mp3 wrote:macman wrote:Remember all your TOP producers you want your beats to sound like process there drums through the following.
API 550B'S £2000
SSL Buss Comp £2200
Neve Pre Amp's (Darkchild) £1800
Neve EQ £2000
Engineer £600 per day
SSL Board
I don't wanna sound like them. That's a cop out.
What's that got to do with a filter that doesn't have the headroom to deal with a normalized kick drum sample?
Honestly, if you're going to represent Akai, then cut out the hyperbole.
Maybe I can save MacMan the trouble...
If a sample is normalized, than it means that any further increase in gain would cause it to clip.
All filters (i.e. regular EQs too) create a resonant peak even when they are cutting a given frequency. In fact, this resonant peak is what makes one filter more popular, or 'sound better' than another. The presence of a resonant peak means that there will be an increase in gain when the filter is used.
So... this means that...
Any perceived headroom that a filter has means that the processor is attenuating the gain on the input (usually via a -10db or -15db and a limiter on the output). This is the same thing as having to turn down the gain (as you described) yourself, but gives you less control. There will simply have to be some attenuation (somewhere) anytime you apply a filer to a normalized sample. The question is whether or not you'll have control over it, whether the filter has a pad on the input, or whether there is a "smart" gain structure to the filter that attenuates only the needed amount (don't we all wish).
Keep in mind that in the digital world, (apart from the db difference between your noise floor and clipping) there is no such thing as headroom. In the analog world, devices have the capability (or lack of) to produce extra strong signals for short periods of time without undesirable distortion, but in the digital world, this concept doesn't work.
To compensate for our (all-too-often) poor gain stage setup, digital processors are often set up to apply attenuation at the input, and makeup gain on the output. When we have control over this, it's great. When we don't, this simply serves to raise the noise floor of the recording.
If processors didn't do this, they'd clip all the time unless we turned down our faders.
For an outside example of this, consider how ProTools and Nuendo use different methods to deal with the same problem:
When you've got a number of audio tracks (in either program), and they're being summed to a Main output, all the bits need to be added up. If all your tracks are peaking around 0db, when they're added up, the resulting signal is going to clip the main output.
Earlier versions of ProTools (maybe the current ones too, but I switched) throw out some of the extra bits out (so it doesn't clip). This is why people used to complain that it sounded bad. It also fueled the sales of external summing bus units.
Nuendo and Cubase and now Sonar use 64Bit summing busses so that all of your 24Bit tracks could actually add up to more than 0 without distorting.
Hopefully this can help put the filter thing into perspective for you.