the maths
Size of 10 minutes
Size of 45 minutes
Constant bitrate
8 Mbps for every second
624 MB
2,808 MB
Variable bitrate
4.8 Mbps average
384 MB
1,728 MB
Difference
3.2 Mbps of average
240 MB
1,080 MB
A talking head does not need action-movie bitrate for the still parts. VBR knew that. My preset did not.
The use case that made this matter to me was not one lesson; it was a season of them. A client archive of weekly recordings priced at the CBR figure plans for 2,808 MB a session. Priced at the VBR average it plans for 1,728 MB, and across a season that gap decides whether one drive or two goes into the quote. That planning exercise, done for a full wedding season, is in The Season Storage Guide. The mistake to avoid is planning from the preset's headline rate when the encoder was set to vary it. Both totals in the table came out of the video file size calculator, run once at the fixed rate and once at the measured average.
Predict A VBR File Before The Full Export
Step 1: export a short sample. Render sixty seconds from the busiest part of the video and from a quiet part, with the exact VBR settings you plan to use. One minute from the middle of the source tells you more than any guess.
Step 2: read the sample's average bitrate. Open the sample's properties and note the data rate, or weigh the file and work backwards: megabytes times eight, divided by seconds, gives the total bitrate it actually used.
Step 3: run the full duration at that average. Put the average bitrate and the full runtime into the calculator. My sample predicted 384 MB for ten minutes at a 4.8 Mbps average, and the finished export landed on 384 MB.
Step 4: leave a margin for busier content. A sample from a quiet lesson underprices a highlight reel. If the full video mixes in motion the sample lacked, price the archive from the higher sample and keep the difference as headroom. On mixed projects I now take two samples as a matter of routine, one from the quietest minute and one from the busiest, and I budget the storage from the busy one.
When I Still Choose CBR
Variable bitrate is my default for finished files, and constant bitrate keeps two jobs in my workflow. The first is live streaming, where the platform ingest expects a steady rate and a VBR spike can stall the stream; encoding references such as Stream's glossary make the same split, VBR for stored files, CBR for live delivery. The second is any delivery with a hard size ceiling stated in advance, where I need to promise a number before the render starts.
Finished uploads and archives: VBR or constant quality, sized by a test export
Live streams: CBR, because the ingest wants a steady rate
Fixed-ceiling deliveries: CBR, or VBR with the average verified by a sample
Storage quotes: priced from measured averages, never from preset headlines
One more setting deserves a warning, because many editors place it beside the other two: constant quality, labelled CRF or RF in tools like HandBrake. It drops the bitrate target and holds a quality level instead, which makes its file size the least predictable of the three. Two episodes of one series at a single quality setting can land hundreds of megabytes apart, because a busy episode simply costs more at the same quality. I use it for personal archives with no ceiling over them. I never quote storage from it without a test export first.
The dropdown I changed by accident that week is now the first thing I check in any export panel. A preset name tells you the resolution and the target. The rate control line tells you what the file will weigh, and it is the only line that ever knew.
Questions People Ask About This Page
Why Do Two Exports Of The Same Video Have Different File Sizes?
Because one used constant bitrate and the other used variable bitrate, or a quality-based mode. A constant bitrate export writes the target rate for every second, simple or complex. A variable bitrate export spends fewer bits on still scenes and more on motion, so the average lands below the target and the file comes out smaller, on a talking-head video often 20 to 40 percent smaller.
Is VBR Worse Quality Than CBR?
Not at the same average size. Variable bitrate moves bits from easy scenes to hard ones, so motion keeps its detail while still frames stop wasting data. Constant bitrate can actually look worse in a busy scene, because its fixed budget has been partly spent on the quiet minutes before it. The trade runs the other way for predictability: only the constant bitrate file has a size you can state exactly in advance.
Can I Predict A VBR File Size Before Exporting?
Not exactly, and that is the honest limitation. You can estimate it well: find the average bitrate from a short test export or a finished file's properties, then multiply that average by the duration in seconds and divide by eight. My ten minute test export averaged 4.8 Mbps on video, which predicted 384 MB against the 384 MB the full export produced.
Should I Export CBR Or VBR For YouTube?
Use variable bitrate or a constant quality mode for uploads, because YouTube re-encodes everything and a quality-led file gives its encoder the most to work with at a sane size. Save constant bitrate for live streaming, where the platform ingest expects a steady rate and a spike can stall the stream.
Facts, Checked Against Trusted Sources
HandBrake's documentation explains that with an average bitrate setting the encoder varies the bits given to any one portion of a video while holding the average, and it recommends two-pass encoding so the encoder learns which scenes are complex before it allocates bits. Source: HandBrake Documentation, constant quality vs average bit rate
Stream's encoding glossary notes that variable bitrate adapts to scene complexity and produces smaller files, while constant bitrate keeps a fixed rate that suits live streaming and real-time delivery. Source: Stream, CBR vs VBR glossary
Written by 
Reviewed by 
Leave A Comment