bigshot
Headphoneus Supremus
I had an interesting experience today. For the last few weeks I've been irritated by a high frequency ringing distortion on movies encoded as MKVs. I chalked it up to the videos being encoded at too low a bitrate. I've also noticed it every so often in my music server which is encoded AAC 256 VBR, which is odd because I did extensive testing to come up with a codec and bitrate to avoid that sort of thing.
I rarely play CDs any more, but I had just gotten some opera CDs and hadn't gotten a chance to rip them yet, so I popped them in my player. There was that compression artifact sound again. I knew that couldn't be on a CD, so I pulled up the parametric equalizer and went through the bands and sure enough, there was the problem.
When I had run the automatic equalization using the mike and test tones, it had gotten close on the mids and top end, but had made a hash of the balance at the crossover from the mains to the subwoofer. I had gone in and manually tweaked the curve for the low end, but I hadn't touched the upper frequencies.
When I pulled up the band controlling the very top, I found that the automatic EQ had boosted 4dB above the zero line at about 12kHz with a width of about 4kHz. That spike was right up near the edge of audibility, but the boost was causing clipping way up high... Which sounded exactly like a compression artifact. It was a pretty small boost, so it only went splat in the peak levels. That made it intermittent enough for me to attribute it to the file instead of the response curve. Many of my AAC files are normalized down to even out the volume levels, so it only clipped on certain files, not all of them.
I flattened out the boost- who needs a boost up there? And lo and behold, the artifact went away. Now I'm wondering if some of the distortion I've attributed to compression is actually clipping in the highs. With hot mastered music, it wouldn't be hard.
I'd advise that everyone who uses automatic equalization should check to make sure that theIr curve is applying corrections subtractively, and not boosting above the zero line. I never would have imagined that automatic EQ would do that, but it sure as heck did.
I rarely play CDs any more, but I had just gotten some opera CDs and hadn't gotten a chance to rip them yet, so I popped them in my player. There was that compression artifact sound again. I knew that couldn't be on a CD, so I pulled up the parametric equalizer and went through the bands and sure enough, there was the problem.
When I had run the automatic equalization using the mike and test tones, it had gotten close on the mids and top end, but had made a hash of the balance at the crossover from the mains to the subwoofer. I had gone in and manually tweaked the curve for the low end, but I hadn't touched the upper frequencies.
When I pulled up the band controlling the very top, I found that the automatic EQ had boosted 4dB above the zero line at about 12kHz with a width of about 4kHz. That spike was right up near the edge of audibility, but the boost was causing clipping way up high... Which sounded exactly like a compression artifact. It was a pretty small boost, so it only went splat in the peak levels. That made it intermittent enough for me to attribute it to the file instead of the response curve. Many of my AAC files are normalized down to even out the volume levels, so it only clipped on certain files, not all of them.
I flattened out the boost- who needs a boost up there? And lo and behold, the artifact went away. Now I'm wondering if some of the distortion I've attributed to compression is actually clipping in the highs. With hot mastered music, it wouldn't be hard.
I'd advise that everyone who uses automatic equalization should check to make sure that theIr curve is applying corrections subtractively, and not boosting above the zero line. I never would have imagined that automatic EQ would do that, but it sure as heck did.