This whole separate preamp and power amp thing.

[hideout]

I posted this comment in Ideascale.  What do you guys think?

 

I read somewhere that Line 6 is hesitant about this idea because when you separate the preamp and power amp and stick a device between them, you change the character of the interaction between those two units - thus changing the sound of the modeled amps. This is true even when you use a buffer, I think and I think I can attest to this. I had a Fender Blues Deluxe some years back and I did notice a change in the way the amp sounded and felt when I put something in the loop.

That said, what if Line6 simply created several modeled power amp models with Fender, Marshall,Vox and Mesa etc. topologies or maybe one power amp module that's reconfigurable to those topologies? After all, do any of us really know how a Mesa (or whatever) Power section really sounds like on its own? I doubt it.

[MusicLaw]

Ben Adrian, Line 6's Chief Helix Sound Designer, Amp and FX Modeler (not his official title yet covers much of what he does) has mentioned this feature has been contemplated and has not been dismissed. I do not know where it may be on Line 6's feature To Do List. I am optimistic that something like may well appear in future f/w updates, now that the Helix Core effort has been largely completed and is ready to be deployed in forthcoming f/w update(s). 

 

To be clear, I do not speak on Line 6's behalf and I have no inside knowledge. I only know what Ben and Line 6 have posted and stated at NAMM Events, etc. 

[codamedia]

15 hours ago, hideout said:

... or maybe one power amp module that's reconfigurable to those topologies?

 

I do like the idea of one "power amp block"... rather than separating every single amp. 

1. It could be as simple as choosing tube options (El84, 6V6, Single Ended, 6L6, EL34, KT66)...
2. or as complicated as choosing PI Options (Cathodyne, Long Tail, etc...), Tube Options (as mentioned earlier) and Output Transformer Options.
3. With option #2, if you knew much about the amp you like you could built your power amp to match. EG: Long Tail PI, EL34 Tubes, Partridge Transformer would give you a close replication of a Marshall Plexi or JCM 800.

In combination with the pre-amp models this would be very useful. 

 

[hideout]

8 minutes ago, codamedia said:

I do like the idea of a "power amp block"... rather than separating every single amp. 

1. It could be as simple as choosing tube options (El84, 6V6, Single Ended, 6L6, EL34, KT66)...
2. or as complicated as choosing PI Options (Cathodyne, Long Tail, etc...), Tube Options (as mentioned earlier) and Output Transformer Options.
3. With option #2, if you knew much about the amp you like you could built your power amp to match. EG: Long Tail PI, EL34 Tubes, Partridge Transformer would give you a close replication of a Marshall. 

In combination with the pre-amp models this would be very useful. 

 

 

 

Sounds kinda like this idea I posted - https://line6.ideascale.com/a/dtd/Amp-Builder-Creator-App/841518-23508#idea-tab-details

[codamedia]

7 minutes ago, hideout said:

Sounds kinda like this idea I posted - https://line6.ideascale.com/a/dtd/Amp-Builder-Creator-App/841518-23508#idea-tab-details

 

Yes - in case I muddied the waters, I was agreeing with you :) 

I have upvoted your idea! 

[hideout]

33 minutes ago, codamedia said:

 

Yes - in case I muddied the waters, I was agreeing with you :) 

I have upvoted your idea! 

 

Nah, man. You were clear.  Thanks for the upvote!

The more I think about it, the more it makes sense to just model the power sections.  I think you have the better idea.

I love it!

[Lachdanan0121]

I would love separate power amp blocks. However, after reading above, I would have to concur. I would prefer a single block, with all the adjustable parameters (mentioned in above posts) in that single block.

 

This leads to much more robust, and flexible automation options when recording. I would like that. I even automate slight differences with the ir block, by automating between similar IRs. (because one IR sounds better on this lick, and the other IR sounded better on that lick)

 

I think this route could lead to some cool tonal opportunities.

 

I also think it would be better to get the power amp block (like stated above), instead of drip fed specific power amp models over a long period of time.

[arislaf]

Totally agree. Even power amps separated would give as some crazy combinations.

[roscoe5]

There are some great separate rack preamps and power amps from Marshall, Mesa, Engl, VHT/Fryette, etc.  It would make sense to model them separately as new models. How about a Helix "Rack 'em up" update ;)

[BigJayBrian]

I play/monitor myself through a power amp which ads some grit and color... That the xlr signal that I send to the pa lacks.

Having a separate pre- and poweramp and the possibility to put a "send" in between could very usefull in that case, and I'm sure there must be many more uses for it

[phil_m]

On 2/15/2019 at 5:34 PM, hurghanico said:

Considering which is the most powerful amp/fx modeling device currently on the the market since it could have the most resources to do very complex tasks it could be interesting to know what its designer Cliff Chase said on the same separate-power-amp topic, since probably it is the same reason why also Line6 didn't implement separate power-amps in their devices:

 

The Amp block models the preamp as well as the power amp. Parameters for both are provided. The virtual power amp modeling can be deactivated but there's no way for the user to separate both sections.

"It's not impossible but it has implementation difficulties. The main problem is that the amp block is nonlinear and therefore oversamples the data. Any effect inserted between the virtual preamp and power amp would need to also run at the oversampled rate which means many times the CPU usage. For example, if the amp block is running 8x oversampled then the CPU usage for any effect inserted would by 8x as much (I'm not going to disclose our actual oversample rate). The other way is to dowsample back to native sample rate, run the effect(s), and the upsample again. No problem right? Except the no-free-lunch theory gets in the way. Downsampling and upsampling add latency."

If only a power amp model is required, for example when connecting an external preamp, the Tube Pre amp model can be used, with its tone controls at default.

"If you only need power amp modeling, for example when using an external pre-amp through the Axe-Fx, use the Tube Pre model. It uses the Vintage tonestack which is flat when the tone controls are at noon."

original article here: http://wiki.fractalaudio.com/axefx2/index.php?title=Amp_block#Preamp_and_power_amp_sections_can_not_be_split

----------------------------------------------

the following short and interesting article explains in a quite easy manner why oversampling is needed in amp sims:

http://www.earlevel.com/main/2017/05/26/guitar-amp-simulation/

 

That makes sense, and I think the whole oversampling thing would come into play for Line 6 as well. Ben Adrian gave a different explanation as to why they don't have the power amps modeled separately in this post: https://www.thegearpage.net/board/index.php?posts/23845504/

 

Quote

Here's the scoop. As one might imagine, we model the preamp and the power amp sections of amps. What is not readily apparent to most users is that we are also modeling the power supply. The audio in the amp moves through preamp and into the power amp, as we all know. The power supply feeds the amp in reverse. It delivers power to the output transformer first, then the power tube screens, then the phase inverter, then the preamp from back to front. This means that preamp sag depends on how the power amp is performing.

In a single amp model, we can have bi-directional communication. In a preamp only, we have a behind the scenes trick to maintain the sag as if it was hooked up to a power amp, without using the DSP resources of having a power amp model. There's currently no way in our architecture to have a power amp model talk in an audio-backwards direction to a preamp model. Also, if there are two preamp models and two power amp models, there's has to be a bit of code to untangle the logic of which preamp is controlled by which power amp. In addition to that, every preamp is different, so there has to be some kind of logic to untangle what to do when a preamp that needs four nodes of power supply is only hooked up to a power amp with three nodes of power supply.

So it's not impossible, and it's definitely on our list of stuff to do, but it's much more complicated than most people think to mush two amp halves together and have the new system perform realistically.

 

 

[mookyshoots]

I think the more useful approach would be to model some of the more popular tube rack power amps (Marshall, Mesa, VHT/Fryette, Soldano, Rivera, Engl). This way, you could build a "virtual rack" in Helix or as I would love, run an actual tube preamp into the modeled tube power amp for recording and thus eliminate the need for a load box. Yes, not the same as a full head but close enough.