Editing aac files

[johnb]

There are various methods of cutting AAC-LC files without decoding/re-encoding (which is to be avoided if at all possible) but it is extremely unusual to find one that can also introduce fade ins/outs.

I have recently come across one programme that can to do just that, while (according to the website) not decoding/re-encoding: mp3TrueEdit which is available for both Windows and Mac computers. The cost is $11.99 but there is a fully functioning trial version can be used for 14 days.

mp3directcut seems to me to be more functional and somewhat easier to use, even though mp3TrueEdit has a more flashy screen. However, mp3TrueEdit is the only programme I have seen that can do fades and edit without any loss of audio quality.

The fade in/out is set by default at 12dB, which seems rather odd. The fade in/out can be repeated to increase the amount faded but it is much easier to change to default to something like, say, 60bB in the "Options".

It can also save the aac file as m4a.

(I should mention that mp3TrueEdit has one infuriating quirk when you first run the software - it shows a sequence of text bubbles explaining what the various controls do, even though they are fairly self evident. This seemed to go on for ever, with me getting increasingly annoyed.)

Anyway, I thought I would post this in case someone might be interested.

PS It also seems to be able to edit m4a files, as long as they contain AAC-LC.

[David-G]

Thanks John, that is very interesting. If I am not interested in the fade capability, is there other such software I ought to consider?

[johnb]

David, if you don't need the fade in/out the mp3directcut will do the job and it is much easier to use once you get used to the interface. It doesn't save to m4a though.

(I usually use mp4box to put AAC files into a m4a container.)

[Dave2002]

The significant feature is not having to convert from and also back to aac. There will always be one conversion to a playable form as the track is played (usually PCM, but could be DSD). If file size and formats are less of a problem, than audacity should be able to do the job (i.e. crossfades), and the resulting file could be FLAC or similar. The (one way) m4a to PCM conversion in audacity would be done by the ffmpeg plug in. http://audacity.sourceforge.net/help...ma-proprietary

There would be a (hopefully small) quality loss if the resulting tracks were saved in any lossy compressed format, including aac.

Any significant differences using the one way approach would then be down to how well ffmpeg decodes the aac data file.

The suggested aac editing tool may be the way to go if aac source and target files are a requirement.

[Bryn]

It might be worth keeping an eye open for further developments re. level adjustments, etc. in Mp3DirectCut. I don't feel a need for a fade facility, so will be sticking with the freeware mp3DirectCut.

[David-G]

Any significant differences using the one way approach would then be down to how well ffmpeg decodes the aac data file.

Does this mean that there is more than one way of decoding the file? Is there not a single algorithm to do that? How does one know if software that reads aac files uses "good" or "bad" decoding?

[Dave2002]

Does this mean that there is more than one way of decoding the file? Is there not a single algorithm to do that? How does one know if software that reads aac files uses "good" or "bad" decoding?

I don't know - I'm not even sure if the aac process has been fully published. In the case of mp3 encoding/decoding there is a broad brush specification, which does allow different encoders/decoders, and in that case the decoders were specifically designed to be simple with relatively little room for variations.

With the design of encoders/decoders the developers do have to make decisions about whether more effort goes into encoding or more into decoding. For "off line" material - recordings - it is possible and realistic to have complex encoders, which means that the overall quality can be acceptably good with simple decoders. However this is not always possible, for example with real time communications systems where any delays in encoding would present problems, so compromises have to be made. Faster processing due to several factors, including faster processors, will undoubtedly have reduced the impact of some of the problems in encoding/decoding, and we see that in digital TV.

To return also to your original question, the simple answer is "I don't know", but even if there were a single algorithm, this could include lines of code such as "if there is still 10 milliseconds left before outputting the next sample, do the following, else emit the sample now".
This would be an algorithm - i.e. a well defined rule, but the results would depend on other factors and the speed of processing. An algorithm could also be heavily parameterised, and the parameters used could give different quality results.

Others may know more specifically about aac. Re mp3 playback, some decoders have previously claimed to give audibly better results, though I believe most are now fairly comparable.

[johnb]

I've often wondered about this too.

The Squeezebox system uses a patched version faad to decode AAC streams - going by the version of faad it is very likely to be faad2 rather than faad. (The patches are to enable it to decode ALAC streams and they don't affect AAC.) So, foolishly or otherwise, I use the unpatched version of the same programme.