MP3 bitrate explained: 128, 192, and 320 kbps
Bitrate is how many bits the encoder spends on each second of audio. OfflineMP3 does not ask you to pick one. Every MP3 it writes is 192 kbps, constant bitrate, 48 kHz, MPEG-1 Layer III. This page explains that choice next to the two rates people usually compare it with.
What the number measures
A kilobit is 1,000 bits. A rate of 192 kbps is 192,000 bits each second, which is 24,000 bytes each second. Over a minute that is 1,440,000 bytes, about 1.37 MiB, before the few extra bytes a padded frame can add. 128 kbps is about 0.92 MiB per minute. 320 kbps is about 2.29 MiB per minute. Those figures describe the audio stream. They are not a promise about how large the video you started from was, because the video also contained pictures.
The rate is not a quality score printed by a lab, and it is not a hearing test. It is a budget. A larger budget gives the encoder more room to represent the sound. It does not turn a lossy format into a sample-for-sample copy of the decoded audio. Two files at the same rate can still sound different if the encoder, the source, or the music is different. This site does not publish listening-test scores.
Constant bitrate, and the frame the page checks
Constant bitrate means each frame is aimed at the same rate. Variable bitrate would let quiet passages spend less and busy passages spend more. OfflineMP3 does not write a variable-bitrate file. The verifier walks every MPEG frame header. A frame has to be MPEG-1 Layer III, the sample rate has to be 48 kHz, and the bitrate index has to be the 192 kbps slot. A single frame at another rate fails the file, and the result is not offered for save.
An MPEG-1 Layer III frame holds 1,152 samples. At 48 kHz that frame covers 24 milliseconds. At 192 kbps and 48 kHz the frame is 576 bytes when the padding bit is off, and 577 bytes when it is on. The page uses that size to know where the next frame starts. If the bytes run out in the middle of a frame, verification fails. This is why "the encoder finished" is not yet "you can save".
128, 192, and 320
128 kbps is a common rate for speech and for older music files, where a smaller download mattered more than the last bit of high-frequency detail. Cymbals, reverb tails, and two voices at once are where a tight budget is easiest to hear. It is a real, widely used rate. It is not the rate this tool writes.
192 kbps is the rate OfflineMP3 always writes, for both mono and stereo. It is a middle setting in the MPEG-1 Layer III table, which also contains 128 and 320. There is no slider because the product currently has one output contract, and the verifier rejects anything else. Stereo at 192 kbps has the same total budget as stereo at 192 kbps from any other constant-bitrate encoder. It does not get a private boost from running in the browser.
320 kbps is the top of that MPEG-1 Layer III table. It spends more bits, and the file is larger. It is still lossy. Moving from 192 to 320 does not restore sound that an earlier AAC or Opus encode already discarded, and this tool will not make that file for you. If you need a different rate, you need a different encoder. OfflineMP3 will not quietly substitute one.
Why 48 kHz is a separate knob
Sample rate is how many samples exist each second, not how many bits describe them. The output is 48,000 samples each second per channel. A CD-style 44.1 kHz file is a different sample rate. This converter does not offer 44.1 kHz. Inspection accepts a source track at whatever positive sample rate the file declares, then the encoder resamples to 48 kHz on the way to the MP3. The output check requires the string 48000. A file that somehow came out at another rate would be rejected.
Channel count is separate again. A mono MP3 at 192 kbps spends the whole budget on one channel. A stereo MP3 at 192 kbps splits that same total budget across two channels. The tool does not raise the rate when the source is stereo, and it does not lower it for mono. A 5.1 source becomes a stereo MP3 at the same 192 kbps after the downmix. The low-frequency effects channel is omitted before that, so it does not consume bits in the MP3.
What a higher rate does not fix
Bitrate does not make an unsupported file supported. A video with AC-3, FLAC, Vorbis, MP3 audio inside the video, or AAC Main still fails inspection, whatever rate you might have wanted. Bitrate does not add a soundtrack to a silent file. A container with no audio track is refused before the encoder starts.
Bitrate also does not skip verification. A perfect-looking 192 kbps stream can still fail the sample-count check or the decoded-duration check described in the troubleshooting guide. If that happens you do not get a 192 kbps file to keep. The rate is a property of files that pass, not a way to force a pass.
If you are choosing between 128, 192, and 320 for some other project, decide from the source you have and the size you can store. On this site the decision is already made: 192 kbps constant bitrate, 48 kHz, one track, lossy.
Read next
How the encode fits into the task: MP4, MOV, and WebM to MP3.
What lossy leaves behind: Lossy and lossless audio.
Short answers: FAQ.