Effects Overview: ----------------- The linear compressors were designed to be used subtly with a low ratio and high threshold. They should be free of harmonic distortion and give crystal clear results. They behave differently than your typical compressor and beginners may find it finicky. The "nlc" compressor is more of a "set it and forget it" type of processor that adapts to the music. The effects work in 64 bit floating point precision so there is headroom available and you can exceed 0 dB while passing between effects. Use the effects make up gain (see below) to set the level returning to mixer. A tip: If the same effect appears twice for the same channel, the second entry will "overwrite" the first one. This is useful if you want to try a new setting without losing your old ones! Linear Compressor: ------------------ Unlike the common non-linear compressor, this compressor attenuates by a common factor, the ratio, regardless of how high the signal is above threshold. Therefore, the maximum gain reduction is tied directly to the ratio chosen. For example, a ratio of 2:1 will apply a gain reduction of 6 dB to the entire signal once the threshold is tripped. Therefore, ratios between 1.1 and 2.0 are recommended to achieve typical gain reductions. The compressor is applied to a channel by starting a line in the faders.txt file with "comp", then tab followed by the channel number. Then with tab between fields, first is threshold in dB (range of 0 to -144), then attack in milliseconds, then release in milliseconds, then ratio (for example, "3" means 3:1). comp 1 -18 10 150 2 means insert compressor on channel 1 with a -18 dB threshold, 10 msec attack, 150 msec release, and 2:1 ratio. The "comp" compressor responds to peak levels and is ideal for drums and other percussive instruments. The release was designed with a degree of "stickiness" so that it can be used on non-percussive material as well. For vocals, bass and other instruments try the rms sensing compressor (see below). Start a line with "rms" instead of "comp." Note that the rms compressor by nature will require a lower threshold to be triggered. Both types of compressors may be used on the same track and the rms compressor receives signal first despite the order you place them in the faders file. Non-Linear Compressor: ---------------------- Traditionally a compressor is a non-linear device, with a gain reduction based on how far above threshold the input signal is. Mixer4's "nlc" non-linear compressor works this way. It also does attack and release settings automatically, responding to RMS measurements. The engineer can adjust the minimum RMS window to scan. The larger the window size, the more slowly the compressor will react. This is extremely useful when the source has very inconsistent volume. A good starting value is 200 samples. Begin a line with "nlc", then tab, then channel, then tab, then threshold, then tab, then window size in sample, then tab, then ratio. For example, nlc 1 -12 200 3 means insert the non-linear compressor on channel 1, with a threshold of -12, a window of 200 and ratio 3:1. The ratio in this compressor can be set deeper than the linear compressor with good results. Limiter: -------- The limiter is extremely useful when you need immediate control of unwanted peaks above a certain threshold. All the other compressors in Mixer4 have some lead in time before full compression occurs which is generally more natural. The ratio of the limiter is fixed at 8:1, but the threshold is adjustable. The limiter starts working slightly below threshold, essentially allowing up to 8 dB signal over your threshold to "fit" just below your threshold. In other words, if you keep your gain reduction below 8 dB, your output signal will be below your threshold. If you are mastering, the limiter can be applied to mix and then the trim function applied in order to add gain, thus giving you a significantly louder track. An average RMS value of -14 dB is recommended for rock, and lower values such as -18 dB or -20 dB for acoustic or classical music. To add the limiter start a line with "lim", then tab, then channel number, then tab, then threshold. For example, lim (tab) 3 (tab) -4 means add the limiter to channel 3 with a threshold of -4 dB. High Pass Filter: ----------------- A gentle high pass filter can be placed before the compressor stage (and is not linked to the compressor) by inserting "hp" at the start of a line. Then enter a tab and input the channel number. Then a tab and input the center frequency. hp (tab) 1 (tab) 100 means insert a high pass filter on channel 1 with a cutoff frequency of 100 Hz. Reverb: ------- The reverb is a simulation of a room with reflections from all 6 surfaces. There are two varieties. The one called "rev" takes reflections and recirculates them so that the density of reflections increases with time. The other called "trev" is the original reverb that consists of up to 66 discrete taps (11 from each surface). Both reverbs can be used simultaneously and their return levels set differently. "trev" might be better in a case where you want more of a delay type effect. Both reverbs allow you to input a room length, while the other dimensions are calculated automatically and are based on an ideal ratio. To add the recirculating reverb to a track start a line with "rev", then tab, then channel number, then tab, then the send level. For example, rev 1 -12 means send track 1 to the recirculating reverb at a level of -12 dB. The discrete tap reverb is accessed in a similar way. trev 1 -12 means send track 1 to the non-recirculating reverb at a level of -12 and with 4 tap cycles. To adjust room size and return level (actually the master send level), start a line with "rtn" then tab, then enter the return level, then tab then enter the room length in feet. rtn -6 75 means the recirculating reverb return level is -6 dB and the room length is 75 feet. trtn -6 75 11 means the non-recirculating reverb return level (actually the master send level) is -6 dB and room length and 11 tap cycles. Note you must add an additional parameter of 1 to 11 to specify the number of tap cycles. The actual number of reflections would be 6 times this number. There can be a significant difference in processor load depending on this number. For the non-recirculating reverb it is necessary to keep the room length under 360 feet for a sample rate of 44100 and proportionately less for higher sample rates if you use all 11 tap cycles. To turn the reverb off temporarily simply alter the "rtn" and "trtn" codes by putting parenthesis around them or something in front of it like an "x". The sends may be left alone. Finally, you may choose the position in the room in 3d space where the source (same as the listener) is located. Begin a line with "pos" then enter the back-front, left-right, and floor-ceiling ratios in a percent format, tab spaced. For the back-front position, for example, 0 means fully against the back wall and 100 means fully against the front wall. In the other dimensions, 0 is fully left and 0 is the floor. To maximize the variety of reflections, it is recommended to not be directly in the center of the room. For example, pos 33 43 29 means place the listener 33% of the way between the left and right walls, 43% of the way between the back and front walls, and 29% between the floor and ceiling. This would put the listener closer to the left wall, back wall, and floor. Using a setting of 50 50 50 will produce a very boring effect because you will have the same timing intervals for left/right etc. So, use an off- center approach for best results. The "width" of the "rev" reverb can be changed by adding a line starting with "width", followed by a tab and then a number from 0 to 100. Zero would produce the former reverb sound, where there are clearly stereo reflections, but ones from the ceiling and floor are in both ears. By default the width is now set to 50 which will tend to will throw the ceiling and floor reflections in one ear or the other, thus lowering reflections perceived as coming from the center of the mix. The maximum width is 100. Filtering the Reverb: --------------------- Beginning in version 1.15 you may apply a high pass and low pass filter to the reverb send. This will affect both rev and trev reverbs. This is useful, for example, when sending a whole drum kit to the reverb and you do not want the bass drum's low frequencies to cause rumbling as it bounces around the room. Also, you can make a more subtle reverb by filtering out some of the high frequencies. To add a high pass filter (to remove low frequencies) use the command "rhp" followed by the frequency in Hz, tab separated to the faders.txt file. To add a low pass filter (to remove high frequencies) use the command "rlp". Both can be used together, for example, to narrow the bandwidth of the signal sent to the reverb to primarily between 200 Hz and 7000 Hz, you enter rlp 7000 rhp 200 Parametric Equalizer: --------------------- Four bands of parametric EQ are available, plus high and low shelves. Begin a line with "band1", "band2", "band3", or "band4". Then tab and channel number, tab and frequency in Hz, tab and gain in dB, tab and value for Q. For example, band1 4 4200 5 1 means add a peaking eq to channel 4 with center frequency of 4200 Hz with a gain of 5 dB and a Q of 1. To add a high shelf or low shelf start a line with "hshelf" or "lshelf" respectively and follow the same format above. Effect Makeup Gain: ------------------- The effects processing uses 64 bit floating point math which essentially means the quality will be consistent no matter what overall level enters the effects chain. After applying effects such as compression, the signal level will drop as desired, but when returning to the main mixer the signal is converted back to fixed point data. The upshot is that to retain the highest audio quality you should add some makeup gain to the signal if it drops too low (more than 6 dB for example, during effects processing). It would be better to use effects makeup gain to bring up a low signal than to apply the trim function in this situation. However, if you are not using effects on a channel at all then there is no point in using effects makeup gain since the signal stays in the pristine fixed point format and the "fxg" function is not available to the channel anyway. Reverb does not count as an effect in this sense. To apply effect makeup gain start a line with "fxg", then tab , then channel number, then tab, then gain in dB, then enter. For example, fxg 1 8 means add 8 dB of gain to the signal of channel 1 after effects processing. You may also reduce the volume using the makeup gain function such as when boosting an already hot signal with EQ. Master Effects: --------------- Beginning with version 1.06 you can apply effects to the stereo signal after all channels have been summed. Simply use "M" instead of a channel number. At this time you cannot use the reverb on both the channels and master simultaneously. Nor can you use rev on the channels and trev on the master, for example, because both reverbs are linked. There is no such limitation with the other effects. Among other possibilities, it gives the user the ability to produce a peak limited and balanced master all within the same fader file as the mix. See also the volume automation section for automating the master output volume. Changing Effect Order: ---------------------- With some minor exceptions, the order of effects can be changed in version 1.10 or later. In the faders.txt file, use the command "ord" for order. This is followed by a colon and then the name of the effect. Then, enter tab and the channel number. Then, another tab and the effect position. For example, ord:comp 3 1 This means change the order of the compressor on channel 3 to position 1, the first processor to receive the signal. There are 11 positions available. If two effects are in the same position then the effect that will run first will be the one that normally runs first in the default setup. Before you change the order, all effects are in position 11, and EQ's process before compressors. The exceptions are that the reverb and limiter are not movable on the master fader channel. And the fxg, effect makeup gain, is fixed as the last processor on all channels.