|
|
Post by shoetique on Apr 6, 2024 17:02:12 GMT -6
I've been hunting high and low and cannot figure out what exactly the SC toggle on the GssL I built does. I understand that it is a side chain, and have read through the documentation on Gyraf's page here, but I don't know what this means in sonic terms.
I've used the comp pretty extensively and it sounds to me that when SC is engaged it preserves the bass and uses the mid/high end to determine when compression takes place. Is this an accurate assessment of how this works or do my ears deceive me?
|
|
|
|
Post by prene1 on Apr 6, 2024 17:35:58 GMT -6
I've been hunting high and low and cannot figure out what exactly the SC toggle on the GssL I built does. I understand that it is a side chain, and have read through the documentation on Gyraf's page here, but I don't know what this means in sonic terms.
I've used the comp pretty extensively and it sounds to me that when SC is engaged it preserves the bass and uses the mid/high end to determine when compression takes place. Is this an accurate assessment of how this works or do my ears deceive me?
You are correct. I forgot what frequency it activates. |
|
|
|
Post by poppaflavor on Apr 7, 2024 11:40:48 GMT -6
I've been hunting high and low and cannot figure out what exactly the SC toggle on the GssL I built does. I understand that it is a side chain, and have read through the documentation on Gyraf's page here, but I don't know what this means in sonic terms.
I've used the comp pretty extensively and it sounds to me that when SC is engaged it preserves the bass and uses the mid/high end to determine when compression takes place. Is this an accurate assessment of how this works or do my ears deceive me?
WooHoo what a doozy of a sidechain description. I especially like how gyraf describes it as 'surprisingly simple' :-) gyraf.dk/gy_pd/ssl/ssl.htm#:~:text=The%20anatomy%20of%20this%20compressor%20is%20surprisingly%20simpleThis was a head spinner, but here is my (probably completely incorrect) take on it just so I can get it out of my still-spinning head: Instead of two GR SC signals each independently going into the VCAs for Ch1 and Ch2 of the SSL Bus Comp, there is a single SC feeding both VCAs for Ch1 and Ch2. This SC appears to be a summation of 2 different DC signals: #1- The signal derived from a Mid signal from the original stereo input, meaning input L+R (likely also -3dB correction) converted to DC. This is audio signal that is missing side information. It is feed forward to this SC VCA, but then it is feedback to the actual GR VCAs. #2- The Threshold-derived DC signal that modulates the Mid feed forward signal described in #1. This is not audio information, it is the threshold voltage and thus offsets the SC signal strength according ot the threshold. This is neither feed forward nor feedback. In terms of audio sonics, I imagine then that the comp will compress lower freq material more since the lower freq material tends to live on the mid channel more so than the side channel. Also, the more compression occurs and the more that makeup gain is applied to stay at unity gain, the more the makeup DC will be a component of the SC signal and thus the more the comp will be feedback instead of feed forward so that it is smoother and less peaky/grabby. So the design seems to emphasize punchy low end content (from summing the SC to mid) and a adaptive response to compression as more threshold is applied. In this way, just looking at the schematics it seems that it might possibly be the opposite what you describe. That the mid and high frequencies are not big determinants of the compression triggering since (especially the highs) are more represented on the sides depending on stereo width of the program material. It looks like the lows might be more determining the triggers since the side chain is comprised of the mids, where the lows live. Not trying to contradict your experience, just trying to naively read this schematic. What a fun read! |
|
gyraf
New Member
Posts: 11 
|
Post by gyraf on Apr 10, 2024 10:24:33 GMT -6
If you're not deliberately using the sidechain filter insert point (for creating e.g. a sc. lowpass), the GSSL reacts linearly and equally to the whole audio spectrum. The only filter pole we have in the sidechain is the 33K/100pF at the proxy VCA recovery buffer - and this gives something like a 48KHz lowpass, hardly influential
..dropped you an email, saw this question late..
/Jakob E.
|
|
|
|
Post by shoetique on Apr 10, 2024 17:11:53 GMT -6
Thanks for taking a look everyone! Here is a copy of the response Jakob sent to me. It really helped me understand what was going on here: --- The GSSL does not have a sidechain filter as-is (the desk never had..) - but there is a wire link where you can insert a capacitor of choice and get a 6dB/oct highpass for the sidechain signal This works because at this point in the circuit the sidechain signal is still AC/audio, i.e. this point is just before the sidechain rectifier makes it DC. So a filter here kinda makes the compressor sidechain progressively blind to frequencies below the selected cutoff. Now, at what frequency is our sidechain low-cut for a given capacitor, and how should we dimension for a frequency of e.g. 100Hz? (Note that "100Hz" actually means "a filter with a -3dB point at 100Hz", not like entirely removing everything..) For dimensioning a first-order RC filter you'll need to know the load resistance we are looking in to: In this case, the filter's load is the combination of the two 20K resistors going to the rectifier inputs. These terminate in 0 Ohms (into virtual-ground inverting inputs), so their combined load is 20K//20K, which is 10K Ohm, 10.000 Ohm Using e.g. the calculator at www.muzique.com/schem/filter.htm - input'ting 10000 ohms and wanted 100Hz it replies "0.159uF" - i.e. 159nF Closest easy-to-find capacitor value is 150nF, which, when entered into the same calculator gives you 106.2Hz - close enough for rock music :-) --- Based on this information it looks like the PCB Grinder Kit I used opted for a .1 uF cap which results in a 159.2Hz high pass filter when the SC toggle is engaged. |
|
