Home |
Search |
Today's Posts |
#1
|
|||
|
|||
Wavegain
Dear all,
I have a recently started experimenting with wavegain (replaygain for wave files). It looks to be a very useful tool but there are some things that I am comfused about. I understand that it uses psychoacoustic analysis to measure the perceived loudness and adjust the gain on that basis. My questions: does it apply a single gain factor across the entire file or does it analyse the file on a frame-by-frame basis and apply different amounts of gain in different parts of the track. Secondly, does it apply the same gain across all frequencies or does it take into account the differing perceived loudness of different frequencies. Furthermore, I tested wavegain on a sample files that I have (a frequency sweep from 20Hz to 20KHz, all at equal amplitude). I noticed that it reduced the overall amplitude of the audio track quite significantly even though there was no chance of clipping. My question: is it possible to apply wavegain on a music file so that it uses the full range of possible values, i.e. performs some kind of normalization. I would appreciate any help on these issues. Many Thanks RG |
#2
|
|||
|
|||
"rg" wrote in message
Dear all, I have a recently started experimenting with wavegain (replaygain for wave files). It looks to be a very useful tool but there are some things that I am comfused about. Have you read: http://www.hydrogenaudio.org/forums/...opic=24527&hl= ???? I understand that it uses psychoacoustic analysis to measure the perceived loudness and adjust the gain on that basis. My questions: does it apply a single gain factor across the entire file or does it analyse the file on a frame-by-frame basis and apply different amounts of gain in different parts of the track. Secondly, does it apply the same gain across all frequencies or does it take into account the differing perceived loudness of different frequencies. Reference says: "MP3Gain then writes to the global gain header of every block, adjusting them all up or down by the same amount in order to achieve those Replaygain values." Applying different amounts of gain would be compression, which is not what the software does. Furthermore, I tested wavegain on a sample files that I have (a frequency sweep from 20Hz to 20KHz, all at equal amplitude). I noticed that it reduced the overall amplitude of the audio track quite significantly even though there was no chance of clipping. My question: is it possible to apply wavegain on a music file so that it uses the full range of possible values, i.e. performs some kind of normalization. Normalization would undo what the software does. You can have one or the other but not both at the same time. |
#3
|
|||
|
|||
Hi,
Thanks very much for your reply and thank you for the link. With respect to applying gain across differing frequency ranges, my understanding is that since perceived loudness is different for different frequency ranges (as shown in the Fletcher-Munson curve), by applying the same gain across all frequencies the actual perceived gain will be different for different frequency ranges. I was just curious how it works in practice. The reason I asked about the normalization is because I was really surprised at how much reduction in amplitude there actually was. The wavegained track had an amplitude about 1/5th of the original, and the original was already half of maximum (on a linear scale). Many Thanks, RG "Arny Krueger" wrote in message ... "rg" wrote in message Dear all, I have a recently started experimenting with wavegain (replaygain for wave files). It looks to be a very useful tool but there are some things that I am comfused about. Have you read: http://www.hydrogenaudio.org/forums/...opic=24527&hl= ???? I understand that it uses psychoacoustic analysis to measure the perceived loudness and adjust the gain on that basis. My questions: does it apply a single gain factor across the entire file or does it analyse the file on a frame-by-frame basis and apply different amounts of gain in different parts of the track. Secondly, does it apply the same gain across all frequencies or does it take into account the differing perceived loudness of different frequencies. Reference says: "MP3Gain then writes to the global gain header of every block, adjusting them all up or down by the same amount in order to achieve those Replaygain values." Applying different amounts of gain would be compression, which is not what the software does. Furthermore, I tested wavegain on a sample files that I have (a frequency sweep from 20Hz to 20KHz, all at equal amplitude). I noticed that it reduced the overall amplitude of the audio track quite significantly even though there was no chance of clipping. My question: is it possible to apply wavegain on a music file so that it uses the full range of possible values, i.e. performs some kind of normalization. Normalization would undo what the software does. You can have one or the other but not both at the same time. |
#4
|
|||
|
|||
rg wrote:
Hi, Thanks very much for your reply and thank you for the link. With respect to applying gain across differing frequency ranges, my understanding is that since perceived loudness is different for different frequency ranges (as shown in the Fletcher-Munson curve), by applying the same gain across all frequencies the actual perceived gain will be different for different frequency ranges. I was just curious how it works in practice. The reason I asked about the normalization is because I was really surprised at how much reduction in amplitude there actually was. The wavegained track had an amplitude about 1/5th of the original, and the original was already half of maximum (on a linear scale). Many Thanks, RG "Arny Krueger" wrote in message ... "rg" wrote in message Dear all, I have a recently started experimenting with wavegain (replaygain for wave files). It looks to be a very useful tool but there are some things that I am comfused about. Have you read: http://www.hydrogenaudio.org/forums/...opic=24527&hl= ???? I understand that it uses psychoacoustic analysis to measure the perceived loudness and adjust the gain on that basis. My questions: does it apply a single gain factor across the entire file or does it analyse the file on a frame-by-frame basis and apply different amounts of gain in different parts of the track. Secondly, does it apply the same gain across all frequencies or does it take into account the differing perceived loudness of different frequencies. Reference says: "MP3Gain then writes to the global gain header of every block, adjusting them all up or down by the same amount in order to achieve those Replaygain values." Applying different amounts of gain would be compression, which is not what the software does. Furthermore, I tested wavegain on a sample files that I have (a frequency sweep from 20Hz to 20KHz, all at equal amplitude). I noticed that it reduced the overall amplitude of the audio track quite significantly even though there was no chance of clipping. My question: is it possible to apply wavegain on a music file so that it uses the full range of possible values, i.e. performs some kind of normalization. Normalization would undo what the software does. You can have one or the other but not both at the same time. foobar has wavegain |