AFMG Network Forum
AFMG Network Forum Index -> AFMG SysTune -> Creating a normalization curve for system tuning
Post new topic  Reply to topic View previous topic :: View next topic 
Creating a normalization curve for system tuning
PostPosted: Thu Jul 23, 2015 3:50 am Reply with quote
Niick
Member
 
Joined: 07 May 2015
Posts: 25
Location: Portland, OR USA




Ok, I hope I describe this idea accurately. First of all, I'm used to using other programs where things like mic calibration files and "house curve" or " target curves" for system tuning is a freq and dB only text file. With SysTune, if I wanted to create a target curve to use as a guideline for EQ settings for ultra high end car audio system tuning, couldn't I create that curve bu measuring the TF of a processor, and set the processor's EQ to the desired shape, then load the measured TF into the normalization plug-in? What I don't understand, is how using this method would be different, if at all, from other types of programs where the "house curve" or normalization file you choose only had freq. and dB data. Will the measured phase in the electrical domain of the processor used to create my "house curve" affect in any way the final result?

I hope that makes sense. Thanks.
View user's profile Send private message
PostPosted: Mon Jul 27, 2015 3:03 pm Reply with quote
AFMG Pedro Lima
Forum Moderator
 
Joined: 05 Jun 2010
Posts: 266
Location: Germany




Hi Niick,

Yes, you can load the measured curve from the processor and use it as template. You can also create a desired TF curve with the Advanced EQ feature.

The normalization filter will always consider phase. In some cases that's important aspect of it, in some other cases not so much. What it does there is a (mathematicall) complex division. Whereas other programs consider only magnitude.

If you are using the curve only as a template (or guideline) to match the measured frequency response, then phase does not really matter. Wink

Which file type would you like to load to use as guideline for the eq? How would you genrate it or where have you gotten it from?

The ability to load simple text files is expected to be a new feature in future versions. Which should allow entering mic calibration files more easily. But I don't how that would help you load a target curve for the EQ. Can you give more info on that? Wink

Kind Regards,
Pedro Lima
View user's profile Send private message
PostPosted: Tue Jul 28, 2015 4:36 am Reply with quote
Niick
Member
 
Joined: 07 May 2015
Posts: 25
Location: Portland, OR USA




Pedro,

Thanks for the reply, so here's what I'm up to. I've recently attended a training that focused on car audio system tuning, as this is what I do for a living, and it is in this capacity that I use SysTune. There aren't any other installers or shops that I am aware of who use anything other than a basic RTA program for system tuning. Therefore, I have no-one to learn from or who I can ask for help in these matters, which is good, because there's no on to tell me "that won't work......that's not how you do it...." Smile

I'm trying to raise the bar of what's possible to achieve in car audio systems, from a "sound quality" standpoint, and from the point of view of tuning time. So, it's pretty much universally agreed upon that a system with a perfectly flat response AS MEASURED WITH AN RTA, doesn't sound good. Taking this into consideration, and keeping in mind what I've read about how frequency dependent windowing more closely mimics human perception, I'm trying to find a means of setting the TFC window to where the resulting magnitude trace more closely mimics the way we perceive relative loudness per frequency.

Here's what I'm thinking, an amplifier SHOULD have a flat freq response, what goes in, should be what comes out. But once you move into the acoustical domain, the presence of reflections affects the overall sound that we hear, so is it possible for me to find a window time (TFC) that will, when used to measure a system from the listening position, and that system is tuned to flat, sound good? And if so, what would the graph look like in spectrum mode, where time isn't taken into consideration?

By the way, these systems I'm talking about all have signal processing that allows me to align in time each individual drive unit, to preserve things like stereo imaging and soundstage.

In order to answer this question of TFC window time, I thought I might take a car, play filtered pink noise thru it, and move thru the different bands, adjusting the EQ until I reach a point of equal PERCIEVED loudness. I would set up some type of repeatable experiment and use as many people as I can to determine these EQ settings. Once I got the system SOUNDING equally load at all frequencies, I thought then I might take a spectrum measurement, note the curve, which I have a feeling will have a higher measured bass response, and a lower treble, then, I thought I might try to start adjusting the TFC window until I get a magnitude trace that reads more or less flat, at which point, I'd note the time settings, and................

Well, I have to actually do the experiment to go any farther. Until then, The target curves I asked about were planning on being used as a way of finding a measured response that sounded good to as many people as possible, then using that response as an end goal for tuning customers cars in time sensitive situations. I find that if I can achieve some kind of pre-determined standard, then customers feel like something has been accomplished, as it's hard enough for them to grasp the concept of paying for something they can't see or touch. what I was thinking, was that it's easier to tune a system to a straight line (hence the normalization idea), other lesser programs call this a "house curve", it's a text file that you enter, once engaged, the RTA trace will subtract the interpolated curve from its reading, so that tuning to a flat line will actually result in whatever curve you entered as your "house curve".

It's an established standard in the car audio world, and although I'm working to define newer, higher standards, sometimes it's helpful to have the ability to perform the same function that others have relied upon for so long. I wasn't sure if using the measured TF of a processor for normalization would have an unintended affect on the magnitude trace, since there is phase info in the processor's TF. In other words, when other programs use frequency only data as a normalization file, would the presence of the processors phase response change the resulting measurement beyond just being the inverse of the magnitude, it's so hard for me to put into words. I have a lot to learn.

Thanks for all your help.

Edit: I don't know why it never dawned on me to just use the advanced EQ function to create my target curves! Duh!!

Sincerely, Nicholas C. Ames
View user's profile Send private message
AFMG Network Forum Index -> AFMG SysTune
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
All times are GMT  
Page 1 of 1  

  
  
 Post new topic  Reply to topic  


Powered by phpBB © 2001-2003 phpBB Group
Theme created by Vjacheslav Trushkin
Variation by CodeWeavers and AFMG