Home |
Search |
Today's Posts |
#1
Posted to rec.audio.pro
|
|||
|
|||
can convolution do this and how?
Here's the problem: I have a song I really like and would like to apply its
reverb to another song. All I have available is the original song, no IR responses or any idea of the type of studio used to produce the reverb. Is there any way I can accurately apply the reverb from the original song onto the new song with convolution? Since no IR is possible, how can I go about this? Thank you, Earl |
#2
Posted to rec.audio.pro
|
|||
|
|||
can convolution do this and how?
"Earl Sanders" wrote in message
Here's the problem: I have a song I really like and would like to apply its reverb to another song. All I have available is the original song, no IR responses or any idea of the type of studio used to produce the reverb. Is there any way I can accurately apply the reverb from the original song onto the new song with convolution? Since no IR is possible, how can I go about this? What you want to do is theoretically doable, but you probably lack the resources to actually do it. What you are trying to do is obtain the impulse response of the virtual or real room that was used to apply reverb to the song you are using as an example. There are a number of ways to obtain the impulse response of a room. One of them is to actually apply an ideal impulse to a room, and measure the room's response to that impulse at some location, which is straight-forward enough. Another way to obtain the impulse response of a room is to apply some known signal to a room, measure the response of the room to that signal, and then do some higher math to find just the impulse response of the room. Part of the higher math is a detailed analysis of the signal before reverb was added to it. The signal you use for this purpose could be music or pink noise or some other signal, but it needs to contain significant amounts of energy at all the frequencies of interest. In your case you are kinda screwed because you don't know exactly what your example song was like before reverb was added to it. That's the key resource that you probably lack. |
#3
Posted to rec.audio.pro
|
|||
|
|||
can convolution do this and how?
Earl Sanders wrote:
Here's the problem: I have a song I really like and would like to apply its reverb to another song. All I have available is the original song, no IR responses or any idea of the type of studio used to produce the reverb. Is there any way I can accurately apply the reverb from the original song onto the new song with convolution? Since no IR is possible, how can I go about this? No, but by listening you can get a sense of how the original reverb sounds, and you can sit down and A-B with your source and reverb unit and try and make them sound the same. Does it sound like a real room? Like a shiny chamber? Like a plate? Is the reverb long or short? Is it bright or dull? There are a lot of parameters so you need to take them one at a time. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#4
Posted to rec.audio.pro
|
|||
|
|||
can convolution do this and how?
"Earl Sanders" wrote in message
Here's the problem: I have a song I really like and would like to apply its reverb to another song. All I have available is the original song, no IR responses or any idea of the type of studio used to produce the reverb. Is there any way I can accurately apply the reverb from the original song onto the new song with convolution? Since no IR is possible, how can I go about this? After looking at the other responses to this point, I think that some of us told you that you were up the creek without a paddle, but we didn't tell you how to find a paddle. Scott did a nice job of telling you what paddles look like. last weekend I was visiting one of the older studios in this city, and listened to their senior guys talk about how they add reverb. The room was dead, but their recordings aren't. First they build up three sets of reverb that sound like what they are looking for in that kind of reverb. One is short reverb (0-8 mS delays) one is medium reverb (8-30) and one is long (30-200 or more). They adjust the delays and the timbre of each kind of delay as desired. Then they add various amounts of each kind of delay, again as desired to get the overall sound they want. Short reverb represents surfaces that are close to the mic and source. Things like floors and nearby walls. Medium reverb represents nearby walls and perhaps the ceiling. Things like the walls of the stage or shell that the virtual performance is in. Long reverb represents things like the other sides and very high ceilings of a very large room. So when you listen to the recording that you are trying to clone the reverb of, think about the room that you hear the music happening in. Is the floor carpeted or polished wood? Is your back to the wall or is it 30 feet back? That relates to short reverb. Carpeting absorbs highs and polished wood reflects it. Do you need to mute the highs in short reverb? Are the nearby surfaces walls draped, plastered or glass? Are they square or angled? That relates to medium reverb. Drapes absorbs highs and glass reflects it. There might also be some mid-bass boom or low-end thickening to of reflections of this kind. Is this taking place in a big room, and if so, is the room a concert hall or a basketball court? That relates to long reverb. |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
What is convolution Reverb | Pro Audio | |||
Convolution Engine | Pro Audio | |||
Convolution Engine | Tech | |||
Plate convolution | Pro Audio | |||
Convolution Reverb q | Pro Audio |