VideoHelp Forum
+ Reply to Thread
Page 4 of 5
FirstFirst ... 2 3 4 5 LastLast
Results 91 to 120 of 122
Thread
  1. A couple of minor modifications to the Now Playing tab. The Title line only displays the title when there's title info in tags. Previously it displayed the file name in the absence of title info, but there's a "File" line already displaying the file name.

    I moved the file type (mp3/aac etc) from the "File" line to the "Codec" line. I think it looks a bit neater.

    I can't decide if I prefer it, but here's a new Playlist Title column too. It no longer appends "file name" when it's displaying the file name. Instead, if a file name is being displayed because there's no Title info in tags, it's displayed in brackets.

    Code:
    Title  ( file name ) - Aligned Left
    $puts(A,$ifequal($meta_num(title),0,'(' %filename% ')',%title%))[$if(%isplaying%,>>>$get(A)<<<,$get(A))]
    Image Attached Files
    Last edited by hello_hello; 20th Nov 2018 at 06:07.
    Quote Quote  
  2. Originally Posted by hello_hello View Post
    A couple of minor modifications to the Now Playing tab. The Title line only displays the title when there's title info in tags. Previously it displayed the file name in the absence of title info, but there's a "File" line already displaying the file name.

    I moved the file type (mp3/aac etc) from the "File" line to the "Codec" line. I think it looks a bit neater.

    I can't decide if I prefer it, but here's a new Playlist Title column too. It no longer appends "file name" when it's displaying the file name. Instead, if a file name is being displayed because there's no Title info in tags, it's displayed in brackets.

    Code:
    Title  ( file name ) - Aligned Left
    $puts(A,$ifequal($meta_num(title),0,'(' %filename% ')',%title%))[$if(%isplaying%,>>>$get(A)<<<,$get(A))]

    Ah nice one, I used your previous one but I couldn't figure out how to get the default columns back lol, this foobar is complicated for some parts but I do like it, I have moved away from windows media player now lol plus how did you get the graphic equalizer bars going at the right hand side?
    Quote Quote  
  3. Are you referring to the Peak Meter? It's the Peak Meter on the right.

    I added the Splitter Element under each tab of the Tabs Element, then added various elements on the left side, and always the Peak Meter on the right.
    I like having the meter there to show me the output, but I think it's a bit fugly, and I don't want it to be intrusive, which is why I have a tab called "Meter". It's the same as the "Now Playing" tab, but with the Peak Meter large enough for the channel names at the bottom to be readable. I was sick of resizing the Peak Meter, so now I switch tabs instead.

    Actually.... my "Meter" tab is almost the same as my "Now Playing" tab, but above the Peak Meter is the R128 Meter displaying the text. It's also a bit fugly.

    FYI, If you right click on the Peak Meter element, you can run it in full screen mode.
    Image Attached Thumbnails Click image for larger version

Name:	fb2k1.gif
Views:	78
Size:	15.8 KB
ID:	47241  

    Click image for larger version

Name:	fb2k2.gif
Views:	76
Size:	31.8 KB
ID:	47242  

    Last edited by hello_hello; 20th Nov 2018 at 10:34.
    Quote Quote  
  4. Been messing about with foobar big time tonight, I have my interface the way I want it, excellent program, need to sort out font colors and stuff but amazed how fast I'm learning it, usually I wouldn't be bothered with so many things to edit and customize but deffo love it, would be good if there was an actual column out there so that you could see if a track is going to say track will be clipped during track gain alteration, but then again there is that option apply gain and prevent clipping but would be good to see it as mp3 gain does, it analyzes the track and afterward it tells you if it would be applying clipping, anyway tell me what you think, need to make some adjustments but with your now playing text, it helps a lot
    Image Attached Thumbnails Click image for larger version

Name:	Untitled.png
Views:	73
Size:	117.7 KB
ID:	47245  

    Click image for larger version

Name:	Peak.png
Views:	72
Size:	875.0 KB
ID:	47247  

    Last edited by circulationds; 20th Nov 2018 at 19:18.
    Quote Quote  
  5. Your setup is looking pretty good, although the Peak Meter displaying so many channels in the 1st screenshot confuses me. The playing file appears to be stereo. Were you using a DSP to upmix on playback, or something similar?

    I know what you mean about determining if there'll be clipping after adjusting. I generally just apply the adjustment and then check for peaks above 0dB and for those, adjust them down if need be (I don't worry about peaks just a fraction over maximum (up to around 0.5dB). Originally, I added the "Selection Properties" element to the GUI and enabled the display of ReplayGain info, because when multiple files are selected, it stops displaying any TrackGain info and only shows the highest peak for the selected files (as "total peak"). If it's greater than 1.0000, then at least one file has a peak above maximum and I'd know to look for the culprit(s). For a long time, that's how I made checking the peaks after adjusting a bit easier, but now we have the Track Peak info displaying as a Playlist column it's not too hard to check each file after adjusting. Anyway... the 1st screenshot below shows my "File Info" tab with the "selection properties" element displaying the loudest peak, although that tab is a bit redundant now. I should remove it.

    Have you played around with the Facets library viewer yet? Once you start managing your song library you'll probably want to. The two columns on the left of my fb2k setup are Facets columns. If you don't have it added to the GUI as an element, you can still use it via the Library/Facets menu. By default, it opens up something like the second screenshot below. The columns interact, so for me clicking on an Artist name in the left column causes the Title column to display just the tracks by that artist (as per the 3rd screenshot), but that's just how I prefer it. It's quite configurable. fb2k's Library Viewer/Playlist combination of doing things can take a little getting used to, but now I can't imagine doing it any other way.

    Facets has a filter that can be used to effectively divide your media library into sections if you wish to. For example, I have original CD rips in a different location to the MP3 versions. To display them as individual media libraries I've created filters that only show the contents of certain folders (and their sub-folders). It's easy as you don't have to specify the full folder path, just a section of it that's unique, so I have two Filters (configured under Media Library/Facets/Filter). One filter's syntax is something like "My Stuff/MP3s" and the second is "My Stuff/CD Rips". The Filter drop down menu makes it easy to switch between them.

    A library viewer is the way you'd normally add stuff from your media library to a playlist, because you can't play anything unless it's in a playlist. By default, Facets switches between being a Library viewer and a Playlist viewer when you select a playlist tab. I don't like that behaviour so I have it configured to always display the contents of my media library. That way, the left side of my fb2k configuration just displays the media library contents and the right side just the playlist contents. My brain seems to prefer it that way.
    Image Attached Thumbnails Click image for larger version

Name:	fb2k1.gif
Views:	82
Size:	37.7 KB
ID:	47248  

    Click image for larger version

Name:	fb2k2.gif
Views:	52
Size:	9.0 KB
ID:	47251  

    Click image for larger version

Name:	fb2k3.gif
Views:	59
Size:	31.8 KB
ID:	47252  

    Last edited by hello_hello; 21st Nov 2018 at 02:03.
    Quote Quote  
  6. Originally Posted by hello_hello View Post
    Your setup is looking pretty good, although the Peak Meter displaying so many channels in the 1st screenshot confuses me. The playing file appears to be stereo. Were you using a DSP to upmix on playback, or something similar?

    I know what you mean about determining if there'll be clipping after adjusting. I generally just apply the adjustment and then check for peaks above 0dB and for those, adjust them down if need be (I don't worry about peaks just a fraction over maximum (up to around 0.5dB). Originally, I added the "Selection Properties" element to the GUI and enabled the display of ReplayGain info, because when multiple files are selected, it stops displaying any TrackGain info and only shows the highest peak for the selected files (as "total peak"). If it's greater than 1.0000, then at least one file has a peak above maximum and I'd know to look for the culprit(s). For a long time, that's how I made checking the peaks after adjusting a bit easier, but now we have the Track Peak info displaying as a Playlist column it's not too hard to check each file after adjusting. Anyway... the 1st screenshot below shows my "File Info" tab with the "selection properties" element displaying the loudest peak, although that tab is a bit redundant now. I should remove it.

    Have you played around with the Facets library viewer yet? Once you start managing your song library you'll probably want to. The two columns on the left of my fb2k setup are Facets columns. If you don't have it added to the GUI as an element, you can still use it via the Library/Facets menu. By default, it opens up something like the second screenshot below. The columns interact, so for me clicking on an Artist name in the left column causes the Title column to display just the tracks by that artist (as per the 3rd screenshot), but that's just how I prefer it. It's quite configurable. fb2k's Library Viewer/Playlist combination of doing things can take a little getting used to, but now I can't imagine doing it any other way.

    Facets has a filter that can be used to effectively divide your media library into sections if you wish to. For example, I have original CD rips in a different location to the MP3 versions. To display them as individual media libraries I've created filters that only show the contents of certain folders (and their sub-folders). It's easy as you don't have to specify the full folder path, just a section of it that's unique, so I have two Filters (configured under Media Library/Facets/Filter). One filter's syntax is something like "My Stuff/MP3s" and the second is "My Stuff/CD Rips". The Filter drop-down menu makes it easy to switch between them.

    A library viewer is the way you'd normally add stuff from your media library to a playlist because you can't play anything unless it's in a playlist. By default, Facets switches between being a Library viewer and a Playlist viewer when you select a playlist tab. I don't like that behavior so I have it configured to always display the contents of my media library. That way, the left side of my fb2k configuration just displays the media library contents and the right side just the playlist contents. My brain seems to prefer it that way.

    Not played around with the facets yet! I will have a look into that at some point as I want to add my now that's what I call music collection to foobar and organize them albums, with the first screenshot that was a graphic equalizer instead of the peak meter, I selected the wrong one when uploading the images to show you, the files playing where my converted movie audio files, I just wanted to see the ones that where over the 1.0000 peaks. I checked the files, they seem to be ok, and I can't notice clipping in them, to be honest. and the peak meter gain when the movie was at high volume didn't show on the peak meter above 0db, it was just below so I know that there is some give in some files. I will have to look into that more but I have been using your matrix mixer to encode 5.1 to 2 channel AAC then after i do that, i scan them files add the tags then I apply 70DB to alter file volume with lower adjustments to prevent clipping according to peak information, then apply that to file content then i use the final normalization with the rocksteady plugin with the following settings.

    Amplify -3
    Full Amplification 15%
    Maximum Amplification 10DB
    Amp Gain Time 10MS
    Use Smart Limiting


    the reason why I am using the rocksteady after converting just the 5.1 to 2 channel using your matrix settings, there was a couple of files that I couldn't hear the dialogue all that well after adjusting the final volume, so I used the rocksteady to convert to final conversion and it's fine, I had to do all the 5.1 audio files again because I was on peaks at 2.0000 which where far too high and I did notice the clipping when watching films the other night. I have re-encoded them files and checked them, some have peaks just below 1.0000 and slightly over 1.0000 but when playing loud parts of the files and viewing the peak meter there not going past the 0DB mark and they sound alot better than they did before.
    Last edited by circulationds; 21st Nov 2018 at 05:30.
    Quote Quote  
  7. Are you converting to AAC, then scanning with ReplayGain and converting again to compress? If there's two conversions in your process the first one should ideally be to a lossless format, but I'm not sure I'm understanding correctly.

    To confirm... you have been downmixing and compressing at the same time, and in that order? As per the 1st screenshot below? The audio is still being downmixed before it's compressed but as a single process. The result should be the same as doing it in two stages. If you want to adjust the volume after compressing you can downmix, compress and convert to a lossless format, or you can convert directly to AAC and adjust their volumes losslessly, but to me it doesn't make sense to adjust the volume with ReplayGain before compressing as the compression will change it.

    Unless.... you're adjusting the volume to 70dB(?) before compressing because you think the Rock Steady plugin works better that way. If so you can still do it in a single step (after scanning the source files and saving the Track Gain data). The ReplayGain processing in the second screenshot would adjust the volume to 70dB first, followed by downmixing, and then compression.

    For the record, if you scan 5.1ch audio to determine the volume, it doesn't translate directly to the volume of a stereo version after downmixing. I wish there was a way to scan 5.1ch audio and use the ReplayGain info to adjust the volume by the correct amount while downmixing to achieve a particular target volume, but unfortunately it doesn't work that way. For the purpose of using the ReplayGain data to adjust the volume before compressing though, the audio would be adjusted to the same volume, so the compressor would have a similar amount of headroom to play with, which I assume is the main reason adjusting the volume before compressing.

    Personally, I'd just reduce the volume by a fixed amount, and then there'd be no need to run a Replay Gain scan before downmixig. As per the 3rd screenshot.
    I'd downmix with the Matrix Mixer's "normalise" option enabled, followed by the Amplify DSP to reduce the volume by a fixed amount, followed by the compression. To prevent the Winamp Bridge from limiting any peaks before it converts the audio to 16 bit, the loudest peak should be -6dB or less. Generally the volume for 5.1ch audio should be somewhere in the vicinity of -23 LUFS (around 83dB) so a fixed reduction of 10dB should be plenty, and it should keep the peaks safe from the Winamp Bridge's limiter. How much you reduce the volume is up to you, but I'd be trying to do it all in a single step as much as possible. There's no reason why the result should be any different to doing it in two or more steps, assuming the same volume adjustments are applied etc.
    Image Attached Thumbnails Click image for larger version

Name:	fb2k1.gif
Views:	80
Size:	16.5 KB
ID:	47255  

    Click image for larger version

Name:	fb2k2.gif
Views:	75
Size:	20.9 KB
ID:	47256  

    Click image for larger version

Name:	fb2k3.gif
Views:	76
Size:	17.0 KB
ID:	47257  

    Last edited by hello_hello; 21st Nov 2018 at 08:26.
    Quote Quote  
  8. Originally Posted by hello_hello View Post
    Are you converting to AAC, then scanning with ReplayGain and converting again to compress? If there's two conversions in your process the first one should ideally be to a lossless format, but I'm not sure I'm understanding correctly.

    To confirm... you have been downmixing and compressing at the same time, and in that order? As per the 1st screenshot below? The audio is still being downmixed before it's compressed but as a single process. The result should be the same as doing it in two stages. If you want to adjust the volume after compressing you can downmix, compress and convert to a lossless format, or you can convert directly to AAC and adjust their volumes losslessly, but to me it doesn't make sense to adjust the volume with ReplayGain before compressing as the compression will change it.

    Unless.... you're adjusting the volume to 70dB(?) before compressing because you think the Rock Steady plugin works better that way. If so you can still do it in a single step (after scanning the source files and saving the Track Gain data). The ReplayGain processing in the second screenshot would adjust the volume to 70dB first, followed by downmixing, and then compression.

    For the record, if you scan 5.1ch audio to determine the volume, it doesn't translate directly to the volume of a stereo version after downmixing. I wish there was a way to scan 5.1ch audio and use the ReplayGain info to adjust the volume by the correct amount while downmixing to achieve a particular target volume, but unfortunately it doesn't work that way. For the purpose of using the ReplayGain data to adjust the volume before compressing though, the audio would be adjusted to the same volume, so the compressor would have a similar amount of headroom to play with, which I assume is the main reason adjusting the volume before compressing.

    Personally, I'd just reduce the volume by a fixed amount, and then there'd be no need to run a Replay Gain scan before downmixig. As per the 3rd screenshot.
    I'd downmix with the Matrix Mixer's "normalise" option enabled, followed by the Amplify DSP to reduce the volume by a fixed amount, followed by the compression. To prevent the Winamp Bridge from limiting any peaks before it converts the audio to 16 bit, the loudest peak should be -6dB or less. Generally the volume for 5.1ch audio should be somewhere in the vicinity of -23 LUFS (around 83dB) so a fixed reduction of 10dB should be plenty, and it should keep the peaks safe from the Winamp Bridge's limiter. How much you reduce the volume is up to you, but I'd be trying to do it all in a single step as much as possible. There's no reason why the result should be any different to doing it in two or more steps, assuming the same volume adjustments are applied etc.

    My steps are

    convert to Apple AAC lossless file and use nothing but your matrix to downmix to stereo, once they're complete I scan them converted files and apply a gain of 70DB apply gain without clipping, then I convert them files again, ACC lossless using the Winamp bridge with the settings I mentioned above then scan them and apply a gain of 83db without clipping. I don't scan the 5.1 audio before converting with the matrix mixer, I will do your steps for the next files and see how I get on, but I had to reduce some of the bridge settings as stated above as I was experiencing clipping that's why I did them again, after files have been fully completed I do a scan again and any that are going over peak I reduce the overall volume. i will try your step for the next batch and see how they get on, all in there is only about 120 5.1 channel files so it's not a big deal and not as much time to do it as when i was doing it with pazerra followed by box4 the image below is the completed files all adjusted to 83 db and with apply gain without clipping, they seem to be ok but like i said i had to adjust winamps bridge settings to get these results. hope this makes sense i have added one of the completed conversions so you can listen and tell me if it's ok
    Image Attached Thumbnails Click image for larger version

Name:	Finished.png
Views:	71
Size:	616.8 KB
ID:	47258  

    Image Attached Files
    Last edited by circulationds; 21st Nov 2018 at 10:07.
    Quote Quote  
  9. circulationds,

    I ran some more tests myself, starting with some 5.1ch audio from a TV show, and found downmixing followed by a -6dB gain reduction, followed by my usual RockSteady settings worked quite well, resulting in peaks of around -2 dB to -3 dB. When I tried some loud audio from a couple of movies it was a different story though. One was 5.1ch DTS and the other 5.1ch AC3. I scanned them just to see, and the DTS audio had a volume of -17.62 LUFS with a Track Peak of -00.17 dB. For the AC3 audio the volume was -13.90 LUFS with a True peak of +07.03 dB. Chances are the AC3 contained info to reduce the volume on playback, but I have the AC3 decoder configured to ignore it. That's my theory anyway. I might check it tomorrow.

    Anyway, I experimented a bit, trying to get away with not having to reduce the volume much, and without reducing the compression if possible. After a few trial and error runs, the following seemed to do the trick.
    First the matrix mixer in the DSP chain with my usual settings and "normalise" checked.
    Second the Amplify DSP set to -12 dB.
    Third the WinAmp Bridge with my usual Rock Steady settings, with the exception of lowering the "Amplify up to" setting to -6dB.

    AmplifyTo -6
    FullAmpTo 30
    MaxAmpDB 15
    AmpGainTime 10
    SmartLimit yes
    JointStereo yes
    WindowWidth 75
    RelPos 100
    CalcRmsPer 32

    The end result for the stereo output was a volume of -25.04 LUFS and a Track Peak of -00.61 dB for the DST source, and for the AC3 it was -23.12 LUFS and a Track Peak of -00.87 dB.
    They both only just came in under maximum for the peak values, but I suspect the two sources I was playing with would be pretty close to worse case scenario, so I think those settings would be fine most of the time. At worst, the volume should only need to be reduced by a few dB more.

    I've just downloaded your sample but it's pretty late here now so I'll listen tomorrow and report back.

    Cheers.
    Quote Quote  
  10. Originally Posted by hello_hello View Post
    circulationds,

    I ran some more tests myself, starting with some 5.1ch audio from a TV show, and found downmixing followed by a -6dB gain reduction, followed by my usual RockSteady settings worked quite well, resulting in peaks of around -2 dB to -3 dB. When I tried some loud audio from a couple of movies it was a different story though. One was 5.1ch DTS and the other 5.1ch AC3. I scanned them just to see, and the DTS audio had a volume of -17.62 LUFS with a Track Peak of -00.17 dB. For the AC3 audio the volume was -13.90 LUFS with a True peak of +07.03 dB. Chances are the AC3 contained info to reduce the volume on playback, but I have the AC3 decoder configured to ignore it. That's my theory anyway. I might check it tomorrow.

    Anyway, I experimented a bit, trying to get away with not having to reduce the volume much, and without reducing the compression if possible. After a few trial and error runs, the following seemed to do the trick.
    First the matrix mixer in the DSP chain with my usual settings and "normalise" checked.
    Second the Amplify DSP set to -12 dB.
    Third the WinAmp Bridge with my usual Rock Steady settings, with the exception of lowering the "Amplify up to" setting to -6dB.

    AmplifyTo -6
    FullAmpTo 30
    MaxAmpDB 15
    AmpGainTime 10
    SmartLimit yes
    JointStereo yes
    WindowWidth 75
    RelPos 100
    CalcRmsPer 32

    The end result for the stereo output was a volume of -25.04 LUFS and a Track Peak of -00.61 dB for the DST source, and for the AC3 it was -23.12 LUFS and a Track Peak of -00.87 dB.
    They both only just came in under maximum for the peak values, but I suspect the two sources I was playing with would be pretty close to worse case scenario, so I think those settings would be fine most of the time. At worst, the volume should only need to be reduced by a few dB more.

    I've just downloaded your sample but it's pretty late here now so I'll listen tomorrow and report back.

    Cheers.

    No worries, I have checked some of my converts and they seem to be fine with the new settings I posted, checked that 10 Cloverfield compared to the other one and it's definitely better than the previous one, I know you don't understand why I'm doing it in a two-step process but I have to make sure the files are ok as I want to delete the originals but keep the extracted original audio just in case after I have a satisfying result with the converted movies as I don't see the point in keeping the originals if I get the converts to a satisfactory result. I did try your method with the downmix to 5.1 then use the amplify and then rocksteady in the Dsp line and I was getting the same results, but it's just me I want to make sure, but either way I'm enjoying using foobar anyway so it's not a big deal to me if I redo the converts i just like to perfect the work i do on them, bit of OCD lol
    Last edited by circulationds; 22nd Nov 2018 at 08:40.
    Quote Quote  
  11. Originally Posted by circulationds View Post
    Originally Posted by hello_hello View Post
    circulationds,

    I ran some more tests myself, starting with some 5.1ch audio from a TV show, and found downmixing followed by a -6dB gain reduction, followed by my usual RockSteady settings worked quite well, resulting in peaks of around -2 dB to -3 dB. When I tried some loud audio from a couple of movies it was a different story though. One was 5.1ch DTS and the other 5.1ch AC3. I scanned them just to see, and the DTS audio had a volume of -17.62 LUFS with a Track Peak of -00.17 dB. For the AC3 audio the volume was -13.90 LUFS with a True peak of +07.03 dB. Chances are the AC3 contained info to reduce the volume on playback, but I have the AC3 decoder configured to ignore it. That's my theory anyway. I might check it tomorrow.

    Anyway, I experimented a bit, trying to get away with not having to reduce the volume much, and without reducing the compression if possible. After a few trial and error runs, the following seemed to do the trick.
    First the matrix mixer in the DSP chain with my usual settings and "normalise" checked.
    Second the Amplify DSP set to -12 dB.
    Third the WinAmp Bridge with my usual Rock Steady settings, with the exception of lowering the "Amplify up to" setting to -6dB.

    AmplifyTo -6
    FullAmpTo 30
    MaxAmpDB 15
    AmpGainTime 10
    SmartLimit yes
    JointStereo yes
    WindowWidth 75
    RelPos 100
    CalcRmsPer 32

    The end result for the stereo output was a volume of -25.04 LUFS and a Track Peak of -00.61 dB for the DST source, and for the AC3 it was -23.12 LUFS and a Track Peak of -00.87 dB.
    They both only just came in under maximum for the peak values, but I suspect the two sources I was playing with would be pretty close to worse case scenario, so I think those settings would be fine most of the time. At worst, the volume should only need to be reduced by a few dB more.

    I've just downloaded your sample but it's pretty late here now so I'll listen tomorrow and report back.

    Cheers.

    No worries, I have checked some of my converts and they seem to be fine with the new settings I posted, checked that 10 Cloverfield compared to the other one and it's definitely better than the previous one, I know you don't understand why I'm doing it in a two-step process but I have to make sure the files are ok as I want to delete the originals but keep the extracted original audio just in case after I have a satisfying result with the converted movies as I don't see the point in keeping the originals if I get the converts to a satisfactory result. I did try your method with the downmix to 5.1 then use the amplify and then rocksteady in the Dsp line and I was getting the same results, but it's just me I want to make sure, but either way I'm enjoying using foobar anyway so it's not a big deal to me if I redo the converts I just like to perfect the work I do on them, bit of OCD lol

    Just an update, it has come to my attention, that some movie files have already been tampered with before I have tried to normalize them, I watched my converted scary movie 5 and its track volume is -24.37 LUFS and it's peak is 0.385749 well in range as you did mention a 0.5 maximum peak is ok but in certain scenes there was a big massive distortion issue in the movie. I scanned the original file, it is coming up at a "wapping" -10.82 LUFS and a peak value of 1.835156 so now I know that I'm not making any mistakes it has been the original files that are too high and there's nothing I can do to fix them apart from getting replacement originals. I have opened the files with mktoolnix as I thought maybe there were tags in them and could remove them but nothing, least I know now that when I was trying your settings with the rocksteady plugin, no matter how much I adjusted them, it wouldn't fix the issue as the originals have already been clipped, so there for ruined by whoever done the editing on them. Going to try audacity as I know that has a clipping fix on it, worth a try anyway.


    Just fixed the scary movie 5 file with clip fix in audacity, so that's good to know I can fix original files are clipping, there is only a few.
    Last edited by circulationds; 23rd Nov 2018 at 11:20.
    Quote Quote  
  12. I remember saying I tend not to fuss too much over peaks just a tad above maximum. +0.5 dB would probably be the most I'd live with. Maybe a tad more if I was in a good mood at the time.
    I think that's roughly 1.06000 as a percentage. Is that what you meant by a maximum peak of 0.5?

    The file you're scanning with peaks above maximum..... don't forget some lossy formats do store peaks above 0dB without actually clipping them. Are you sure the peaks are physically clipped rather than being clipped on playback? Do they looked clipped in Audacity if you lower the volume? Audacity will probably import the audio as 32 bit float so I don't think the peaks would be clipped while it's being imported as such, although I'm not 100% what Audacity does.
    Quote Quote  
  13. Originally Posted by hello_hello View Post
    I remember saying I tend not to fuss too much over peaks just a tad above maximum. +0.5 dB would probably be the most I'd live with. Maybe a tad more if I was in a good mood at the time.
    I think that's roughly 1.06000 as a percentage. Is that what you meant by a maximum peak of 0.5?

    The file you're scanning with peaks above maximum..... don't forget some lossy formats do store peaks above 0dB without actually clipping them. Are you sure the peaks are physically clipped rather than being clipped on playback? Do they look clipped in Audacity if you lower the volume? Audacity will probably import the audio as 32 bit float so I don't think the peaks would be clipped while it's being imported as such, although I'm not 100% what Audacity does.
    Yeah that's what I mean a maximum peak of 0.5, I remember you saying that lossy can store peaks above 0DB but when i played the movie both on tv and pc the same scenes where distorted really bad, like over volume, I played my normalized audio file as well and the same was happening on tv and pc

    yeah they do look clipped in audacity, there are only about 20 files that need working, will finish them off today then all the movies are done, when i open the audio file in audacity you can tell that they are clipped as you do a scan for clipped sections and then use fix clip, that's what i did, will only use audacity for any i come across in the future as it is time-consuming and with foobar i just set my converts up before bed and let them do there thing.
    Quote Quote  
  14. Originally Posted by circulationds View Post
    Originally Posted by hello_hello View Post
    I remember saying I tend not to fuss too much over peaks just a tad above maximum. +0.5 dB would probably be the most I'd live with. Maybe a tad more if I was in a good mood at the time.
    I think that's roughly 1.06000 as a percentage. Is that what you meant by a maximum peak of 0.5?

    The file you're scanning with peaks above maximum..... don't forget some lossy formats do store peaks above 0dB without actually clipping them. Are you sure the peaks are physically clipped rather than being clipped on playback? Do they look clipped in Audacity if you lower the volume? Audacity will probably import the audio as 32 bit float so I don't think the peaks would be clipped while it's being imported as such, although I'm not 100% what Audacity does.
    Yeah that's what I mean a maximum peak of 0.5, I remember you saying that lossy can store peaks above 0DB but when i played the movie both on tv and pc the same scenes where distorted really bad, like over volume, I played my normalized audio file as well and the same was happening on tv and pc

    yeah they do look clipped in audacity, there are only about 20 files that need working, will finish them off today then all the movies are done, when i open the audio file in audacity you can tell that they are clipped as you do a scan for clipped sections and then use fix clip, that's what i did, will only use audacity for any i come across in the future as it is time-consuming and with foobar i just set my converts up before bed and let them do there thing.
    Finally, all done now I can move on to my music and just do any movie file when I get them, was a pain as I had so many to work on but now caught up and plex is upto date for my tv, big thanks hello_hello helping me use foobar and getting these video's sorted. much appreciated
    Image
    [Attachment 47318 - Click to enlarge]
    Quote Quote  
  15. Originally Posted by hello_hello View Post
    I remember saying I tend not to fuss too much over peaks just a tad above maximum. +0.5 dB would probably be the most I'd live with. Maybe a tad more if I was in a good mood at the time.
    I think that's roughly 1.06000 as a percentage. Is that what you meant by a maximum peak of 0.5?

    The file you're scanning with peaks above maximum..... don't forget some lossy formats do store peaks above 0dB without actually clipping them. Are you sure the peaks are physically clipped rather than being clipped on playback? Do they looked clipped in Audacity if you lower the volume? Audacity will probably import the audio as 32 bit float so I don't think the peaks would be clipped while it's being imported as such, although I'm not 100% what Audacity does.


    Hi mate,

    Thanks for sending the text display, I will have a look at them sometime today, I have been messing around with movies again as I forgot what to do and was getting clipping issues with a couple, but have found a good set up now and happy with them, maybe you can give them a spin and tell me what you think?

    Extracted 5.1 audio, scanned and updated the tags, Converted them using your matrix normalized settings, and used the following conversion qaac.exe;qaac64.exe/ --ignorelength -s --no-optimize --no-delay -V 91 -o %d - I didn't add the -N-- for normalizing as you have that already with the matrix mixer settings, when I had that enabled with the N-- i noticed that the volume kept fading in and out but stopped now as I didn't include the normalization from the encoder

    and my steps as follows for the final conversion,

    Set RG with info to -6DB

    in the DSP chain i have the following set up

    Image
    [Attachment 48692 - Click to enlarge]


    Image
    [Attachment 48693 - Click to enlarge]


    Image
    [Attachment 48694 - Click to enlarge]



    Seems that the clipping issues were caused by too much bass so i have reduced it a little using the graphic equalizer and seems to have sorted that out, I thought maybe you would like to try the new settings I have tried out.
    Quote Quote  
  16. Member
    Join Date
    Apr 2018
    Location
    Croatia
    Search Comp PM
    You are using heavy compressor.
    Quote Quote  
  17. Originally Posted by richardpl View Post
    You are using heavy compressor.

    Yeah I'm noticing that now, lol

    working on a new method, it's all fun anyway
    Last edited by circulationds; 17th Apr 2019 at 06:59.
    Quote Quote  
  18. There's a few things I'd change, but first.... the matrix mixer normalising and the QAAC normalising are two different things. For the matrix mixer, it reduces the volume according to how you're downmixing and the number of channels being combined. It reduces the volume using a set formula so there shouldn't be any clipping after the channels are combined. QAAC's normalising does peak normalising. It runs two passes, finds the loudest peak on the first pass, then adjusts the over-all volume when encoding to peak normalise (peaks at 0dB). I never use QAAC's normalising when "compressing". Maybe I once had a conversion preset saved..... but if I did, it's long gone.

    Your ReplayGain setting is telling fb2k to reduce the volume to 6dB below the standard ReplayGain volume when there's ReplayGain data (83dB). That's a bit pointless when compressing, because the volume will change again anyway. If there's no ReplayGain data, it does nothing. If you just want to reduce the volume by a fixed amount, I think the first screenshot below is correct. It'll reduce the volume by 6dB while ignoring any ReplayGain data, but I'd disable ReplayGain and just use the Amplify DSP for that.

    Edit: I kind of remember you were keen to adjust all audio to the same volume before compressing? Is that why you adjust to ReplayGain's 83dB first?
    If so, the second screenshot below will adjust to 83dB when there's ReplayGain info, or reduce the volume by 6dB when there isn't. If you're following ReplayGain with the Amplify DSP set to -6dB, that'd make it a 12dB reduction in total. If you want it to be 12dB you can just set it to -12dB in the ReplayGain configuration and not use Amplify at all. I think you might be doubling up unintentionally there.

    I don't know what's causing the crackling, unless the audio was clipping pretty hard when it was converted to 16 bit, or because you disabled smart limiting, or maybe because you adjusted the RMS options. I know they're labelled RMS, but if memory serves me correctly, they have dual functions, the secondary function being "crackles". If you want to test that theory, the defaults are 75, 100 & 32. And maybe only convert one file at a time. The WinAmp DSP Bridge sometimes crashes fb2k if I don't.

    With "Full Amp Up To" set to 50%, you've told RockSteady it can apply your specified maximum amplification until the volume reaches -6dB. I have it set to 30%, which is roughly -10dB. It'll increase the chances of volume pumping at 50%, but it should compress a bit more. After the "Full Amp Up To" volume, RockSteady scales back the amplification until it's zero at the "Amplify Up To" volume. It probably wouldn't hurt to set "Amplify Up To" to -3dB to give RockSteady a bit of wiggle-room.

    You can leave the Matrix Mixer in the DSP chain (and configured for downmixing to stereo) when converting stereo audio too. If there's only two channels in the audio stream, and left and right aren't "matrixed" (they just stay in left and right), Matrix Mixer will simply pass them through untouched. Not that it matters, but I thought I'd mention it.

    So if all you want to do is reduce the volume by 6dB before compressing, you can use

    - Amplify (-6dB)
    - Matrix Mixer (normalising enabled and downmixing to stereo)
    - WinAmp Bridge

    in the DSP chain, and leave ReplayGain disabled. Otherwise you might as well set the ReplayGain option for the pre-compression volume you prefer and then you won't need Amplify.
    Image Attached Thumbnails Click image for larger version

Name:	fb2k rpg1.jpg
Views:	105
Size:	20.1 KB
ID:	48700  

    Click image for larger version

Name:	fb2k rpg2.jpg
Views:	86
Size:	20.1 KB
ID:	48701  

    Last edited by hello_hello; 17th Apr 2019 at 22:41.
    Quote Quote  
  19. Originally Posted by hello_hello View Post
    There's a few things I'd change, but first.... the matrix mixer normalising and the QAAC normalising are two different things. For the matrix mixer, it reduces the volume according to how you're downmixing and the number of channels being combined. It reduces the volume using a set formula so there shouldn't be any clipping after the channels are combined. QAAC's normalising does peak normalising. It runs two passes, finds the loudest peak on the first pass, then adjusts the over-all volume when encoding to peak normalise (peaks at 0dB). I never use QAAC's normalising when "compressing". Maybe I once had a conversion preset saved..... but if I did, it's long gone.

    Your ReplayGain setting is telling fb2k to reduce the volume to 6dB below the standard ReplayGain volume when there's ReplayGain data (83dB). That's a bit pointless when compressing, because the volume will change again anyway. If there's no ReplayGain data, it does nothing. If you just want to reduce the volume by a fixed amount, I think the first screenshot below is correct. It'll reduce the volume by 6dB while ignoring any ReplayGain data, but I'd disable ReplayGain and just use the Amplify DSP for that.

    Edit: I kind of remember you were keen to adjust all audio to the same volume before compressing? Is that why you adjust to ReplayGain's 83dB first?
    If so, the second screenshot below will adjust to 83dB when there's ReplayGain info, or reduce the volume by 6dB when there isn't. If you're following ReplayGain with the Amplify DSP set to -6dB, that'd make it a 12dB reduction in total. If you want it to be 12dB you can just set it to -12dB in the ReplayGain configuration and not use Amplify at all. I think you might be doubling up unintentionally there.

    I don't know what's causing the crackling, unless the audio was clipping pretty hard when it was converted to 16 bit, or because you disabled smart limiting, or maybe because you adjusted the RMS options. I know they're labelled RMS, but if memory serves me correctly, they have dual functions, the secondary function being "crackles". If you want to test that theory, the defaults are 75, 100 & 32. And maybe only convert one file at a time. The WinAmp DSP Bridge sometimes crashes fb2k if I don't.

    With "Full Amp Up To" set to 50%, you've told RockSteady it can apply your specified maximum amplification until the volume reaches -6dB. I have it set to 30%, which is roughly -10dB. It'll increase the chances of volume pumping at 50%, but it should compress a bit more. After the "Full Amp Up To" volume, RockSteady scales back the amplification until it's zero at the "Amplify Up To" volume. It probably wouldn't hurt to set "Amplify Up To" to -3dB to give RockSteady a bit of wiggle-room.

    You can leave the Matrix Mixer in the DSP chain (and configured for downmixing to stereo) when converting stereo audio too. If there's only two channels in the audio stream, and left and right aren't "matrixed" (they just stay in left and right), Matrix Mixer will simply pass them through untouched. Not that it matters, but I thought I'd mention it.

    So if all you want to do is reduce the volume by 6dB before compressing, you can use

    - Amplify (-6dB)
    - Matrix Mixer (normalising enabled and downmixing to stereo)
    - WinAmp Bridge

    in the DSP chain, and leave ReplayGain disabled. Otherwise you might as well set the ReplayGain option for the pre-compression volume you prefer and then you won't need Amplify.

    I have tried my old settings and what you have recommended, are AC3 files a problem, as I'm noticing I'm getting clipping on these sorts of audio files, would it be better to get dts files instead? I've just tried a dts and same thing so that was pointless, I have added the extracted audio can you please have a look at it for me and tell me what settings you used to stop it clipping, I have gone through all your steps and i honestly don't know what i am doing wrong?hello_hello.rar it clips really bad at 1:40:43
    Last edited by circulationds; 18th Apr 2019 at 16:03.
    Quote Quote  
  20. When you said "clipping" were you referring to something you can hear? I assumed you were, but are you referring to a ReplayGain scan of the compressed/encoded audio reporting peaks above 0dB? If it's the latter, that's what lossy audio does. The only way to avoid it is to reduce the volume by 3dB before it's encoded. At least when using RockSteady to compress, you can follow the compression with the Amplify DSP to reduce the volume.

    I took a FLAC file with a peak of 0dB and ran it through RockSteady while encoding to different formats. Then I did the same again, but I followed RockSteady with the Amplify DSP to reduce the volume by 3dB. Here's the scan results.
    Image Attached Images    
    Last edited by hello_hello; 18th Apr 2019 at 16:34.
    Quote Quote  
  21. i will upload a few video's sorry for the messing about
    Last edited by circulationds; 18th Apr 2019 at 19:59.
    Quote Quote  
  22. Last edited by circulationds; 18th Apr 2019 at 20:00.
    Quote Quote  
  23. tried to upload video's and remove replies
    Last edited by circulationds; 18th Apr 2019 at 20:01.
    Quote Quote  
  24. god help lol
    Last edited by circulationds; 18th Apr 2019 at 20:06.
    Quote Quote  
  25. Originally Posted by circulationds View Post
    i will upload a few video's sorry for the messing about
    here you go second video is without mouse pointer, i downloaded a free desktop recorder, hope you can see what I'm doing now?? as you can probably hear, the volume is rising and lowering all the time and clipping no matter what i do?
    Image Attached Files
    Last edited by circulationds; 18th Apr 2019 at 20:47.
    Quote Quote  
  26. If you're referring to the thin lines bouncing up and down above the main meter bars, they don't indicate clipping. The peak meter responds quite quickly, so it's hard to tell how loud the peaks actually are. Those thin lines hold and display the peak level for a little longer to give you time to see it. If they go over 0dB, that's an indication of clipping, but I'm not sure I saw that happen.

    RockSteady is only 16 bit, so like a 16 bit wave file it can't output values above 0dB. If it's the last DSP in the chain, 0dB is the limit. Take a look at the screenshots below. I used the ReplayGain section to give the audio a 20dB boost before downmixing. The matrix mixer's normalising expects audio at a standard volume, so if it's boosted it won't be reduced enough to prevent clipping. You can see what the peak meter is displaying with and without RockSteady. With RockSteady in the chain the peak did sound a little distorted, but there wasn't any horrible "digital" clipping.

    I heard some obvious "volume pumping" in the Creed audio towards the end. The commentator's speech was going up and down. That's because it's barely above the background music, and it's what's causing RockSteady to adjust the volume. The speech is going along for the ride.
    That can be a side effect of compression and is one reason why I never include the LFE channel when downmixing. If you'd care to upload a sample of the audio from that point I can have a play, but using my RockSteady settings that sort of volume pumping can happen on occasion, although I find it very untypical. The only way to fix it would be to slow the compressor's response time or reduce the compression. The former will make the increasing and decreasing of the volume more noticeable. A bit like turning the volume down when it gets loud and then slowly turning it back up again. That's how the normalising built into MPC-HC and ffdshow works, which is why I don't like it. The compression built into most hardware players (night mode, or something similar) is usually quite slow so there might be less of the "Creed effect", but you can often hear low volume background sounds slowly increase and decrease in volume between loud foreground sounds.
    You could try giving the centre channel a bit of a boost in the matrix mixer. Put it in the left and right channels at 1 instead of 0.707, but I don't think it'll make much difference.

    The ffdshow screenshots should help show what the matrix mixer's normalising does as I'm pretty sure it works the same way (although the numbers are a bit hard to read).

    #1 is an un-normalised mix (ffdshow doesn't reduce the surround channels by default and it includes the LFE channel).
    #2 is with normalising enabled.
    #3 is how I usually run it, if it was un-normalised
    #4 is how I usually run it, normalised.

    As you can see, 2 and 4 are different because the normalising matrix is adjusted for the missing LFE channel and the boost to the centre channel. The centre channel boost is mostly just placebo though, because after the audio is compressed it doesn't make much difference.
    Image Attached Thumbnails Click image for larger version

Name:	Clipboard01.gif
Views:	63
Size:	37.0 KB
ID:	48726  

    Click image for larger version

Name:	Clipboard02.gif
Views:	65
Size:	39.7 KB
ID:	48727  

    Click image for larger version

Name:	1.gif
Views:	52
Size:	22.2 KB
ID:	48728  

    Click image for larger version

Name:	2.gif
Views:	64
Size:	22.3 KB
ID:	48729  

    Click image for larger version

Name:	3.gif
Views:	52
Size:	21.9 KB
ID:	48730  

    Click image for larger version

Name:	4.gif
Views:	93
Size:	22.3 KB
ID:	48731  

    Quote Quote  
  27. Actually, screenshots 3 and 4 aren't exactly how I normally run it, as I reduce the surround channels to 70% (-3dB), but I forgot to do it when taking those screenshots. As you can see, the remaining channels get a slight boost as a result (compared to #4 above).
    Image Attached Thumbnails Click image for larger version

Name:	5.gif
Views:	52
Size:	22.3 KB
ID:	48732  

    Quote Quote  
  28. Originally Posted by hello_hello View Post
    Actually, screenshots 3 and 4 aren't exactly how I normally run it, as I reduce the surround channels to 70% (-3dB), but I forgot to do it when taking those screenshots. As you can see, the remaining channels get a slight boost as a result (compared to #4 above).
    I have done this setting, had to do a few tweaks with Winamp DSP and I edited the matrix settings to the settings you have in picture 3. when I was showing you the clip of creed it wasn't to show that it was clipping, I know that if audio goes over 0DB it will clip but that creed audio wasn't going over 0DB and it was clipping, anyway here is a video of my new settings, not quite what I wanted but will see how they do on my tv

    Not sure how i can send you that specific part of the clip with all channels, above i have added the whole extracted 6 channel audio
    Image Attached Files
    Last edited by circulationds; 19th Apr 2019 at 08:50.
    Quote Quote  
  29. Originally Posted by circulationds View Post
    Originally Posted by hello_hello View Post
    Actually, screenshots 3 and 4 aren't exactly how I normally run it, as I reduce the surround channels to 70% (-3dB), but I forgot to do it when taking those screenshots. As you can see, the remaining channels get a slight boost as a result (compared to #4 above).
    I have done this setting, had to do a few tweaks with Winamp DSP and I edited the matrix settings to the settings you have in picture 3. when I was showing you the clip of creed it wasn't to show that it was clipping, I know that if audio goes over 0DB it will clip but that creed audio wasn't going over 0DB and it was clipping, anyway here is a video of my new settings, not quite what I wanted but will see how they do on my tv

    Not sure how i can send you that specific part of the clip with all channels, above i have added the whole extracted 6 channel audio
    I feel such an idiot, all this time your were telling me that the rocksteady DSP only accepts 16Bit, i have just noticed it in the conversion stage lol just converted the file and it sounded better, there was no clipping at all Image
    [Attachment 48737 - Click to enlarge]


    now i'm getting confused with rock steady, i created a wav file that was 16bit and when i play it and have the rocksteady dsp on, adjustments don't do anything?
    Last edited by circulationds; 19th Apr 2019 at 11:57.
    Quote Quote  
  30. I thought I heard some "crackles" but I wasn't sure if it was part of the audio or not.

    Changing that setting shouldn't make a difference, although as a rule, bit depth shouldn't be an option for lossy encoders. If they're configured as lossy, it won't be. fb2k will automatically output the correct bit depth according to what's set in the encoder configuration. That bit depth setting in your screenshot wouldn't have an effect until after the DSPs.

    For lossless encoding, the bit depth setting allows you to convert a 24 bit lossless source to 16 bit etc, and when it's set to "auto" the output bit depth is the same as the input bit depth. For lossy sources, the output bit depth is 16 bit when it's set to "auto" and the output is lossless.
    So if the encoder is configured as lossless, "auto" and 16 bit should do the same thing.

    When the encoder is set to lossy, the bit depth setting shouldn't be shown when you select the encoder. The bit depth is only controlled by the bit depth specified in the encoder configuration. I checked, and even though RockSteady is 16 bit, if a lossy encoder is configured for 32 bit, the bit depth of the audio sent to the encoder is 32. Chances are, the WinAmp DSP bridge outputs 32 bit even though you're using a 16 bit WinAmp DSP. It'd make sense for it to do so, as that'd keep the rest of the DSP chain 32 bit.

    You can see what's being sent to the encoder by checking the Console log. This conversion included RockSteady. I still don't know how that setting could have caused or prevented those clipping sounds, given the peaks didn't look like they exceeded 0dB.

    Converting: "D:\Matchbox Twenty - 3 AM.mp3"
    Destination: "E:\Matchbox Twenty - 3 AM.m4a"
    CLI encoder: qaac.exe;qaac64.exe
    Destination file: E:\Matchbox Twenty - 3 AM.m4a
    Encoder stream format: 44100Hz / 2ch / 32bps
    Command line: "C:\Program Files\foobar2000\encoders\QAAC\qaac.exe" --ignorelength -s --no-optimize --no-delay -V 91 -o "Matchbox Twenty - 3 AM.m4a" -

    By the way, under Advanced/Tools/Converter, there's an option to get rid of the transcoding warning popups. Which is also odd now I think about it. If the encoder configuration is set to lossless, I don't think there is a transcoder warning, yet I'm sure I saw one in your video.

    PS I haven't listened to the video yet. I thought it was the audio sample at first, but I'm out of time for now anyway. I will look at it later.

    I still think the crackles might be related to this:

    Position in RMS window: Specifies where the current sample (the one which gets amplified now) is positioned within the RMS window. A value of 0 means the beginning and 100 means the end of the window. A value lower than 100 (recommended about 75) will make amplification more responsible to sudden changes in the sound level. If you use values highter than 75, turning Smart limiting on would be a good idea. And now, some BAD NEWS: WinAmp's plugin mechanism doesn't allow RockSteady to know when playing is stopped or started or when seeking occurs. Also, if the Position in RMS window is lower than 100, RockSteady has to delay the output sound a little. The combination of these two facts means that you will hear audible clicks if you seek inside a song, or if you stop and then start playing again. So, the recommended value is 100.
    Image Attached Thumbnails Click image for larger version

Name:	Clipboard03.gif
Views:	86
Size:	54.3 KB
ID:	48740  

    Last edited by hello_hello; 19th Apr 2019 at 13:24.
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!