So after a ton of testing I have found that capturing amps at edge of breakup results in less gain and dynamics than are actually found in the amp. This is especially true when the amp signal is taken directly before the speaker so that it can be used as an amp model and paired with any speaker/mic block. My detailed testing has been with a Victory Sheriff 44 into a Suhr Reactive Load IR in which case I’ve taken about 30 captures at varying levels on INST LEVEL AND IN 1 LEVEL (but I have used other amps in similar cases with similar results e.g. friedman small box, suhr PT15).
I have the amp Gain set on Channel 1 at about 1 o’clock, Bass at 10 o’clock, Treble all the way up, Presence at noon. When plugging directly into the amp there is a great ability to get clean sounds when playing lightly yet gain saturation increases significantly as pick attack is increased. With the Quad Cortex captures the sound is pretty similar in tone (a bit more bright and slightly scooped), but loses about 70% of the gain saturation with increased picking force.
Adjusting the INST LEVEL does make a difference when testing the reference sound, but seems to make no difference in the capture sound. Adjusting the IN 1 LEVEL only seems to adjust the overall volume of the final capture but not the gain or ability to respond dynamically as the amp does. The only way to get close to the way the amp sounds is to either A. Add gain after the capture (although this makes the capture much less dynamic than it actually sounds when using the real amp), or B. to increase the gain on the amp (e.g. from 1 o’clock to 3 o’clock), which does get closer to the dynamic response and gain, but has a bit of a different sound signature because the amp is set differently. What
I am guessing is happening here is that the Quad Cortex is not pushing its test frequencies through at a high enough volume to actually capture the way the amp actually responds to the guitar. This is true of both single coil and humbucker pickups. I’m wondering if I need to run a level booster between the Quad Cortex an the amp in in order to get the appropriate input gain for the capture. Do you have any advice or ideas on how I can approach this? I really do love the unit so far, but this has taken hours and hours of testing and I cannot find a reasonable solution thus far.
Anyone having similar issues? Anyone not having this issue? Any suggestions?
What you observed makes sense, like the QC capture is modeling some limited dynamic range delta round the input and capture settings. Its possible the limits are based on the number of layers and nodes in the deep neural network used to capture the model. These limits might result from memory utilization, training time, and perhaps most importantly, latency resulting from running the captured model in real time.
With the auto level it would be all over the place depending on what I played. Usually if I just played a bunch of different things like chords, single not lines, etc. It would bring the level down to -10dbish. If I just chugged on an major chord as hard as possible like in the QC instructional video it would go to around +5 db. But the INST level made no difference in the captured sound
I’m having a similar issue. I’m getting massive gain loss on my captures, to the point where I fear there may be something wrong with my quad cortex.
Most direct example is when capturing the hi gain channel on my fortin sigil. If I dime the gain on the amp and capture it the corresponding neural capture has about the same amount of gain as the amplifier with the gain set to 9 o’clock.
I can compensate during the set up by boosting the instrument input volume, but that doesn’t seem to translate to the actual capture.
Curious if anyone else is having the same extreme instance of this and how they might be working around it
Yes that sounds just like what is happening to me - the input gain does nothing in relation to the captured sound. I notice it a bit more when I go direct out of my reactive load than when I use an IR. But the whole point is that I want to capture the amp and then be able to run it with multiple IRs. I wonder if I should be adjusting the overall input gain of the QC to help offset this…
I had different experiences capturing pedals vs. an amp and cab. For pedals, if I followed the instructions and boosted the instrument and capture levels to get -10dB or so, I found the gain and volume for the capture block was too high. Setting the instrument and capture levels to 0dB seemed to result in a capture block that behaved more like the pedal.
But for amps, it was the opposite, If I didn’t get the levels up on the input and capture, then the level in the capture block for the amp was too low.
I must say I don’t really understand how to optimize these levels to get a capture that is natural and matches the gain and volume of the captured device.
I am having this exact same issue, where edge of breakup amp captures are coming through with too little gain, especially on DI captures.
I’m also finding captures to have a lot more low-end/mud/boominess than the actual amps. This becomes particularly apparent in patches with multiple amp capture blocks, e.g. for Wet Dry Wet.
Not necessarily just less gain, but also a little loss of “texture”. I would do profiles of DI all the time and started noticing some dullness which I believe is associated with the DI process. I think more gain covers it a little better but I’ve done Captures now with the same amp using a mic and cabinet versus a DI and the mic and cabinet convey the more nature feel and reaction whether low or high gain. I’ve since stopped making DI captures and only use a mic and cabinet because of this. I don’t want to retrace what I did with profiles of many amps to find them lacking. Try the test out for yourselves and see if you find the same. It’s not a matter of the QC its because the line out doesn’t carry the accurate detail of the amp cab and mic reaction so when you add back in another cabinet there is a little something missing.
What DI are you using to make your DI capture? Make sure your DI doesn’t have a “cabinet” simulation eq on the output when you are using the amp input pad settings. I’ve had great results with my Two Notes Torpedo Captor Loadbox/Attenuator/DI. I just make sure the speaker simulator is off. You can find me on the neural app: sleiweke
I’ll share a preset and a capture of my mid 60’s Vox AC30 captured direct. If you need a good blue alnico cab IR let me know, I can email the one I’m using to you.
Several DI’s used over many years. I just did a JJ Jr last night using a Suhr Reactive Load compared to the XLR out (cab sim) compared to a mic on a cab. I’m not saying the DI is bad. What I am saying is the DI does not provide the same feel and sound as the mic and cab. It might be the interaction between the cab and amp are more accurately captured when used together. There is enough that it’s noticeable in the final result so I will stop doing DI captures and only do Amp+cab+mic captures. I get much better results than I ever did with the Kemper, which is why I went to DI profiles then. I started out with the QC making DI captures, but after comparing them with Amp/cab/mic captures I’ve decided to not use DI captures for my personal amps.
I would say try the comparison for yourself before thinking I’m all wet. I’ve checked this on multiple amps so far with the same outcome.
I will say that doing the XLR capture and then comparing it to the result was 100%. It must be that the cab sim in the XLR being static is easier to replicate than the interaction from the cab/mic/amp as the frequency waves are ran through the capture process?
I’m with you @bshaw92. I think that capturing the entire chain as one capture is much more accurate and feels better to me. I have been experimenting with DI captures for my own curiosity.
So in your last post:
“I will say that doing the XLR capture and then comparing it to the result was 100%. It must be that the cab sim in the XLR being static is easier to replicate than the interaction from the cab/mic/amp as the frequency waves are ran through the capture process?”
What did you mean? That you liked the di capture with the amp sim enabled better than all the other options?
No. The JJ Jr has an XLR on the amp with an amp simulation that Friedman made (analog circuit). I was saying that I captured it that way as well and it was indistinguishable from the amp. When I clicked back and forth you couldn’t tell a difference.
I also find that capturing the whole signal chain is the best way to get as close as possible although it is still missing something. Going direct out of the amp pre speaker is tricky because the sound you are capturing is obviously harsh and unpleasant since there is no speaker impacting the sound, so trying to A/B with the “original” sound after capture is hard to do. Further I find that doing the A/B capture test does not actually reflect how the amp actually sounds. Something about running all the cable and signal through the QC sounds different than plugging directly through the amp. But then again, pluggin directly into an amp with 10 ft of cable sounds and fees different than 10ft of cable into a buffer into 10ft of cable and then the amp, even though people claim buffers are transparent, they all do alter the signal especially in feel and compression - there are some that are pretty transparent but they are rare - most have an obvious impact…
Also note that once your capture is done make sure the input 1 level is up based on your guitar. If not you will hit the QC with a lower level and hear less distortion and dullness. FYI
I’m finding that the best direct amp captures are when i’m right at the threshold of failing sanity check. The hotter the signal the more faithfully the capture process nails the gain on the source device.