Hey all - new user here, signed up specifically to share the following observations, and join the very good discussions on here...
Following up on discussions back in October about gain staging, Entropia sounds and the HL compressor:
I also had some issues getting to grips with Entropia. Things tended to feel over-compressed, and I struggled to dial a nice non-metal edge-of-breakup tone in. Messing with different drives in Anagram I felt my results are inconsistent and I tried to dive deeper. I now think different drives treat the dry/wet blend knobs and/or clean / processed level knobs differently.
To test this, I put a gain block at the start of the signal chain, kicking the signal to a very hot but juuust not clipping ~ -5dB as indicated by the meter on the left. Next I put different drive blocks, and observed the output levels as indicated by the meter on the right for different settings of the drive block.
I found three main classes of drives:
- B3K, VMT, AO, Duality Fuzz, Green PiRate Fuzz
- The clean blend seems to bypass the drive block's model
- For these drives, the block's level/volume knob did not have an effect on blend full dry. A hot input gives a hot dry output even when the level/volume is turned to zero.
- The wet output levels at full level/volume knob is hotter than mid-scale and varied a bit between drives, but was lower than my very hot test signal. Around -12dB to -8dB were typical.
- If you have a very hot signal chain and only a little bit of wet blend going, your signal level changes with blend.
- Microtubes X, Chinchilla, Atomic Sans, Bee Kolme, Arc Dreamer, Entropia
- The clean blend seems to be part of the drive block's model
- The level/volume affects also a full-dry blend, but the details vary a bit between drives:
- On some drives (eg Chinchilla), the signal on full dry blend had the very hot input signal levels, while the full wet blend was around mid-scale for 12 o'clock volume/level
- On some drives, the blend had no effect on signal level. Dry/wet and between were both as hot as the hot test signal (eg Atomic Sans, Arc Dreamer), on some the output was around mid-scale at 12 o'clock level knob for a hot input (Bee Kolme)
- Entropia, Microtubes X
- These anyway don't have a classic wet/dry bypass, but rather frequency crossovers that can let some signal bypass. These more complex bypasses will do ..."something"... with hot signals. Hard to determine.
Also, it seems this behaviour of the drive blocks actually changed a bit over the various firmware updates. Eg my notes from my first tests on version 1.8 put the PiRate into the second class together with eg Chinchilla, rechecking this on 1.9 while writing this post puts it clearly into the first class with eg Duality. So ymmv.
Now: what to do with this?
I think it shows it's important to understand gain staging best practices within the Anagram. I presume the models are optimised to ~mid scale levels, since many gravitate to this as output level when left to defaults. So probably it's good practice to place a gain block at the start to achieve this (until Darkglass gives us a global input level setting...).
Also, adding gain blocks to boost the level in front of some block to achieve a certain effect will probably lead to unpredictable results for different blocks, and maybe also between firmware updates.
Maybe @IvanBassist, this would be a good topic to explain in a deep-dive video?
