[SoundKlang]

Okay, here are more samples.

Example 3 - 747 on a (sampled) Grand Piano

Objective:
Reduce the crispness in the high notes, even out the dynamics, bring the ambience to the foreground without darkening the overall sound.

Comment:
The basic idea was to bring the ambience to the foreground by keeping the compressor permanently working in the first half and then opening up in the end. In the result the ambience was less emphasized than wanted but still okay. For the low compression ratio a lower threshold was necessary and even with the low threshold needed all possible input gain inclusive +10dB option and a raise in level in the DAW was necessary. In the original the high notes were played with high velocity. They were too loud and too crisp. This was reduced by the compression, too.
I am beginning to understand why people wrote about hours of knob tweaking ... this is the best compromise I was able to find and still the compression and the loss in detail is too obvious. Maybe okay for use in a pop arrangement when the piano is masked by other instruments, but probably not for solo piano. Then on the other hand this is trying to do something, which I guess the 747 was not intended for.

Avalon Settings:
- Input Gain: full input gain inclusive +10dB option (see above), TSP off for a clearer signal
- Compression: Sidechain On, 160Hz (bass note) and 2.5kHz (high note) set to being less(!) detected, fast Attack and slow Release, Ratio 2.5:1 (very difficult to find a setting which was working but not obvious), minimum Threshold
- Equalizer: Off
- Output Gain: used for level matching


Chain:
RME AIO Analog Out (still no better DA) > Vovox Cables > Avalon Vt 747sp > Mogami Cables > Forssell MADC-2 > AES/EBU In
Pyramix Native V7
Dry Signal originally rendered @192kHz/32bps; then rerecorded through the chain without the 747 @192kHz/32bps; then downsampled to 44.1kHz/16bps with SRC (Linear Phase Filtering) and Redithering (Requantization to 16 Bit and High Pass Noise Shaping) Mixdown
Wet Signal @192kHz/32bps from Real Time Mixdown; then downsampled to 44.1kHz/16bps with SRC (Linear Phase Filtering) and Redithering (Requantization to 16 Bit and High Pass Noise Shaping) Mixdown and Redithering Mixdown to 44.1kHz/16bps with Linear Phase Filtering and Dithering with Requantization to 16 Bit and High Pass Noise Shaping


Example 4 - 747 on a (sampled) Drum Groove

Objective:
Try parallel compression with the 747.

Comment:
The 747 does not offer the possibility to mix the dry and the processed signal. True parallel compression is not possible with the 747. For these examples a lot of compression was used and the result saved as a first wet file (04 Drums_Wet*.wav). Then this first wet file was mixed with the original (04 Drums_Wet*Parallel.wav), paying attention to an exact sample alignment of both files. This way parallel compression is possible with the 747.

Optimum - Assuming that the ambience of the snare is the most wanted aspect of parallel compression on drums, the sidechain filters were set to the kick and the body of the snare in a way that the detection circuit reacts very much to the snare hits and only a little to the kick. There is a very delicate balance between the threshold, the ratio and the release time. In a way like "the lower the threshold the smaller the ratio must be" and "the higher the ratio the more gain reduction, which then needs shorter relase times". But there is a certain amount of release time necessary, so there should not be too much gain reduction. But less gain reduction is achieved by a smaller compression ratio, while a compression ratio too small does not deliver the wanted sound (the closer it gets to 1:1 = no compression at all). Also, when the Hihat gets damaged, either the ratio or the release time should be decreased. After the threshold is set, the most work is done by comparing different settings of ratio and release time without damaging the HiHat and after mixing them with the original file. Since the transients will be provided by the dry file, there seems to be no need to use a slower attack time. All in all the result achieved here seems interesting and would be much more flexible when not compressing the whole drumkit, but the kick and the snare independently from each other.

Extreme - Here extreme compression was used similar to that used in examples of other compressors doing parallel compression on rock drums. To achieve this sound a very high ratio was necessary, which negatively affected the HiHat. Also, even with the fastest attack time, the compressor seemed to miss a transient every now and then. It got better with a longer release time, but too long release times did loose the ambience of the snare. This was the first approach to parallel compression on drums with the 747. The low sound quality of the result made it necessary not to try doing with the 747 what other devices do but again to use the 747 within the boundaries of its abilities. Further tweaking then led to the previously described "Optimum" example.


Avalon Settings:

(Opt)
- Input Gain: full input gain inclusive +10dB option (see above), TSP off for a little more transparency
- Compression: Sidechain On, SC Listen off, 60Hz/-8dB and 800Hz/+12dB, fastest Attack and almost fastest Release, Ratio 8:1, minimum Threshold
- Equalizer: Off
- Output Gain: used for level matching

(Xtreme)
- Input Gain: full input gain inclusive +10dB option (see above), TSP off for a slightly better transient control
- Compression: Sidechain On, 80Hz and 2.7kHz, fastest Attack and Release, Ratio 20:1, minimum Threshold
- Equalizer: On, cut at bass bell and shelving bands, cut at high shelving and bell bands, EQ before Comp
- Output Gain: used for level matching


Chain:
RME AIO Analog Out > Vovox Cables > Avalon Vt 747sp > Mogami Cables > Forssell MADC-2 > AES/EBU In
Pyramix Native V7
Dry Signal originally rendered @192kHz/32bps; then rerecorded through the chain without the 747 @192kHz/32bps; then downsampled to 44.1kHz/16bps with SRC (Linear Phase Filtering) and Redithering (Requantization to 16 Bit and High Pass Noise Shaping) Mixdown
Wet Signal @192kHz/32bps from Real Time Mixdown; then downsampled to 44.1kHz/16bps with SRC (Linear Phase Filtering) and Redithering (Requantization to 16 Bit and High Pass Noise Shaping) Mixdown and Redithering Mixdown to 44.1kHz/16bps with Linear Phase Filtering and Dithering with Requantization to 16 Bit and High Pass Noise Shaping


Example 5 - 747 on a Recorded Synth Soundscape

Objective:
Get some analog warmth with the 747. Even out the jumps in loudness.

Comment:
This was way more difficult to achieve than expected. What is analog warmth? To me: less cold high mids, more midrange overtones, some slightly excited shimmering highs, a denser yet transparent texture like from driving a tube. In this example there is some organic movement in the sound. It sounds like at least 3 layers. All these details, the inner structure of the sound should still be visible, just less analytic. Again the TSP was offering some warmth but could not be driven too much before clouding the result a little. And what to do with the high bottle whistle sound which is by definition a little cold but essential to the soundscape? And the huge jumps in level when the drones roll in? The compressor had a lot to do here but should be as little visible as possible. The result is a compromise by trying to find a way between all these different demands - dynamic range more controlled, more round and sustained bass and mids, less high mids. Some shimmering highs like offered by some tube gear (e.g. in a way by the high band of the UA 610) would be desirable but not possible with the Avalon 747. The result achieved here stays definitely less Hifi than the original, more like old analog gear, but is this analog warmth? Not sure about this one, I like it somehow. Probably good enough to add some more originalness into digital synth legend clones. (Again, the original may sound more interesting at first due to more level in the high mids and the much more dynamic in the drones.)


Avalon Settings:
- Input Gain: not too much to stay transparent, +10dB option Off, TSP on
- Compression: Sidechain Off for more transparency, slow Attack and Release for a more stable less visible action, Ratio 20:1 together with a Threshold set for about 7dB gain reduction for best transparency by affecting the quieter parts as little as possible
- Equalizer: On, slight raise at bass and low mid bell bands, slight cut at high bell and shelving band (not the 'air')
- Output Gain: used for level matching


Chain:
RME AIO Analog Out > Vovox Cables > Avalon Vt 747sp > Mogami Cables > Forssell MADC-2 > AES/EBU In
Pyramix Native V7
Dry Signal originally recorded at lower resolution and now rendered @192kHz/32bps; then rerecorded through the chain without the 747 @192kHz/32bps; then downsampled to 44.1kHz/16bps with SRC (Linear Phase Filtering) and Redithering (Requantization to 16 Bit and High Pass Noise Shaping) Mixdown
Wet Signal @192kHz/32bps from Real Time Mixdown; then downsampled to 44.1kHz/16bps with SRC (Linear Phase Filtering) and Redithering (Requantization to 16 Bit and High Pass Noise Shaping) Mixdown and Redithering Mixdown to 44.1kHz/16bps with Linear Phase Filtering and Dithering with Requantization to 16 Bit and High Pass Noise Shaping