Recovering a failed software RAI, where does parallel -j0 dd to?

I'm trying to go through
https://raid.wiki.kernel.org/index.php/Recovering_a_failed_software_RAID
echo $DEVICES
/dev/sda1 /dev/sdb1 /dev/sdc1 /dev/sdd1
parallel -j0 dd if={} of=/dev/null bs=1M ::: $DEVICES

dd: error reading '/dev/sdd1': Input/output error
23510+1 records in
23510+1 records out
24652464128 bytes (25 GB, 23 GiB) copied, 173.685 s, 142 MB/s

Or is that just doing a scan and just checking which drive are bad? I
sure would like to get that 25GB back. Is that possible?

I also see this:
ddrescue -r 3 /dev/old /dev/new my_log
ddrescue -R -r 3 /dev/old /dev/new my_log

Can I just use a mount point for /dev/new?