gregorio
Headphoneus Supremus
- Joined
- Feb 14, 2008
- Posts
- 6,910
- Likes
- 4,138
[1] RRod, my interest is to find out what kind of super audible content exists in typical commercially recorded music and how much of it there actually is. I'm not looking for extreme examples designed to favor ultrasonics. I want to experiment to find out for myself what super audible frequencies add to music.
[2] I'm interested is hearing for myself exactly what they are claiming makes such a difference.
[2a] Your technique for pitching the stuff above 20kHz down to be able to hear it is very interesting. I'd like to do that with typical music that contains ultrasonic content and see how important it really is. Specifically, I'm interested in finding out what the relative volume level is to the rest of the signal, and what those frequencies sound like pitched down. I'm looking for some perspective on this.
1. That entirely depends on the genre and how the music was recorded, what mic's were used and how close to the source sound they were placed. With an orchestral recording for example, mic'ed from an audience position, there would be nothing above 20kHz except unwanted noise/interference (in fact there would typically be little/nothing, besides noise, above about 16kHz). However, commercial orchestral recordings are not recorded from only an audience position, much more closely positioned "spot mics" are also employed and these mics can record some ultrasonic content from some orchestral instruments, although much of it is likely to be mechanical or other unwanted sound. How much ultrasonic content is in the final mix will depend on what those spot mic's picked-up and how much of those spot mic inputs have been mixed with the main (more distant) mics. Popular music is mostly recorded with close mic'ing and therefore we're likely to find more ultrasonic content but again, it depends on what mics were used and how those mic inputs have been processed.
2. Besides the fact it's inaudible anyway, you're not even going to be reproducing that ultrasonic content without some sort of super-tweeter setup (as I believe you use speakers predominantly).
2a. Pitch shifting might sound like a good idea, to get around the problem of both reproducing and hearing ultrasonic content but in my experience it's typically ineffective and misleading. Even using relatively expensive, sophisticated commercial pitch shifters, pitch shifting by a few semi-tones is transparent but beyond that, the audible artefacts increase. Pitch shift by 2 or more octaves, which is what's necessary, and you'll end-up with a very significant amount of artefacts. While some of those artefacts will likely sound obvious, others will be harmonically related and you won't be able to hear what's original signal and what's just pitch shifting artefact. In other words, using the pitch shifting approach is going to be at least somewhat misleading and likely, very misleading. The other potential approach is to slow the recording down, for example, play a 96kS/s file as a 48kS/s file (which some audio editors/DAWs will allow), this would effectively shift the pitch down by one octave and then only incur one additional octave of pitch-shifting artefacts instead of two. However, this too can be very misleading as short, almost transient length notes will sound more like sustained notes, a significantly different thing as far as human hearing perception is concerned. In short, there's no reliable way of knowing "what those frequencies sound like".
The HD Tracks charts surprised me because not only was there noise up there, you could clearly see the redbook roll off before the noise starts. That made me wonder if any of my SACDs actually contain any musical super audible frequencies at all.
It's certain that at least some (if not most) early SACDs were effectively just upsampled redbook. However, you also have to be careful with the statement "you could clearly see the redbook roll-off". Are you sure you're clearly seeing a redbook roll-off, are you sure it's not a mic roll-off, a synth or sampler roll-off, an EQ roll-off applied for aesthetic reasons or even the natural roll-off of the instrument/s? Most commercial mixes are a simultaneous combination of all these types of inputs/channels and it can be virtually impossible to tell the difference between this and a rebook roll-off (44.1kS/s anti-alias filter). More than one engineer I know of has fallen foul of this sort of problem and been accused of just upsampling redbook, when in fact that was not the case.
G