|
|
|
07-12-2007, 10:47 AM
|
#1
|
Human being with feelings
Join Date: Jul 2007
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
|
Normalization (FIXED)
After recording from a mic or inserting a media file, normalizing the clip almost always takes the levels above 0dB (sometimes as much as +3dB). Shouldn't Reaper find the highest peak and move it to no more than 0dB?
Of course I can re-adjust the normalization by dragging down on the clip but I think that this must be a bug. Anyone else find this?
FIXED (1.885)
Last edited by Dstruct; 08-15-2007 at 04:35 AM.
|
|
|
07-14-2007, 08:22 PM
|
#2
|
Human being with feelings
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 8,062
|
It seems that the track's volume fader position is also taken into account...Doesn't seem very logical to me... So if you have the track fader at less than 0db, the clip's volume will be normalized to over 0db...
-X
|
|
|
07-14-2007, 08:26 PM
|
#3
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by bachstudies
Can anyone confirm this bug? I really want to support this software by buying a one of the licenses. I also want to recommend this product to my recording musician friends but I can't right now given the normalization errors. If it's not actually a bug just let me know!
|
No doesn't happen here. Works fine, never had any issues.
Are you sure all the faders are at 0dB, no FX?
Also what file type is it wave/mp3/ogg/wavepack?
|
|
|
07-14-2007, 08:30 PM
|
#4
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by Xenakios
It seems that the track's volume fader position is also taken into account...Doesn't seem very logical to me... So if you have the track fader at less than 0db, the clip's volume will be normalized to over 0db...
-X
|
No, at least here it doesn't.
BTW it is also independent of the clip volume hanbdle, so if this is in the +dB range it will also give you overs.
|
|
|
07-14-2007, 08:33 PM
|
#5
|
Human being with feelings
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 8,062
|
Well, in any case, the results of normalization in Reaper seem a bit too inconsistent...
It should work the same every time, that is, look for the maximum peak in the clip (not the whole file, of course!) when it's processing and bring up the peak in the clip to 0db (or perhaps preferably, a user defined level)...Track faders or fx or things like that should not be taken into account, only the raw audio data in the audio file...
-X
|
|
|
07-14-2007, 08:50 PM
|
#6
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by Xenakios
Well, in any case, the results of normalization in Reaper seem a bit too inconsistent...
It should work the same every time, that is, look for the maximum peak in the clip (not the whole file, of course!) when it's processing and bring up the peak in the clip to 0db (or perhaps preferably, a user defined level)...Track faders or fx or things like that should not be taken into account, only the raw audio data in the audio file...
-X
|
AFAIK is this exacly how it is currently handled and I see no problem what so ever with it here ...
|
|
|
07-14-2007, 09:02 PM
|
#7
|
Human being with feelings
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 8,062
|
I don't know what I am doing wrong but I very often get undesired results from normalization, too loud or too soft...It's difficult to say when exactly the wrong results appear, as there are so many variables to take into account.
I could try and go through the most common scenarios and report my findings.
-X
|
|
|
07-14-2007, 09:09 PM
|
#8
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by Xenakios
I don't know what I am doing wrong but I very often get undesired results from normalization, too loud or too soft...It's difficult to say when exactly the wrong results appear, as there are so many variables to take into account.
I could try and go through the most common scenarios and report my findings.
-X
|
There might be the following things:
- resampling (file isn't at the same samplerate as project) (dunno for sure though, havent tested)
- file type e.g. mp3 or ogg(?)
- faders or volume handles
- FX
all these could result in the file appearing louder in the mixer.
|
|
|
07-15-2007, 09:16 AM
|
#9
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by bachstudies
It is true that my importing of mp3 files (probably encoded at less than my project settings) could have something to do with it but this difference doesn't seem to affect how Sonar / Cubase deal with imports so why Reaper? The only program that I've seen that really screws with imports at different resolutions etc is Krystal Audio which raises the pitch of some of my mp3s by a tone!
In any case, the first time I used Reaper and noticed the normalization errors was recording into the built-in laptop mic recording at 24-bit/48khz. I hadn't touched the faders / vol controls and normalization went WAY too hot (around +3dB max).
There's definitely something wayward about the normalization process and I continue to see this in most of my recording/importing tasks. Sometimes, yes, the peak reaches exactly 0dB but it is a lottery!
|
Okay mp3 will after encoding always (due to it's encoding) give you a louder output than what the original was. Other software might convert them back to PCM first and then look for the peaks, and I partly agree with you that REAPER should also track the correct maximum, but however this might take mp3s longer to load. Mp3s shouldn't be used for recording nor procession anyway.
Regarding Resampling: It depends on the resampling mode, e.g. point sampling (with no filtering) will give you the exact maximum peaks, but will sound bad, other modes with filtering may have some ripple in their passband and therefor occasionally give you overs up to 3 or 4dBs. Other software might do the normalization AFTER the resampling, but again this will consume CPU and time.
P.S. Does Sonar even support mp3 import at all?
EDIT: Remember everything in REAPER is done on the fly, including decoding opf files (like Mp3) or resampling, not like in other software where it gets converted on import.
|
|
|
07-15-2007, 09:25 AM
|
#10
|
Human being with feelings
Join Date: Jun 2007
Location: Gauteng, South Africa
Posts: 271
|
I noticed the same problem. I only ever use 44K 16BIT wavs (reaper is only my mixing tool)
I'm not happy about the way the normalizer works either. I will try this fader @ zero type workaround, but its definately something I'd like to see resolved.
I usually Normalize immediately after inserting the media, and I still get erratic Overs.
Cannot emphasize how dissapointed I am at either the gap in my education on Normalizing or the way it is implemented here.
With much respect!
art
|
|
|
07-15-2007, 09:49 AM
|
#11
|
Human being with feelings
Join Date: Jun 2006
Posts: 22,567
|
never had a problem with normalize here -- sounds to me pretty obvious there is something else in the chain which is causing the apparent volume increase.
|
|
|
07-15-2007, 10:04 AM
|
#12
|
Human being with feelings
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 8,062
|
No, the wrong volume increase (or decrease) does not come from anything in the signal chain, it can even be seen on the normalized item's peak graphics that it isn't always working correctly. I'm sorry that I am unable to give the exact circumstances when this happens, I did not yet start a categorical testing procedure...
I can understand that Reaper is maybe trying to be "smart" and apply the normalization based on other factors also than only the audio data in the audio file, but it can get a bit confusing for the user...
Proposed solution : a new command "Normalize item audio as a new take". This new command would only scan the peak from the audio data from the file within the range of item, and then calculate a new file with the appropriate gain applied. (And maybe it should also neutralize any gain-related item settings also for the new take!)
-X
|
|
|
07-15-2007, 10:13 AM
|
#13
|
Human being with feelings
Join Date: Jun 2006
Posts: 22,567
|
unneeded, unnessecary clutter, imo.
show real proof there is a problem and im sure justin could fix it without that menu clogging option
|
|
|
07-15-2007, 10:15 AM
|
#14
|
Human being with feelings
Join Date: Jun 2006
Posts: 22,567
|
for instance, send a wav or mp3 of the exact file that is normalizing incorrectly, along with the rpp of the project maybe.
|
|
|
07-15-2007, 11:36 AM
|
#15
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,721
|
Quote:
Originally Posted by Jason Brian Merrill
for instance, send a wav or mp3 of the exact file that is normalizing incorrectly, along with the rpp of the project maybe.
|
yep, this is what we'd need to reproduce
|
|
|
07-15-2007, 01:05 PM
|
#16
|
Human being with feelings
Join Date: Jul 2007
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
|
OK - I imported the first song (Wish I could) from the latest Norah Jones album (Not too Late). I encoded it at a variable bitrate avg of 177 @ 44.1kz and after normalizing in Reaper the max value is +0.3dB on the right channel and 0.0dB on the left.
There are more obvious ones that I can find and post and I also agree with Xenakios that the waves peaks after normalization sometimes go past the edge of the track visually, proving again that it is too hot.
I also repeat that this problem is not just about importing. Recording from mics directly into reaper also makes this happen. I have tried it with both on my parents' bog-standard laptop built-in mic and on my own system using m-audio 2496, behringer preamp and sm58 beta.
I also agree with artkau that this implimentation of normalizing is definitely not right and that perhaps reaper should follow the lead of sonar/cubase/soundforge/name any other DAW where they actually take time to calculate the peaks and process rather than appearing as realtime FX.
|
|
|
07-15-2007, 01:06 PM
|
#17
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Here two files where the normalize function seem to fail: http://www.outburst-audio.com/misc/normalize-me.zip
Saw Up A1 55Hz.wav -> Reaper's meters read +1.9dB when normalized
Square A1 55Hz.wav -> Reaper's meters read +0.2dB when normalized
Last edited by Dstruct; 07-16-2007 at 03:57 AM.
|
|
|
07-15-2007, 01:16 PM
|
#18
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Apart from that there's another bug in Reaper's level-meters in the case of repeating a loop-selection!
See the attached project:
-> play
-> read the meter(s)
=> +1.5 dB !? (reaper "reads" something from the first item - which it shoudln't do)
-> move the second selected item more to the right (a second or so)
-> set the loop-selection to it again
=> -8.8 dB (ok, fine)
conclusion: SEAMLESS metering needed?
Last edited by Dstruct; 08-28-2007 at 11:56 AM.
|
|
|
07-15-2007, 01:18 PM
|
#19
|
Human being with feelings
Join Date: Jun 2006
Posts: 22,567
|
Quote:
Originally Posted by Dandruff
Attached two files where the normalize function seem to fail:
Saw Up A1 55Hz.wav -> Reaper's meters read +1.9dB when normalized
Square A1 55Hz.wav -> Reaper's meters read +0.2dB when normalized
|
confirmed. (after resample -- )
does it show as clipped in a wav editor
??
Last edited by Jae.Thomas; 07-15-2007 at 04:10 PM.
|
|
|
07-15-2007, 01:19 PM
|
#20
|
Human being with feelings
Join Date: Jun 2006
Posts: 22,567
|
Quote:
Originally Posted by Dandruff
Apart from that there's another bug in Reaper's level-meters in the case of repeating a loop-selection!
See the attached project:
-> play
-> read the meter(s)
=> +1.5 dB !? (reaper "reads" something from the first item - which it shoudln't do)
-> move the second selected item more to the right (a second or so)
-> set the loop-selection to it again
=> -8.8 dB (ok, fine)
conclusion: SEAMLESS metering needed?
|
jeez dandy, can ya make another thread? hehe you crazy german bug man!!!!
(glad youre on our side)
|
|
|
07-15-2007, 02:05 PM
|
#21
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,721
|
Quote:
Originally Posted by Dandruff
Attached two files where the normalize function seem to fail:
Saw Up A1 55Hz.wav -> Reaper's meters read +1.9dB when normalized
Square A1 55Hz.wav -> Reaper's meters read +0.2dB when normalized
|
Hmm if I set my audio device to 48khz to match the sources, everything works as expected. If I set it to 44khz, then it does go over. hello, passband ripple (both square and saw have a _lot_ of harmonics)!
So then I turn it to plain linear interpolation, no longer goes over! yes, this confirms passband ripple to me
|
|
|
07-15-2007, 02:10 PM
|
#22
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Quote:
Originally Posted by Justin
Hmm if I set my audio device to 48khz to match the sources, everything works as expected. If I set it to 44khz, then it does go over.
|
i get the overs as written above also when having my card at 48kHz (i'm always running this samplerate) ...
|
|
|
07-15-2007, 04:06 PM
|
#23
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by Dandruff
Attached two files where the normalize function seem to fail:
Saw Up A1 55Hz.wav -> Reaper's meters read +1.9dB when normalized
Square A1 55Hz.wav -> Reaper's meters read +0.2dB when normalized
|
Normalizes jsut fine here (if done right, i.e. no resampling and other stuff involved).
|
|
|
07-15-2007, 04:12 PM
|
#24
|
Human being with feelings
Join Date: Jun 2006
Posts: 22,567
|
Quote:
Originally Posted by Justin
Hmm if I set my audio device to 48khz to match the sources, everything works as expected. If I set it to 44khz, then it does go over. hello, passband ripple (both square and saw have a _lot_ of harmonics)!
So then I turn it to plain linear interpolation, no longer goes over! yes, this confirms passband ripple to me
|
confirmed!!!
set to 48000, no clip. 44100, same as i got above.
looks like its because of something else in the chain in this example....
|
|
|
07-15-2007, 04:14 PM
|
#25
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by Dandruff
Apart from that there's another bug in Reaper's level-meters in the case of repeating a loop-selection!
See the attached project:
-> play
-> read the meter(s)
=> +1.5 dB !? (reaper "reads" something from the first item - which it shoudln't do)
-> move the second selected item more to the right (a second or so)
-> set the loop-selection to it again
=> -8.8 dB (ok, fine)
conclusion: SEAMLESS metering needed?
|
Works fine here (at 32x64 buffers though).
I assume this is due to the fact that REAPERs looping is only block accurate, but hey I'm never ever gonna loop tiny little 0.018 second snippets anyway ... anyone?
|
|
|
07-15-2007, 04:21 PM
|
#26
|
Human being with feelings
Join Date: Jul 2007
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
|
Just tried the Norah Jones song (Wish I could) in Soundforge. It analyses the peak value as -0.216dB which then normalizes correctly to 0dB. Reaper also plays this file back with the correct original peak but over-normalizes by increasing the levels upwards by 0.51dB. It just doesn't make sense! Reaper must be interpreting the peaks wrongly...
And I still don't see any explanation as to why recording with a mic would also cause these problems, unless my parents soundcard which no doubt records at max 16-bit is causing problems in a 24-bit project file (although I think i remember trying 16-bit project too).
Well, I hope with others experiencing the same, we can get to the bottom of this bug.
|
|
|
07-15-2007, 04:27 PM
|
#27
|
Human being with feelings
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 8,062
|
Are you guys always applying normalization to a whole audio file? My problems mostly come when I use normalize on items that are smaller parts of a longer file...(After normalizing such items, the gain can be several decibels higher than expected.) I suspect Reaper somehow sometimes loses track of what it has to look for to calculate the amount of normalization for cut items.
-X
|
|
|
07-15-2007, 10:23 PM
|
#28
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Quote:
Originally Posted by LOSER
Normalizes jsut fine here (if done right, i.e. no resampling and other stuff involved).
|
why doesn't it work properly here?
see the screenshot:
Last edited by Dstruct; 08-28-2007 at 11:56 AM.
|
|
|
07-15-2007, 10:31 PM
|
#29
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Quote:
Originally Posted by LOSER
Works fine here (at 32x64 buffers though).
I assume this is due to the fact that REAPERs looping is only block accurate, but hey I'm never ever gonna loop tiny little 0.018 second snippets anyway ... anyone?
|
32x64 buffers? i've tested it at 256 samples buffersize.
the thing is: i want proper display of such things when editing at sample level.
|
|
|
07-15-2007, 10:48 PM
|
#30
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Quote:
Originally Posted by Xenakios
Are you guys always applying normalization to a whole audio file? My problems mostly come when I use normalize on items that are smaller parts of a longer file...(After normalizing such items, the gain can be several decibels higher than expected.) I suspect Reaper somehow sometimes loses track of what it has to look for to calculate the amount of normalization for cut items.
-X
|
confirmed. very weird things happen here when normalizing resized items.
(tested with filesamplerate = soundcardsamplerate of course)
|
|
|
07-15-2007, 10:52 PM
|
#31
|
Human being with feelings
Join Date: Feb 2007
Location: Oulu, Finland
Posts: 8,062
|
Quote:
Originally Posted by Dandruff
confirmed. very weird things happen here when normalizing resized items.
(tested with filesamplerate = soundcardsamplerate of course)
|
Thanks for confirming! Yes, and these things happen to me also when no change of samplerate or timestretch is in effect...
-X
|
|
|
07-15-2007, 11:29 PM
|
#32
|
Human being with feelings
Join Date: Jun 2007
Location: Gauteng, South Africa
Posts: 271
|
Okay, I'm lost. Ripple what?!
If things are working properly according to the Guru's I'm happy with that. Could a best practice or work around at least be published for a Dunce like me?
Is it settings that need to be changed / matched?
I'm happy to make adjustments to my brain if that is what it takes. Normalization is THAT important to me. I'm just looking for consistency.
Peace!
art
|
|
|
07-16-2007, 01:37 AM
|
#33
|
Human being with feelings
Join Date: Jul 2007
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
|
Ditto. I not convinced Reaper handles normalization anywhere near the same as other DAWs and this is definitely to its detriment.
Ripple Passband - I also want to know why only Reaper has a problem with mp3s etc as other DAWs don't.
There are just too many weird things about the way Reaper normalizes. I have several big projects coming up which rely on normalizing quickly and correctly. As of now I'll have to look to other software. From my experience in the open source world, I know that I don't have to pay money for good software (see Audacity), but with Reaper it just doesn't seem to stand up to competition.
|
|
|
07-16-2007, 03:49 AM
|
#34
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Quote:
Originally Posted by Dandruff
why doesn't it work properly here?
|
Don't know. But I just noticed you deleted the files so I can't test again.
EDIT: Okay still had them (recovered anyway).
First off you got a different file than what you send "Saw Up A1 55Hz-imported-2.wav", but I assume it is identical. So I got +1.84dB Normalization gain here, while you got +3.76dB, which is likely the problem (hmm.. almost exactly twice what I got), anyway can you try to rebuild all peaks and see what that does also try it with the .wav that you originally posted.
Last edited by LOSER; 07-16-2007 at 03:55 AM.
|
|
|
07-16-2007, 04:32 AM
|
#36
|
Human being with feelings
Join Date: May 2006
Posts: 2,373
|
Okay that seems wierd, Dandruffs file normalizes fine, actually a bit too good only -0.1dB. But after some more testing and normalizing lots of files I finally got one to mess up only by 0.3dB though. So I will also confirm this, even though it worked exactly how it should all the time before .
|
|
|
07-16-2007, 04:45 AM
|
#37
|
Human being with feelings
Join Date: Jul 2007
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
|
Ok the saw wav normalizes fine in audacity ( 0dB) but way too hot in Reaper. Definitely Reaper's problem if you ask me. I have also found instances where the normalization peaks under 0dB as well.
|
|
|
07-16-2007, 05:56 AM
|
#38
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Quote:
Originally Posted by LOSER
Okay that seems wierd, Dandruffs file normalizes fine
|
what does the normalize dialog display (gain) on your machine for the saw-wave when normalizing it?
|
|
|
07-16-2007, 06:44 AM
|
#39
|
Human being with feelings
Join Date: Jul 2006
Posts: 12,480
|
Quote:
Originally Posted by Xenakios
Thanks for confirming! Yes, and these things happen to me also when no change of samplerate or timestretch is in effect...
|
Here a project with an example (drumloop) of this problem: http://www.outburst-audio.com/misc/normalize.zip (400kB)
You see three items. Last two items are just resized duplicates of the first item.
Go and normalize each item and watch the level-meters. I've added my results in the item-names. All items don't get normalized properly here.
Last edited by Dstruct; 07-16-2007 at 06:48 AM.
|
|
|
07-16-2007, 06:59 AM
|
#40
|
Human being with feelings
Join Date: Jul 2007
Location: No longer using REAPER and no longer active in the forums.
Posts: 129
|
Confirmed your normalization values. Audacity has no such problems and peaks at exactly 0dB.
|
|
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -7. The time now is 04:21 PM.
|