search

claritychallenge / clarity_CC

Support for Clarity Enhancement and Prediction Challenges (obsolete - see README)
http://claritychallenge.org/the-team
Creative Commons Attribution Share Alike 4.0 International
46 stars 10 forks source link

Fixed typo in installation #5

Closed MKegler closed 3 years ago

MKegler commented 3 years ago

Hi, thanks for setting up this awesome challenge and providing baseline tools!

I tried to fresh-install & test the framework on Linux (Ubuntu 18.04.5 LTS) and came across two minor error/typos. I believe that the below edits should fix it.

MKegler commented 3 years ago

@jonbarker68 thanks for the quick reply. So at the time of testing, I haven't yet downloaded the official data (takes a while), but used the small test set provided in the repo. In this case, maybe it would make sense to re-name the testing data (./data_test/clarity_data/train/rooms/brir/) to follow the same convention (i.e. from brir_xxx to anech_brir_xxx)?

sgraetzer commented 3 years ago

Thanks @MKegler. Since we introduced the new anechoic brirs we haven't updated the test data accordingly. That is, the anechoic BRIRs are missing rather than incorrectly named. The check_data function will fail when applied to the test data.

MKegler commented 3 years ago

@sgraetzer okay, I see. The check_data is indeed failing on test data, but is that the desired outcome (was a bit confusing for me when I first ran the code)?

Either way, apologies for the confusion. I just reverted changes in SceneGenerator.

jonbarker68 commented 3 years ago

No, it's not desired for the check_data to fail :). Looks like we need to add some dummy anech_brir_xxx signals to the test_data to fix this. We'll fix this in the next commit to avoid confusing others. Thanks for spotting this.

sgraetzer commented 3 years ago

Thanks again @MKegler. These issues should be resolved now.