ulmshare: document the BS100BW/BS50BW IQ offset in Known Issues
#97
by oisinn - opened
- ulmshare/README.md +1 -0
ulmshare/README.md
CHANGED
|
@@ -166,6 +166,7 @@ Aggregate only; the subjects are mice, so no PHI applies. Subject attributes are
|
|
| 166 |
- The metadata should include a `brain` anatomy label.
|
| 167 |
- `zea` version for the files is not up-to-date.
|
| 168 |
- Some of the acquisitions are missing a subset of the channels for the first 400 transmits or so. This is also present in the original ULMShare dataset and does not significantly affect results.
|
|
|
|
| 169 |
|
| 170 |
## Ethical Considerations
|
| 171 |
|
|
|
|
| 166 |
- The metadata should include a `brain` anatomy label.
|
| 167 |
- `zea` version for the files is not up-to-date.
|
| 168 |
- Some of the acquisitions are missing a subset of the channels for the first 400 transmits or so. This is also present in the original ULMShare dataset and does not significantly affect results.
|
| 169 |
+
- The I and Q samples in `data/raw_data` are a quarter carrier period apart: Verasonics BS100BW/BS50BW sampling captures I at $t = kT$ and Q at $t = kT + T/4$, and the conversion stored them as simultaneous. The original ULMShare release has the same offset. [zea#672](https://github.com/tue-bmd/zea/pull/672) adds the correction to `bs100bw_to_iq`; these files have not been reconverted with it yet.
|
| 170 |
|
| 171 |
## Ethical Considerations
|
| 172 |
|