I have already started a thread in tubes section that has as purpose finding a way to simulate the "special coloured" "tube sound".There you're all invited to give some information and ask for too, about what tubes do with the signal and how this affects the listener's perception.
This thread is meant to be a brother of that. Here you can too to post anything that can be done with DSP. There we already mentioned some of those things, so I'm asking you all who want to post, first do a scan there to have an ideea.
Thankyou all and respect.
This thread is meant to be a brother of that. Here you can too to post anything that can be done with DSP. There we already mentioned some of those things, so I'm asking you all who want to post, first do a scan there to have an ideea.
Thankyou all and respect.
I find it amusing that it takes a DSP chip with half a million transistors to try and emulate technology 80 years old. Besides, theres no fun in watching a DSP chip.
man...
You've already got it wrong here. Sorry to say that but isn't this a little to arrogant to start with?
Yes is seems like with a n-milon transistors DSP circuit could approach tube's effects. But could would you emulate at same quality, simplicity, size and efficiency a DSP circuit with tubes instead of transistors? Wich way is better in your opinion?
And another thing, with all the respect (I'm yunger than you)
please, be compliant and make a concrete point, we're not discussing here witch can be better we're trying to make both better.
Chating about tubes and emulating their effects is a good thing because it needs that those effects to be describled and may help some ppl understand them better. If you're not interessed you may find other things on this forum, or you coud put your critical ideeas in another light.
You've already got it wrong here. Sorry to say that but isn't this a little to arrogant to start with?
Yes is seems like with a n-milon transistors DSP circuit could approach tube's effects. But could would you emulate at same quality, simplicity, size and efficiency a DSP circuit with tubes instead of transistors? Wich way is better in your opinion?
And another thing, with all the respect (I'm yunger than you)
please, be compliant and make a concrete point, we're not discussing here witch can be better we're trying to make both better.
Chating about tubes and emulating their effects is a good thing because it needs that those effects to be describled and may help some ppl understand them better. If you're not interessed you may find other things on this forum, or you coud put your critical ideeas in another light.
Sorry, but I think you need to take one step back and first determine if there is a tube coloured sound or not, for SE amps I agree that they add so much distorsion that it has an effect on sound and it shouldn't be impossible to simulate that using DSP or other techniques.
However for many other type of tube amps it is debatable if they colour sound or not compared to SS amps, there are many indications of the opposite, for instance it seems that only the very best SS amps give a sound quality as a reasonably good tube amp, this points to that tube amps give less colouration and not more.
For other than HIFI applications there have already been amps designed using SS technique that try to simulate tube sound, if I remember correctly Marshall or some other guitar amplifier manufacturer is selling this already using DSP technique but in this case I think the goal is to mimic the clipping behavior including the sag of a tube rectifier in the power supply.
Regards Hans
tubetvr
Tubes can colour the sound; this is right isn't it? I do not intend to reproduce a specific kind of tube amplifier. I want to go further and sum all benefical effects that were encountered at diffrent kind of tube amplifiers and than when I'll have all the effects that tubes can produce than I'll start putting the DSP plugin together.
You got the ideea?
Tubes can colour the sound; this is right isn't it? I do not intend to reproduce a specific kind of tube amplifier. I want to go further and sum all benefical effects that were encountered at diffrent kind of tube amplifiers and than when I'll have all the effects that tubes can produce than I'll start putting the DSP plugin together.
You got the ideea?
Yes, of course as can transistors but my point is that maybe in many cases it is so that tube amps colour sound less than SS amps, how would you simulate that using DSP? you will still, (as I understand it) use a SS power amp after the DSP.
Yes, I understand completely, you have already made up your mind that tube amps colour the sound in some way more than a SS amp and you want to simulate that, but as I said this only works if your assumption is true, if the beneficial effect of many tube amps is that they don't colour the sound, what are you going to simulate? an amp without colouration? how are you going to do that?, (especially if connected to a SS amp).
Regards Hans
Yes it should be possible in theory to emulate the sound of a tube amp with signal processing. As to how someone would go about doing that is beyond my skill level. Its documented that tubes produce even order distortion where silicon produces odd order. The sound of even order is described as warmer and less harsh. Maybe theres an SDK or something available for programming DSPs on high end sound cards?
Of course there is and I'm glad you brought that. 4 example I use a Sound Blaster Creative Live as hardware device within its EMU10k1 sound processor. As software the KXProject Drivers. They come also with the SDK. If anyone here handy with C++ (unfortunately I'm not) than you could program almost anything.
Here's the link for the drivers (they also run with any sound card that has EMU processor: Audigy series of cards, and E-mu Systems' Audio Production Studio sound-card):
http://kxproject.lugosoft.com/index.php?skip=1
That's their main page if anyone interessed.
i think to get this past the point it's already been discussed to, go and read this, the original posting (at the moment it seems like the same old stuffs being repeated here again).
http://www.diyaudio.com/forums/showthread.php?s=&threadid=35496&perpage=10&pagenumber=1
http://www.diyaudio.com/forums/showthread.php?s=&threadid=35496&perpage=10&pagenumber=1
Every circuit has a transfer function. Even a simple wire. These transfer functions can be measured. So, measure the transfer function of whatever tube amplifier you wish to emulate, and your transistor based DAC/amp. Subtract the two. That should yield a transfer function which will make your DAC/amp sound like the tube amp. Implementing a convolution by transfer function in a DSP is about the most basic thing you can do in a DSP. The code should be readily available.
that is actually a good idea... while you must understand that it is still not going to give a true tube sound because of the highly reactive nature of a tube circuit making it imposible to put in the digital world because of the literally infinite number of variations that occour within a short amount of time and the sometimes unpredictable nature of these variations.
although it is a good idea because it will give a much more tube like sound than the original model proposed by polimorph.
cheers
Question, how do you simulate a Halcro amp using DSP or any other technology? It must be assumed that you don't have access to a Halcro amp or even better amp after the DSP as power amp.
I think this task in some ways is similar to the task of simulating a good tube amp, (not SE) that have very low subjective distorsion.
Regards Hans
I think this task in some ways is similar to the task of simulating a good tube amp, (not SE) that have very low subjective distorsion.
Regards Hans
Hi,
Ask yourself whether it is the amplifying device that's causing the distortion you're after or the topology it is used in?
From what I've read elsewhere so far, you want a particular harmonic distortion spectrum that seems to be correlated to tubes, to be precise: triodes....
Wouldn't it then be a good idea to know what that distortion spectrum is made of before even trying to simulate something out of the blue?
Cheers, 😉
Ask yourself whether it is the amplifying device that's causing the distortion you're after or the topology it is used in?
From what I've read elsewhere so far, you want a particular harmonic distortion spectrum that seems to be correlated to tubes, to be precise: triodes....
Wouldn't it then be a good idea to know what that distortion spectrum is made of before even trying to simulate something out of the blue?
Cheers, 😉
Cameron
Where should I look for that code? And how could I implement the code if I don't know much about C++.... This would be a tough task for me.
Thanks, maybe I'll try to learn something about programming.
Where should I look for that code? And how could I implement the code if I don't know much about C++.... This would be a tough task for me.
Thanks, maybe I'll try to learn something about programming.
Cameron said:
This is overly simplified. Convolution as I know it deals only with linear/filtering effects. A linear system may be described fully by its impulse response in both the time domain and the frequency domain.
A tube amplifier is non-linear, meaning that a measured impulseresponse fed to a convolver will not give the desired result.
most likely, a general tube amplifier may be described by sections of
convolve->nonlinear->convolve->nonlinear...
If you investigate forums for music dsp you will find that the area of tube simulation is complicated and no textbook solutions seems to work very well..
regards
Knut Inge
- Status
- Not open for further replies.