| Columns Retired Columns & Blogs |
May 13, 2009 - 2:18pm
#1
Subsonic Garbage on CDs - Anyone Notice?
| Loudspeakers Amplification Digital Sources | |
Analog Sources Accessories Featured | |
Music Columns Retired Columns | |
Show Reports | |
Features Latest News Community | |
Resources Subscriptions |
© 2024 StereophileAVTech Media Americas Inc., USA
All rights reserved
I've noticed subsonic junk on recordings, especially since I got my killer SVS subs a few years ago. Though I've never seen/heard anything as low as 1 Hz, not that I actually looked. This is not uncommon, especially today with music mixed on smaller speakers. Foot steps on a flexing floor can get into microphones, and even drums can have substantial LF content get onto the recording if the mix engineer has inadequate monitoring.
--Ethan
Sure, we read about the AC coming on or, a truck passing on the street. However, this is so consistent that it seems it must be electronic.
Wow, can't say I ever noticed that. Then again, even my killer sub doesn't output anything at 1 Hz!
--Ethan
Just to be clear, I'm not hearing it obviously, just watching the cones. I can almost count the cycles.
I don't watch my subwoofer.
Anybody else keep an eye on their's?
Dumb question: Can a CD have 1 Hz signal?
Yeah, PCM has no lower limit on frequency response (although whether the equipment is DC coupled or not is a different story entirely).
It can even encode negative frequencies! <grin>
Pedantic point of order - the lowest frequency that PCM can code is dependent on the length of the digital recording. Nothing lower than 10 Hz can be recorded in a 100 millisecond digital recording, etc.
However, as you say the everpresentoutput coupling caps generally set the actual LF extension.
As long as it is 1 second long, or longer.
The longest CD would be about 80 minutes long, or 4800 seconds, so the lowest frequency recordable on a s CD would be 1/4800 Hz or 0.0002 of a Hz.
The longest 44/16 stereo .wav file is about 200 minutes long, so it has considerably lower bass extension.
What was odd about this recent issue is that it is happening so often, I've seen strong woofer flapping occasionaly on various recordings. On the other hand it was repeating at the Joseph/Manley room, HE 2004.
The particular system that I'm looking at now has a Harman Kardon (by Heathkit) power amp with a -3dB point of .1 Hz.
It will obviously pass just about anything at the input.
Looked at the schematics and the main amp and feedback are DC coupled, it only has a huge (2 - 220uF back to back) input cap. It could be something with my rig, but I've seen it with certain CD's on other systems. The Stevie Wonder CD has it in at least one place where I've seen it in other systems.
I have to dig into this more when I have the time.
The point of this thread was to confirm that others are seeing this also and to perhaps suggest that a 5 to 10 Hz filter should be used in the final CD mastering. Perhaps, home systems should have this filter since there is so much material out there already.
Another example of when bad is good, too much bandwidth passes out of band noise signals and therefore less bandwidth is often a good thing.
Tried a different power amp and the main problem with this CD is easy to find. It is within the first few seconds of track 1.
I don't see it repeating as with the other power amp, seems that amp has issues.
My experience is that bass roll-off on pre-signal (line level)level inputs or outputs is directly equatable with the ultimate crushing of dynamics and harmonic structure. This comes from playing with the guts of gear on a daily basis.
So you might imagine, then, that my 'opinion' is that no roll off of any kind is what is the more desirable (with all things being equal, which they never are) situation.
I've many times run with pure DC gear with no safeties of any kind-whatsoever. Like a shotgun with a hair trigger and the trigger guard cut off. I risk failure when amplifiers will go to to a 90VDC + or - full rail condition with no hesitation..and just sit there until the woofers or amplifiers explode- or both!
But that is the price that one pays for getting to some of the more 'ultimate' sonic fidelity.
This sort of gear is rarely sold to the public as the public don't like stuff that even MIGHT explode. As someone who can fool with it and is willing to take chances on catastrophic failure at that level and HAS experienced failure on that level..and having heard the differences..I head toward it-every time. I've had power outages cost me $k's before from this type of system. Power amps with so much capacitance that when the power goes out and the preamp's power supply drains and goes off center to outputting DC..but the power amp isn't drained for another 30 seconds...well..imagine the panic. Woofers that stick out looking like eyeballs on crushed roadkill as you run frantically toward the speaker wires, in the pitch black power outage (the ONLY hope of saving anything) trying to rip them off your fanatically tightened binding posts.
Too late..it's all gone......
The rest of the time, though..it sounds really great. Heh.
I've noticed this on LPs with no low end roll-off filter. It's picking up the "infrasonic" (not "subsonic") frequencies in the warp parts of the LP. With the filter engaged, this is much less cone excursion, and also seems to color the sound in a negative way, unfortuantely.
Sure I've known about the warp issue for many years. It is worse if the arm cartridge resonance is not well placed. The point here is that I would not expect subsonic garbage on a CD.
Here it is, with what I see as the damped, roughly 4 Hz noise hand drawn in red. Oddly, it is only on one channel, and it is nearly 50% of full amplitude at the start. It is easy to estimate as the difference between the two channels.
Here's a picture: http://baselaudiolabs.googlepages.com/SW-TRK1-LOW-FREQ-NOISE.png
Why not? What you're seeing is that Redbook is quite good enough to faithfully capture any 'subsonic garbage' present anywhere in the signal chain of the recording and mastering.
Including in some cases a nasty case of DC bias.
Just taking a basic histogram of the levels on a CD can provide you with some very, err, revealing information.
Clipping, missing codes, DC average levels of +-1000 (although that is rare, thankfully), clipping at levels well below max and min (identified by peaks in the histogram at the extremes, which are not at +- max), and all sorts of other stuff.
One CD I analyzed used 1 of every 3 digital levels. Literally. It had levels 2 +-3*i, for integer i. Don't ask me. No, don't even recall what it was, only noticed it because it sounded, well, err, "interesting".
Been meaning to get back to this thread, however it seems that I need to back up a bit here to help people see my perspective. Am I surprised that Red Book CD can provide ultra low frequency content, no ... this is obvious. I did study analog and digital signal processing at the graduate level ...
So let me back up and explain why I am surprised. With LP the low frequency content is cut in the mastering process in order to increase recording time by reducing the groove width - this is well known. The noise signal - warp is introduced in the playback process and there is obviously nothing in the production phase that can eliminate the warp noise signal.
It is also well know that there are very low frequency sources such as vocal pops, string plucks and pops that need to be filtered to provide a reasonably band limited signal for playback. Digital formats have no noise source such a warp in the playback path and therefore why not digitally high pass filter the final digital data at say 10 Hz? The cutoff could even be specified depending on the program content. I understand that a few sound tracks have 10 to 12 Hz real content, perhaps even one with 8 Hz.
This would eliminate the need for a subsonic filter in the playback path. Indeed, how many preamps have a subsonic filter in the line stage path, if they have one at all?
I'd still like to know where this noise came from:
http://baselaudiolabs.googlepages.com/SW-TRK1-LOW-FREQ-NOISE.png
Looks like the channel with the noise had some clipping on the negative going cycle right before the LF noise, perhaps a poorly designed DC servo thrown off by the overload?
Why not add a digital high pass filter to standard processing for commercial CDs?
I recently produced a CD, for my son who is an electronic musician (mostly synthesizer/laptop sources). We had it mastered by an experienced professional, fortunately, who found some very low garbage which was unintended but a by-product of some of the synthesizer or digital mixing algorithms I guess. You wouldn't know they were there unless you played it on something with a good subwoofer. He of course removed it easily with EQ prior to production. Could it be that this step sometimes gets missed, even in a high-profile release?
IMO the main reason subsonic garbage gets onto released CDs is because of inadequate monitoring. These days many recordings are mixed in untreated bedrooms using speakers that can't reproduce much below 50 or 60 Hz.
Ethan Winer
RealTraps