Why those does over compressing a kick cause the sub to disappear?

[darbo]

Why those does over compressing a kick cause the sub to disappear?

I've been experimenting with compression so I'm not looking for someone to tell me that I'm over compressing the kick, because I know I am over compressing the kick, what I want to know is WHY the lower frequencies end dropping more in volume then the rest what is the reason for this?

Help Much Appreciated

[The Elf]

The time it takes the compressor to react will mean that the initial transient is emphasised, so the click will be much louder than the rest of the signal. As the compressor pulls down the level the boom of the kick will be held low until the signal falls back below threshold.

[darbo]

I'm still a bit confused because when I was experimenting with test I lined up two samples to make the kick an 808 Sub and a higher punchy clicker kick. Then I adjusted the start points so the Sub triggers at the very beginning, it may not be the sound of the skin been beat but it is been triggered at exactly the same time it should be coming through at the same time. So if its what you said and the time it takes the compressor to react how come the sub isn't coming through as much in this time as the rest of the kick? or is it coming through and its just the fact that the our ears need more time to here lower frequencies? I still don't understand

[narcoman]

the lower frequencies arent dropping - what is happening si the overall kick sound is being compressed - ie lowered in amplitude. You'll hear the front , as mentioned above. If you want more of the sub to creep through I'd suggest a tiny bit of compression to smooth as you have it - ie compress less. then put a second compressor to squash a LOT like you want but with a longer attack time.

[jayzed]

Indeed, our ears do take longer to hear lower sounds but that's because they don't exist until a longer time has passed. It's not the ears at fault here, it's physics.

Apart from the theoretical impulse, no sound exists instantaneously. This is the occasionally counter intuitive fact that confuses people about how the nyquist/shannon stuff works and how 44.1 can contain all the frequencies up to 20K without distortion in the digital world.

[darbo]

Ok bare with me

You say I'll hear the front more. But even with a zero attack time and a hardknee it still happens. So even though the comrpressor says 0 attack time is there still a very short delay before it kicks in?

I started looking into the shannon sampling theory too. The first thing that struck me with was if the sampling rate has to be 2 times the frequency bandwidth, so 40,000 samples per second for 20kHz, and CD players and other audio systems are Stereo, would 80Khz, not be an more appropriate sampling rate.
Another thing was well, common sampling rates are 44.1kHz & 88.2kHz or whatever its not 40kHz or 80kHz on the button, is this because low-pass filters used in DAC or not "brick wall" filters?
And how exactly does this relate back to my to this subject. Is it some kinda of Aliasing issue?

[darbo]

I've noticed then when I play the 808 sample of its own it has a quick attack but then when I low pass it to just the sub that attack is missing. So I can understand how a kick with a very short attack time can make the kick sound clicklery. But is the compressor has a 0 attack time, that still confuses, and just to repeat myself, does 0 attack time not necessarily mean 0 attack? is there a short delay

[Hugh Robjohns]

what I want to know is WHY the lower frequencies end dropping more in volume then the rest what is the reason for this?

I think others have already explained.

A compressor takes time to react, and some types of compressor take longer to react than others. That's why people choose opto-compressors (slow) for some things and VCA or FET compressors (fast) for others. These characteristics are mirrored in their software emulations too, of course.

With a high ratio/low threadhold setting on the compressor, it will try to introduce a high level of gain reduction once the signal has exceeded the threshold, but it will take some time to react. As a result, the intial click of the drum will tend to come through relatively unattenuated while the main body of the kick drum will be heavily attenuated.

The result is that you will perceive the low end 'sub' to disappear relative to the clicky initial thwack. This is quite a normal side effect of heavy over-compression with a single band compressor.

Using a faster responding compressor with a very fast attack time will tend to reduce this effect, but not completely.

The other aspect to consider is the release time setting. Very fast release times can also have a significant affect on the perceived low end because the compressor will try to track the fundamental waveform and squash it completely. It's a common problem with bass guitars, but also affects kick drums quite dramatically. Slower release times negate this problem, but don't sound as 'over compressed' and pumpy.

hugh

[Hugh Robjohns]

But even with a zero attack time and a hardknee it still happens. So even though the comrpressor says 0 attack time is there still a very short delay before it kicks in?

The hard knee isn't relevant. The zero attack time may not be all it appears -- it depends on the compressor design. Some provide a 'look-ahead' facility that provides true zero-attack time, but I'm not convinced this will cure your problem either.

As I said above, it may well be your release time is way too fast.

I started looking into the shannon sampling theory too.

A fabulous piece of work... although I think Mr Harry Nyquist would appreciate a mention too But not relevant to your problem with overcompressed kick drums!

The first thing that struck me with was if the sampling rate has to be 2 times the frequency bandwidth, so 40,000 samples per second for 20kHz, and CD players and other audio systems are Stereo, would 80Khz, not be an more appropriate sampling rate.

No. Stereo means two channels -- they both have a nominal 20Hz-20kHz bandwidth, so something above 40kHz per channel is entirely adequate and appropriate.

The bit-rate for a two channel system is twice that of a single channel system, of course. A single 16 bit/44.1kHz wav file has a bit rate of 705.6kb/s where as a stereo one is 1.4112Mb/s.

Another thing was well, common sampling rates are 44.1kHz & 88.2kHz or whatever its not 40kHz or 80kHz on the button, is this because low-pass filters used in DAC or not "brick wall" filters?

Yes. Ideally all unwanted image frequecies should be attenuated well below the noise floor. In a 44.1kHz system, any components at 20kHz in the original signal will appear at 24.1kHz in the unwanted lower image.

the lower the sample rate, the closer the unwanted image frequencies come to the wanted source frequencies -- and if they overlap (called aliasing) they can't be removed.

So choosing a sample rate is all about keeping the sampling images away from the wanted audio, and the bigger that gap the easier things are. The down side is that the higher the sampling rate, the more data there is to move around and store, and the more significant clock jitter becomes.

Back when Sony/Philips were developing the CD format in the very early 1980s 44.1kHz worked out as the most convenient and practical rate for them and hence its familiarity today. Professional users opted for 48kHz for various reasons, one of which was to make the job of the filters easier.

Prior to the Sony/Philips domination, most experimental digital recorders were operating at sample rates of 50 or 60kHz -- and 60kHz would actually have been a far better choice for everyone.

The reconstruction filter in the DAC has to attenuate the unwanted image frquencies to the level of inaudibility which means providing in excess of 100dB attenuation in the space between 20kHz and 24kHz.

Filter slopes are described in terms of dB/Octave and the gap between 20 and 24 kHz is a very small portion of an octave in relation to 20kHz. So the filter ideally has to provide a slope of well over 200dB/Octave and that's not easy. The steepest analogue filter you're likely to come across is 24dB/octave in a synthesizer, or 18dB/Octave in a preamps high-pass filter.

And how exactly does this relate back to my to this subject. Is it some kinda of Aliasing issue?

It doesn't, and no, it isn't.

Hugh

[Hugh Robjohns]

I've noticed then when I play the 808 sample of its own it has a quick attack but then when I low pass it to just the sub that attack is missing.

That's because the 'attack' part is made up of higher frequencies. Removing high frequencies has the effect of making things sound slower and smoother. Boosting hte high frequencies does the opposite -- it makes things sound snapier and faster.

Hugh

[darbo]

I think I understand now.

Thanks to everyone for replying