Has anyone noticed any difference in sound quality between LAME and Fraunhofer for CBR 256 or CBR 320 files? Or are both encoders transparent a these levels?


FYI - I use LAME 3.97 vbr 0 for my personal encodings. I am asking because most of the legal digital audio now available for sale is encoded at one of these 2 settings by one of these 2 encoders. I'm wondering if it is all pretty much the same quality.

I think it's pretty hard / impossible to spot a difference between the two.
You may be able to spot a killer sample in the FhG codec more often than LAME.
Notice that LAME has been the recommended encoder in here for a long while.

Sure there are differences, but no way to say which encoder is better.
It also depends a bit on the version used.

My personal impression, partially backed up by abxing:

a) severe pre-echo problem eig is relatively good with Lame 3.98 and FhG alternate codec (very high quality setting) at 256 kbps. It is very bad with the FhG codecs easily available today (CLI surround encoder, WMP 11 encoder, dbPoweramp encoder).
b) harpsichord attacks can be pretty bad with FhG alternate codec.
c) quality of tonal spots is a tiny bit better to me with FhG encoders than when using Lame.
d) overall quality is very good at 256 kbps with any of these codecs when not caring too much about specific problem scenarios which usually are extremely rare in practice.

iTunes Plus uses 256 kbps AAC and regular iTunes uses 128 kbps AAC (http://www.apple.com/pr/library/2007/05/30itunesplus.html). The iTunes Store is the largest music retailer in the U.S. and it has sold over 5 billion songs worldwide (http://en.wikipedia.org/wiki/ITunes_Music_Store#Market_share_and_milestones).

AAC is not encoded by either LAME or Fraunhofer.

FhG tends to suppress all above 16K even at 320. It also uses simple stereo at higher bitrates.

I would definitely prefer a recent version of Lame at high bitrate to FhG if I had to pick blindly before listening.

WOW Fhg still uses 16k filter even at 320, i know it uses pure stereo at bitrates over 224. I know that the latest sorround Fhg encoder that supports VBR only uses frames upto 256, to avoid the 320 bitrate reservoir ISO limitation.

No. FhG alternate as implemented within Magix Cleaning Lab 2007 loves HF (a bit too much for my taste), but also the current dbPoweramp FhG implementation isn't very cautious about HF beyond 16 kHz.

Yes.

I stand corrected. It has been a long time since I did my testing. I used older versions and tested many downloaded files.

I did not realize dbpoweramps' FhG codec was not the best version of FhG available. Where does one obtain the best version? Is it ossible to tell the different versions? encspot does not seem to be able to.

Not that it matters that much if everything is really good at 256kbps or higher

dbPoweramp's FhG codec is bad with eig, but I wouldn't call it bad just because of this.
I did use it with CBR 192 (highest bitrate using joint stereo) and was happy with the results for practical purposes (not talking about problem samples). CBR 256 is fine as well of course when not looking at eig.

Sure this is valid for other FhG codecs and Lame as well, so in the end it's impossible to say what is the best way to go. Usually everything is fine, and differences in problem sample behavior are hard to evaluate when there's no clear bias towards a certain codec. This is exactly the situation we're in IMO.

A subjective approach is adequate IMO. If at the moment you have a personal bias towards a certain codec, say Lame 3.98b8 for instance, take a series of tracks from your collection which have a strong emotional meaning to you, and listen very carefully to the encoder's result. If everything's fine: stick with your encoder. If you're not totally content move on to another candidate. If you do the careful listenig by abxing chance is high you will be content with your bias codec. Enjoy it.