I recently discovered the August release of the Nero Digital Audio tools, with support for Linux.

My library is archived in FLAC format. Previously I used a perl script to batch convert all of my FLAC files to AAC using FLAC and neroAacEnc.exe (an early release) under Windows XP. I used -lc and -q 0.22 to produce files with an average bitrate of around 92-96 kbps. Sometimes I would get an error message saying "lost sync" and the encoder would bail and move on to the next file, resulting in a truncated (file size smaller) AAC output file.

Now I am able to do the whole operation under Linux using the same perl script, FLAC, and neroAacEnc (the August 6 2007 build).

First of all I noticed that -q 0.22 results in much, much smaller output files, so I experimented and found that -q 0.34 produces the same average bitrate I had before. Did Nero rescale the quality factors at some point?

Next, I have not noticed any of the "lost sync" errors. However, I have noticed that some of the songs that have been copied to my iPod abort playback partway through. When I re-do a song, the file sizes are identical. The new version of the AAC plays all the way through on the iPod. Both versions play all the way through in iTunes. Does anyone have any ideas why this might be happening?

I am using 'neroAacEnc -lc -ignorelength -q 0.34' to encode. I can upload sample files later, one that will play back on the iPod and one that aborts part way through. Does anyone want to take a look?

It would be nice to be able to turn off the progress indicator in neroAacEnc. I run the conversion as a background batch job, and it would be helpful to be able to log STDERR without ending up with a file full of "##%complete" to sift through for actual problems. In fact, it would be nice to be able to turn off the banner and just output errors.

After a bit of diagnosis, re-copying the same AAC file onto the iPod again, it seems that the aborted playback, and the digital "scritch!" errors, were introduced when copying the files onto the iPod.

Has anyone else experienced fairly frequent errors copying to their iPod? This is the first time that I have noticed it. The device in question is a 5th gen 30GB iPod (from before the backlight upgrade). I am using USB 2.0 from a Windows XP machine to do the copying.

I haven't had any issues with either the older February 2007 build (the one that I think actually produces higher quality files) or the August 2007 build. I have some AAC files from both builds and they copy fine to my iPod touch and my 5G 60GB iPod.

I am guessing that something isn't write with the encoding process. Since you are using Windows XP, wouldn't it be better to use a program like foobar2000 and the older February build of Nero AAC to encode all your FLAC files? I have done some ABX testing as well as a couple of other members here and we all agree (not that we have communicated with each other) than the February build produces superior results.

It is funny that you mention the bitrate as my experience is opposite. I find that the August build produces larger files over the February build. Most of my songs are in the general metal category and the February build produces an overall average bitrate of around 150kbps when using -q0.45 but the August build will produce files of around 190kbps.

Like I said, I discovered that the problem is apparently with copying files to the iPod, not with the encoder.

And like I also said, I am using Linux now, not XP. I prefer to use Linux because I have a Linux machine that is on all the time, and I don't want to have to leave the XP box on all the time. It has a bad time dealing with unusual characters, anyway. If the February build included Linux binaries, I would give it a shot.

As for the bitrate, I am not comparing to the February build. I am comparing to the May 26 2006 build.

During previous tuning of the encoder we were forgetting to update mappings for q values. That was finally done in August version and now q mappings correspond to the table for Nero AAC Recommended Settings : http://www.hydrogenaudio.org/forums/index....showtopic=44310
In the future we will conform to these mappings.


We are trying to track down where the problem comes from and we would be grateful to anybody who can contribute by providing us the examples where February version clearly outperforms August version. It is not possible to simply use psychoacoustic parameters from February because many bugs that affect quality have been fixed since then, and psychoacoustic tunings for February version compensated those bugs.