It would also be kind of meaningless if you just allowed the sound enter through one ear and exit through the other, mindlessly. The author obviously needs to pay attention, try to judge every single candidate individually compared to his own mental image of the character. He might want to replay a certain sample if it caught his attention. On the other hand, I imagine he would need to hear only a few seconds of many a sample to immediately know it's no good. This might actually be why he was able to save time. But it's still every single sample.If the estimated time of 3-5 minutes per audio file is correct, it would take the author anywhere between 624 to 1040 minutes. This would be 10.4 to 17 1/3 hours. So that "8 hours later" part wouldn't even mean they finished that.
For anyone wondering, if they were actually done after 8 hours, the audio files would have to average a length of 2.3 minutes.