How do I know a file is encoded?

The tool cannot tell, so the file's origin has to. Anything exported by the mid side encoder or a mid and side processor is encoded with the mid on the left and the side on the right; an ordinary stereo file is not, and decoding it would produce a wrong result that happens to sound plausible.

Is the round trip really lossless?

Yes, provided nothing destructive has been done to the components in between. The two transformations invert each other exactly, so a file that is encoded and decoded without other processing comes back sample for sample.

What happens if I decode an ordinary stereo file?

The arithmetic still runs, but the assumption behind it is false, so the result is not the original recording. The two output channels are built from a sum and a difference that were never a sum and a difference, which produces a narrower and slightly strange image.

Why is the pair split into two tools?

Because most of the use for the transformation is in working on the components between the two steps, and separate tools make that workflow explicit. When the adjustment and the recombination belong together, the mid side processor offers them in a single pass.

Can I use the decoder to widen a mix?

Only if the components are adjusted between the two steps, since the decoder on its own restores exactly what the encoder produced. Scaling the side component before decoding is how width is changed in this representation, and the widener does the same arithmetic without the intermediate file.

Mid-Side Decoder

You Might Also Need

Mid-Side Encoder

Encode stereo to Mid/Side format

Stereo Widener

Enhance stereo field width

Mid/Side Processor

Process mid and side separately

Channel Mixer

Mix and balance audio channels

Mid-Side Decoder turns an encoded pair back into ordinary left and right stereo. It assumes the left channel is the sum and the right is the difference, and rebuilds the original channels from them.

What is Mid-Side Decoder?

Drop an encoded file and process it; the output is standard stereo left and right, restored from the mid and side components.

What the tool does:

It treats the first channel as the mid, which is the sum of the original sides, and the second as the side, which is half their difference.
It reconstructs the two output channels by adding and subtracting the two inputs, restoring each side of the original stereo image.
Both input channels are required, since the reconstruction needs both components.
A file exported by the encoder decodes back sample for sample, which is the property the pair was designed around.
Nothing is added, removed or scaled in the process.

The reconstruction is the exact inverse of the encoding, and the reason it works is worth a sentence because it explains why the format is lossless. The encoder described the original pair by their average and half their difference; the decoder adds the difference to the average to recover one channel, and subtracts it to recover the other, so every sample of the original is reproduced precisely from the two components. That is why a file can be encoded, worked on as components, and decoded back without any loss as long as the processing applied to the components was arithmetic rather than destructive. The tool makes one assumption that has to be respected: it takes the left channel as the mid and the right as the side, which is the convention the encoder writes, and it has no way to detect whether a file really is encoded. Feeding it an ordinary stereo recording produces a plausible but wrong result, since the two channels of an ordinary mix are not a sum and a difference, so the transform is only meaningful on material that went through an encoder or was produced by a mid and side processor. The output is indistinguishable from the original stereo when the pairing is correct, which is the best evidence that the round trip worked.

What it does not do:

No detection of whether the input is actually encoded.
No adjustment of the components before decoding.
No stereo width control, which is the widener's job.
No handling of a mono input.
No record of the transformation is stored in the file.
Output is sixteen bit WAV.

Use it to restore stereo after working on the components as audio: after compressing the mid to control the centre, after equalising the side to change the width's tone, or after any other processing that was easier to apply to the two components than to the two channels. When the components need to be adjusted and recombined in one operation, the mid side processor is the tool that does both, and when only the width needs changing, the stereo widener does that without going through an encoded stage at all. Keeping the two components in separate files during a session is a common working habit, since it makes clear which representation is being handled and prevents an encoded file being mistaken for ordinary stereo later.

How to use Mid-Side Decoder

Drop a file that was produced by the mid side encoder or a mid side processor.
Confirm the convention: the first channel is the mid and the second is the side.
Process it to rebuild ordinary left and right channels.
Compare the result against the original stereo to confirm the round trip was exact.

When to use Mid-Side Decoder vs related tools

Mid-Side Decoder is the tool for the second half of a mid and side workflow, once the components have been processed as audio. Mid-Side Encoder produces the components in the first place, and Mid-Side Processor handles both directions with controls for each part. When the aim is simply a wider or narrower image, Stereo Widener reaches the result in one step.

Privacy & Security

This tool runs entirely in your browser — no data ever leaves your device. There is no server round-trip, no upload, no logging, and no account required. Your input is processed locally using client-side JavaScript and is never stored, transmitted, or accessible to anyone else. When you close the tab, everything disappears.