I am trying to record LPs at 24/96 and then export as AAC in a way that will play on both iTunes and my Sony Xperia with one file. Using the M4A FFMPEG standard option on 500 full quality. It seems iTunes will not play the 24/96 recorded rates but the phone will. Phone specs state 24/192 capabilities. If I record the LP at 24/88200 rate and export as above, Audacity seems to choose a 195K? bitrate(is this a bug), iTunes WILL play that file, but now the phone won’t. I’m would like to record at the 24/96 as I have a huge media server in the listening room with many TBs of storage, so I also do a WAV export for that room to save wear and tear on my LPs. Can we not turn this into another discussion about my choice of bitrates, it’s what I want to do. So I just would like some info on why I’m getting odd bit rates with AAC exports of like 191kpbs or 195Kbps. Of course this all straightens out if I record at 48K samples, I know that. I record on MBP but edit with Windows 10 2.1.3,
As you point out in the second sentence, the FFMpeg libraries are doing all the advanced format heavy lifting. That’s not an Audacity product.
FFMpeg isn’t perfect. Several people have posted about inconsistencies or errors in the libraries and it’s not unusual for the library version to bump up with an Audacity version.
There is an Audacity 2.2.0 waiting in the wings and that might help with your problems.
Koz
I usually don’t use AAC so I can’t help you much. But, iTunes can convert to AAC and you might have better luck with it. I assume iTunes makes “compatible” AAC files that play on anything that can play AACs.
Or, you could try [u]TAudioConverter[/u] which gives you a couple of different AAC encoder options (including FFmpeg).
…Usually, what you’re converting it from doesn’t matter as long as Audacity can open the file. Audacity uses 32-bit floating-point PCM internally, so the only variable should be the sample rate, and 96KHz is not “unusual”.
so I also do a WAV export for that room to save wear and tear on my LPs
You might want to consider FLAC or ALAC (both are lossless). Tagging (Song/artist/album, etc. metadata) is not well-supported for WAV, and as a bonus your files will be almost half the size.
P.S.
Does this help, or are you already doing this?
From the [u]Audacity User Manual[/u]:
AAC Export Setup
Quality: Currently this slider has no effect > with the version of FFmpeg recommended for Windows and Mac. The exported file is always a constant bit rate (CBR) 196 kbps (stereo) or 98 kbps (mono) file. Exporting more than two channels using Advanced Mixing Options is not currently supported.
To specify a different constant bit rate for a mono or stereo AAC file, choose Custom FFmpeg Export in the Export Audio Dialog. > Then use the Open custom FFmpeg format options button to open the Custom FFmpeg Export Options dialog. See the example on that page for details.
Thanks folks. Well as usual, I got things pretty screwed up. Looks like the only thing 24/96 is gonna play nice with is the FLACs and WAVs. Phone does FLAC but just too much data unless I buy a 128GB card I guess. 96KHZ aint compatible with the AAC or MP3 export right now. So I guess I have to decide whether to go back to recording at 48, or letting the software down-sample on the export. I guess iTunes must be doing that on the AAC convert. The bitrates shown on a converted file are weird unless you do the 320k at 44.1. I understand that AAC doesn’t make that much improvement at higher bitrates anyway. so that’s kind of a waste of data in the long run.
So I guess I have to decide whether to go back to recording at 48, or letting the software down-sample on the export. I guess iTunes must be doing that on the AAC convert. The bitrates shown on a converted file are weird unless you do the 320k at 44.1.
I believe that AAC normally uses variable bitrate, and that may be why the (average?) bitrate seems “weird”… It doesn’t make sense to use a super-high bitrate on silence or “simple” sounds that don’t need it since the whole point of compression is to make a smaller file. Variable bitrate is “smarter” and it makes better use of the available bits.
So I guess I have to decide whether to go back to recording at 48, or letting the software down-sample on the export.
Either way, or you can “manually” downsample before AAC encoding. But, it doesn’t make sense to re-record what you’ve already recorded at 96kHz, or to record twice in the future if you feel you need 96kHz lossless for “home listening”.
…If you know how AAC & MP3 work, they throw-away stuff you can’t hear (mostly sounds that are masked by other-louder sounds), so it’s going to throw-away any ultrasonic information that might be present in the 96kHz file anyway.
I understand that AAC doesn’t make that much improvement at higher bitrates anyway.
That’s generally true. If a 192kbps AAC is transparent (sounds identical to the original) a higher bitrate isn’t going to improve the sound. (With most music, 192kbps is probably transparent in a proper scientific, blind [u]ABX Listening Test[/u].) And at some point, you can run into limitations related to the lossy compression format, but unrelated to the bitrate, so increasing the bitrate doesn’t remove the compression artifacts. (It depends on the particular file… Some music is “easier” to compress than other music, and it depends on your ability to hear small-slight compression artifacts.)
And realistically, AAC is a lossy format but any artifacts from a high-quality AAC will be tiny-tiny compared to the noise & distortion from an LP. 
Given that there is no audible difference between 24/96 uncompressed PCM and 16/44.1 uncompressed PCM, it seems very unlikely that there is any audible benefit in higher bit rate or bit depth after lossy compression. In fact, I would not be very surprised if compressed 24/96 to be slightly less good in terms of sound quality than 16/44.1 that has been compressed at the same bit-rate, because the “hi res” version could be wasting bits on data that is inaudible, thus reducing the number of bits available for encoding audible data. It is certainly the case at low bit-rates that reducing the audio bandwidth can improve the subjective sound quality,
A very real advantage of 16/44.1 is that it is almost universally compatible, which as previous posts in this thread illustrate, cannot be said of 24/96.