Understanding Signal Flow
Guys,
With the issue of my USBi out of the way, I would like to ask a question more related to actual signal processing.
Although I am no novice to audio and sound, I am just now trying to sharpen my actual signal processing chops. As I look through the tutorials in SigmaStudio, most things I understand (I think), but there are a few that I don't really understand. I would like to kind of talk through what I understand the signal flow of the "Basic Stereo In/Out" example project included with Studio and have you correct me where I'm wrong and answer a couple of questions.
Okay, so the schematic starts off with the standard input block that is routed into an individual volume slider to control the input level.
The signal exiting the volume slider goes to T connetors where it is split into two equal stereo signals.
One stereo signal is sent, unchanged, to a high pass filter.
The second has both channels added together into one mono signal, and is fed into a low pass filter.
The outputs from those filters, are respectively, routed to individual EQs taylored to their specific frequency ranges.
The outputs from the EQs are routed to independant compressors, were a mild compression ratio is applied at signal levels above aprox. -10dB.
Those signals are sent to independant, one sample delay blocks.
The output of the delay blocks has the "high" signal being summed with mono "low" signal and sent to the DACs.
Am I understanding it correctly?
Questions
In step 4, I know that bass is "non directional" but, what is the advantage of processing this signal as a mono, sum of channels, instead of a stereo signal?
In step 5, being that these signals have just exited filters, what is the advantage of using the "shelf" bands on the EQ blocks? Weren't those frequencies just filtered?
In step 7, what is the purpose of delaying all of the signals by one sample?
Thanks, in advance for any insights shared...
Anyone?
Bueller??
My comments are below in red.
mdot wrote:
Questions
In step 4, I know that bass is "non directional" but, what is the advantage of processing this signal as a mono, sum of channels, instead of a stereo signal?
Two reasons. One: you might be outputting it to a dedicated subwoofer channel anyway, so you might as well mix it down earlier than later. Two: processing in mono requires (typically) a little more than half the DSP resources of processing in stereo.
NOTE: Splitting bass and treble frequencies allows you to do dual-band compression, which helps to avoid pumping due to a loud bass line or kick drum.
In step 5, being that these signals have just exited filters, what is the advantage of using the "shelf" bands on the EQ blocks? Weren't those frequencies just filtered?
This is basically just used as a tone control, either user-configurable or set by a tuning engineer to a hardwired value. It's simply meant to tune the system as a whole on those particular channels.
In step 7, what is the purpose of delaying all of the signals by one sample?
Actually, you'll be delaying them by more than that in practice. This is speaker delay compensation. Each speaker may be a different distance from the listener, meaning the signals from each speaker arrive at the listener's head at a different time, resulting in phase distortion. To correct for this, you delay the closer speakers so that the signals from each speaker arrives at the listener at the same time. The furthest speaker should have no delay. You'll need to do measurements in order to figure out the distance of each speaker from the listener, then calculate the required delay in samples based on the speed of sound and the sample rate.
Thanks for the explinations. It all makes sense now.
Actually, you'll be delaying them by more than that in practice. This is speaker delay compensation. Each speaker may be a different distance from the listener, meaning the signals from each speaker arrive at the listener's head at a different time, resulting in phase distortion. To correct for this, you delay the closer speakers so that the signals from each speaker arrives at the listener at the same time. The furthest speaker should have no delay. You'll need to do measurements in order to figure out the distance of each speaker from the listener, then calculate the required delay in samples based on the speed of sound and the sample rate.
So these values are basically set at a "default" value awaiting specific tuning "in the field"?
One other question (let me know if I should start another thread), I have heard the term "expander" used in different posts, but there is no function block by that name. When I think of an expander, I think of one of those BBE Sonic Maximizers or one of the old 3dbx units. Something that "adds" dynamic range to audio signals. Is that the function being described when I see it used here? If not, how is the "expander" function described here and which functional blocks are used implement it?
That's right.
Our dynamics processors allow flexible curve programming, so they can be gates, compressors, limiters, and expanders, all at the same time! Here's a link to an external website's article showing the differences between these types of curves.
A company called Rane also has some great reference materials on their site. Here's the section about Dynamics Processors. Chapter 6 shows you how Gate, Expander, Compressor, and Limiter can all be implemented in a single cell in SigmaStudio.
Please log in to reply.