GNU Radio, the “Stream to Vector” block

In GNU Radio, the “Stream to Vector” block works with streamed sample data and groups it into fixed-length vectors.

Here’s a clear breakdown of how it works:

---

📡 Input Type

• Stream of individual items (e.g. float, complex, short, byte — depending on how the upstream flowgraph is configured).
  • For example: float32 samples from an SDR, or complex64 baseband IQ samples.

---

🧱 Output Type

• Vectors of fixed length (you specify the length in the block’s parameters).
  • If the vector length is N, then every N incoming scalar items are bundled into a single output vector of length N.
  • Output data type is the same as the input type, just in vector form (e.g. “complex vector”, “float vector”).

---

⚙️ Example

Say the input stream is:

[1, 2, 3, 4, 5, 6, 7, 8, ...]

and the vector length = 4.

The output will be:

[ [1, 2, 3, 4], [5, 6, 7, 8], ... ]

This is commonly used before:

• FFT blocks (which expect vector inputs).
• Correlators or filters that work on blocks of samples.
• Custom processing where a chunk of samples is required together.

---

🧭 Practical Notes

• If the number of incoming items is not divisible by the vector length, the last incomplete vector is held until more samples arrive.
• No resampling or conversion is done — just repackaging.
• If the vector length is too large for your buffer or processing block, it can stall or underflow.

---

✅ Typical use case:
In radio astronomy or SDR work, for example, you might stream complex baseband data from SDRplay RSPduo, use Stream to Vector to create 2048-sample vectors, and then feed those to an FFT block for spectral analysis.