Splitting / doubling signal to feed compressor sidechain, most transparent way?
On a master bus, I'm coming out of an eq, going into a compressor. I wish to send a separate copy of the signal to the compressor's sidechain so I can eq the sidechain separately. On compressors that have sidechain SENDS and RETURNS, this is a breeze... just stick the eq between the send and return, done. ALL compressors should have this!!!! But, sadly, even a lot of great compressors out there have just sidechain inputs only (no sidechain sends), thus you need to get a copy of the signal from elsewhere to feed the sidechain. So now, in a critical master bus path, I have to wonder how to split my master stereo signal without in any way harming it. Is there some kind of ultra high-end splitter that can be gotten that 100% preserves the original signal? I am aware that some mastering consoles have features that allow for this, but I was hoping to avoid buying a super expensive mastering console just to do this one operation. Thanks.
Just a quick workaround idea, if you have multiple AD connected from your daw, you could dupe the track in the daw and send one out either to an external unit to EQ before your sidechain input, or perform the EQ for the sidechain within the daw itself. Alternatively you might have a signal routing system via your software that would allow you to dupe the daw signal to different AD.
That might not always help if your compression is happening later in your analogue chain, but I felt it worth mentioning. It would certainly avoid having to split signals in the analogue domain if there were pretty light EQ before the comp that didn't effect the overall detection much, or if the EQ in your analogue domain was fairly easy to replicate on your software EQ for the purposes of feeding the sidechain.
Y cable. Barring some kind of really screwy impedance mismatch, what could be more transparent than a piece of wire?
But isn't it bad to take a signal and merely split it? Don't you lose half the energy going to each of the two destinations? As well, doesn't this indeed mess with the impedance? I never tried it but was always told not to. On a master bus or mastering chain, surely you don't want to mess around and do anything that might even slightly "degrade" the signal.
Man, I really wish more compressors had sidechain SENDS as well as returns. This is becoming an important feature in my world here.
Hey guys, longtime lurker here, learned a lot from all you and just figured I'd bump the thread and share a serviceable (although as-of-yet personally untested) bit of advice I came across, courtesy of DriverGuru on another website: Splitting an Audio Signal: Please explain this to me once and for all... - Ars Technica OpenForum . (I hope it's legal to post external links like this, if not I apologize to mods in advance, my third grade teacher should've beat me for not reading the directions first)
When splitting balanced signals:
The lower the source (output) impedance, and the higher the load (input) impedance, the higher % the original signal gets passed on, and with lower % feedback/distortion, except with the qualifier that this greater impedance difference creates a higher chance for noise pickup from surrounding equipment (I think because of the higher voltage flow drawing in artifacts from surrounding space with greater attraction or something).
I think recommended "you'll be fine" impedance ratios are anything above 10:1 load:source impedance, seeing as according to wikipedia that ratio is the working definition of when an impedance-bridging situation is occurring. That being said, the higher the ratio the better, unless it's so high that you're picking up noise.
So putting this together with the y-splitter idea, you will be A-OK to y-split a signal into 2 outputs as long as they are either going into two balanced inputs on the same device where the ratio before splitting was at least 20:1 (since it will be halved after the split), OR they are going into two balanced inputs on separate devices, both of which would have had the same pre-split ratios of at least 20:1.
If they are both over 20:1 but the input impedances of each vary by more than a little bit, I don't have any more knowledge than that the signal will be hotter relatively on the input with a higher impedance than the other, with the effect increasing as relative input impedance ratio increases. Additionally, I would imagine distortion may develop in this case, in addition to just the differing levels, but I am not certain of this by a long shot.
Anyway, that's what I (think I) learned from the referenced post, so credit to DriverGuru on that forum. Just figured it may be of use.