DTS/AC3 to Apple AAC?

[kkiller23]

I'm trying to figure out how to convert DTS/AC3 files to Apples AAC? I've been using foobar and nero's AAC, but people keep telling me i'm wasting my time and to use apples instead. Is there any updated apple encoders that work with foobar? or just any at all for windows that i don't need to install itunes/quicktime?

[poisondeathray]

qaac can work with foobar

[kkiller23]

qaac can work with foobar

Are there any tutorials on how to use qaac with foobar? I'm trying to convert ac3 5.1 to AAC-LC 2.0 128kb/s

So what do i actually need installed to use qaac?

[poisondeathray]

I don't use it with foobar, but there is a step by step pdf with pictures you can download that shows you how to use it with foobar

search for foobar2000 and qaac , it should be the 1st hit

[kkiller23]

Hm, that guide seems to be for lossless. I can't figure out how to encode aac using qaac lol


Does anyone know what i have to put in to get aac-lc 128kb 2.0?

[poisondeathray]

I normally just use the command line

If you want a GUI for qaac encoding AAC, you can use megui

I don't use foobar much so I don't know how to configure it.

[kkiller23]

I normally just use the command line

If you want a GUI for qaac encoding AAC, you can use megui

I don't use foobar much so I don't know how to configure it.

How do you use qaac with megui?

Do i have to rename it neroAacEnc.exe or something haha

[poisondeathray]

in the audio input, load your audio

in the encoder settings, select " qaac scratchpad" then push "config" button to configure the options . Once done, push queue. Then go to the queue tab and push start to begin

[kkiller23]

in the audio input, load your audio

in the encoder settings, select " qaac scratchpad" then push "config" button to configure the options . Once done, push queue. Then go to the queue tab and push start to begin

Ah, i had to go into options first and tick the qaac box and then restart megui for it to show up. Thanks a ton!

[jagabo]

What about ffmpeg? I used this in a batch file:

ffmpeg -i %1 -ac 2 "%~n1.aac

to convert .DTS to .AAC. It appears to have worked. I don't know what exactly it did to downmix 6 channels to 2 (all are present in the output though). It also worked for AC3, MP2, WAV, even an MPG file with MP2 audio.

[kkiller23]

hm, the audio is really low going from dts to qaac in megui. Should i normalize peaks to 100 as well? Also what decoder should i use in megui?

Also @ jagabo. I didn't want to use ffmpeg because i wanted apples codec because everyone said they have the best quailty. I've test a few movies using qaac and it does sound great for the low bitrate. It's just really quite. I normalized peaks to 100 a few time and that made it louder. I'm just not sure if i should be doing that or not. I'm also not sure what decoder to use in megui. I was using NicAudio, but FFAudioSource sounds a little better to my ears when playing the file back in vlc.

[jagabo]

I've test a few movies using qaac and it does sound great for the low bitrate. It's just really quite. I normalized peaks to 100 a few time and that made it louder. I'm just not sure if i should be doing that or not.

If ffmpeg's aac encoder is anything to judge by, aac doesn't change the volume level. If some other program is giving you louder audio it's because it's automatically normalizing to a higher volume level.

[kkiller23]

I've test a few movies using qaac and it does sound great for the low bitrate. It's just really quite. I normalized peaks to 100 a few time and that made it louder. I'm just not sure if i should be doing that or not.

If ffmpeg's qaac encoder is anything to judge by, qaac doesn't change the volume level. If some other program is giving you louder audio it's because it's automatically normalizing to a higher volume level.

I was talking about in megui. I clicked the normalize peaks thing in megui and it made it louder in playback with vlc. Is it okay to normalize peaks to 100 in megui for dts/ac3? It won't distort the sound or anything, right? Also, do you know what the decoder thing is in megui and what the best one to use is?

[hello_hello]

The script MeGUI uses for downmixing uses some sort of "normalising matrix" which effectively reduces the volume most of the time, so using the other normalising too (increasing peaks to 0dB) is probably a good idea.

I've never quite understood the reason for it.... I'd guess it's to ensure there's no possibility of clipping when the channels are combined, but I think.... for those who use ffdshow's audio decoder... if you enable ffdshow's mixer filter and select stereo output, then check the "normalize matrix" box, you'll get the same or a similar drop in level as when downmixing with MeGUI.

Generally for decoding you'd pick the first decoder in the list. MeGUI will try it and if it fails it'll move on to the next one and so on. You'd generally only start with a decoder other than the first one in the list if you know you need to for some reason... maybe DirectShow decoding as an example.

For the record, foobar's multichannel to stereo DDSP just downmixes without applying a further gain reduction. If you want to ensure the downmixed file doesn't contain peaks above 0db, according to Dolby (for AC3) a further gain reduction of 7.5db should be applied. I've checked a lot of encoded audio after downmixing and I think a 6db reduction will do the job 99.99999% of the time. I recall another poster here saying he uses 4.5db. That's one reason I still use foobar2000 for downmixing if I do it. I dislike traditional "normalizing" because the peaks in every audio stream will be different, so some will be "normalized" more than others and you can end up with different levels... which doesn't matter for movies but it drives me a little nutty when it comes to re-encoding audio from episodes of TV shows. So.... foobar2000... downmix to stereo.... and throw in a gain reduction of 6db while converting. The peaks won't all be at 0dB, but the volume of each encode should be the same relative to the original each time. Well that's my method anyway.

From what I've read, qaac is better then nero at silly low bitrates, but once the bitrate is sufficient I doubt think you'd tell them apart. Ever read an encoder comparison which compared encoders at a bitrate you're likely to use? Unless you do use very low bitrates....

I just use the default Q.50 for Nero. Can qaac use a similar quality method for encoding? I've never used it.

[kkiller23]

I was using q.45 with Nero, but thats the problem. I can't figure out how to encode qaac with foobar so that is why i was using MeGUI. If i apply the normalizer it seems to sound fine.

Here are the settings i'm using with qaac in megui



do they look okay or am i doing something wrong? I'm not really an audiophile. So i don't know a ton about encoding audio. I usally just used the default settings when encoding with foobar. The only thing i did was apply the downmix to 2.0. I've never really messed with dB or anything like that. I figured if i'm going to start using the better encoder i might as well do it right.

[hello_hello]

That looks fine. The "too quite" downmix is due to the way MeGUI downmixes. It creates an AVIsynth script to do the downmixing. If you're keen you'll find it in the log file but there's nothing you can do to change it and you probably shouldn't anyway. Selecting the "normalize peaks" option will effectively get MeGUI to increase the volume again (peaks at maximum).

I'm pretty sure the volume will only drop when downmixing. If you convert multichannel to multichannel or stereo to stereo you should get the same volume as the original... unless of course you also check the "normalize peaks" option, in which case it'll be errrrr.... normalised too.

[kkiller23]

That looks fine. The "too quite" downmix is due to the way MeGUI downmixes. It creates an AVIsynth script to do the downmixing. If you're keen you'll find it in the log file but there's nothing you can do to change it and you probably shouldn't anyway. Selecting the "normalize peaks" option will effectively get MeGUI to increase the volume again (peaks at maximum).

I'm pretty sure the volume will only drop when downmixing. If you convert multichannel to multichannel or stereo to stereo you should get the same volume as the original... unless of course you also check the "normalize peaks" option, in which case it'll be errrrr.... normalised too.

Looking at the MeGUI logs i was able to get the command line and using "--ignorelength --threading -V 75 - -o %d" in foobar works. It seems to sound pretty good with just applying the "Downmix channels to stereo" Active DSP.

This is the output

General
Format : MPEG-4
Format profile : Apple audio with iTunes info
Codec ID : M4A
File size : 78.2 MiB
Duration : 1h 34mn
Overall bit rate mode : Variable
Overall bit rate : 115 Kbps
Encoded date : UTC 2013-08-10 19:51:01
Tagged date : UTC 2013-08-10 19:52:42
Writing application : qaac 2.19, CoreAudioToolbox 7.9.8.3, AAC-LC Encoder, TVBR q73, Quality 96

Audio
ID : 1
Format : AAC
Format/Info : Advanced Audio Codec
Format profile : LC
Codec ID : 40
Duration : 1h 34mn
Bit rate mode : Variable
Bit rate : 114 Kbps
Maximum bit rate : 171 Kbps
Channel(s) : 2 channels
Channel positions : Front: L R
Sampling rate : 48.0 KHz
Compression mode : Lossy
Stream size : 77.1 MiB (99%)
Encoded date : UTC 2013-08-10 19:51:01
Tagged date : UTC 2013-08-10 19:52:42

It sounds pretty good, but the voice are pretty low still. So it looks like the best bet is just using megui with the normalizer since that makes it the loudest. Since i'm no pro with foobar and always just used basic settings in it.

[hello_hello]

I was using q.45 with Nero, but thats the problem. I can't figure out how to encode qaac with foobar so that is why i was using MeGUI.

Open foobar2000's converter setup and when you add a new encoder, select "Custom" from the encoder drop down list, then navigate to wherever qaac is on your PC. There may already be an encoder listed in the encoder file section as I think when you select Custom foobar2000 adds the last used encoder.... just replace it by navigation to qaac.

Change the extension to m4a or whatever you use and replace anything in the command line area with this:

--tvbr 70 - -o %d

The number in the above (70) specifies the quality. I assume it works the same way (same values) as MeGUI's qaac encoder setup. Change it to whatever quality setting/value you want to use. The above does quality based VBR encoding. If you want to also do CBR encoding for example, you'd need to create a different encoder preset with a different command line.

The format is lossy and I'd guess 32 is the highest supported bit mode.

Change the display stuff at the bottom to whatever you want displayed for the converter preset and click on OK.

If you want to, you can of course repeat the process to save multiple encoder configurations which use different quality settings. Maybe include the quality in the display section when you setup each encoder preset so you can tell them apart if you do.

Don't hold me to any of the above. It came via Google and a consensus of the correct command line to use (which made sense to me). I don't have qaac so i can't test it, but I'm fairly confident it's correct.

If you give it a go, please report back to let us know if it worked.

[hello_hello]

You beat me to it....

-ignorelength wasn't part of my command line consensus but it makes sense. I've used it in the NerosAAC command line with foobar2000 and it seems to do the same thing for both encoders.

--threading mightn't be needed. I'd have to think about that one. If you convert more than one file at a time foobar2000 uses a thread for each job.... I don't know whether the encoder also using multithreading is a clever idea or not. Or for that matter whether the "do not convert using multiple threads" in foobar's encoder setup refers to using multiple threads for a single encoding job.

The full list of command line options is here if you need it: https://github.com/nu774/qaac/wiki/Command-Line-Options

[kkiller23]

Well i tried "--ignorelength --threading -N -V 75 - -o %d" adding the -N "Normalize" and it still seems pretty low compared to the MeGUI out put. I might just stick with MeGUI for encoding qaac.

[kotuwa]

guys, how to swap Left and Right Channels when encoding AC-3 to AAC with QAAC?
what are the switches?
Can it be done with GUIs like MeGUI or FOOBAR?
Is there any way or any tool that can do this in AAC tracks without reencoding?

[poisondeathray]

guys, how to swap Left and Right Channels when encoding AC-3 to AAC with QAAC?
what are the switches?

See the help

--chanmap <n1,n2...> Re-arrange channels to the specified order.
For N-ch input, you take numbers 1,2..N, and
arrange them with comma-seperated, to the order
you want.
For example, "--chanmap 2,1" swaps L and R.

[hello_hello]

Can it be done with GUIs like MeGUI or FOOBAR?

You can use whatever command line you like when setting up a foobar2000 encoder preset, but you can't change the MeGUI command line.

There's a link to the list of command line options for QAAC two posts above yours.

[kotuwa]

Thanks...
What is the order of channels in 5.1?
is it L C R, Side: L R, LFE
1 2 3 4 5 6 respectively? or?

[hello_hello]

According to the info at the bottom of this page, different formats use different channel orders. However......

Regardless of the format, I think when 5.1ch audio is decoded, it's usually decoded using the order for 5.1ch WAV (when decoded with a PC). If it's being re-encoded, the encoder will remap the channels according to the order it requires if necessary. The upshot of which is, if you're messing with the channel order, assume it's the order for 5.1ch WAV.

FL, FR, FC, LFE, RL, RR

So to swap left and right, you'd swap channels 1 & 2.

For QAAC and 5.1ch audio you'd need to add this to the command line to swap the left and right channels: --chanmap 2,1,3,4,5,6

The order for 5.1ch WAV is the same order foobar2000's output meter uses for multi-channel audio. It's also the same order used by the Matrix Mixer I mentioned here.

[kotuwa]

thank you!

Is RL and RR swapped in --chanmap 2,1,3,4,5,6 !?... or just FL and FR?

(for me 2,1,3,4,5,6 order look like nothing swapped! :/)

[hello_hello]

thank you!

Is RL and RR swapped in --chanmap 2,1,3,4,5,6 !?... or just FL and FR?

It should be just FL and FR.

If you're using foobar2000, why not try the Matrix Mixer DSP I mentioned in one of the other threads where you asked a similar question? That way you can swap channels regardless of the encoder. You can even stick the DSP in the playback chain and swap the channels around in real time while you listen to the playback.