The error here is complaining about 32-bit samples, so you don't need to lose 8 bits of data, but 24.

You need to use some tool that understands 32-bit sample TIFFs - eg. Imagemagick, Imager and use that to convert to 8 bit samples.

eg.

convert input.tif -depth 8 output.tiff

or

use Imager;
my $im = Imager->new;
$im->read(file => "input.tif") or die $im->errstr;

my $im8 = $im->to_rgb8;

$im8->write(file => "output.tif") or die $im->errstr;
[download]