Is QuickTime any better than Nero nowadays?

[Anakunda]

Hello ! I see many audio tools use NeroAac by default. But Nero wasnot updated for years there comes question if it's still better encoder than encoders that are regularly upddated ie. QuickTime or Fraunhofer. I remember some comparison where qtaacenc beaten Nero completely. Also what's your opinion for good encoder for movie audio,ie. audio with 40-80 kb/s per channel, and with witch bitrate it gives transparent sound with fully preserved psychoacoustic?

[sneaker_ger]

http://forum.doom9.org/showthread.php?t=161957

[nibus]

I think Quicktime is better, and the new listening tests also reflect that opinion.

I use qaac.exe to convert to AAC - qtaacenc.exe doesn't correctly remap multichannel audio for the Quicktime encoder, but qaac.exe does. Plus it's regularly updated.
http://sites.google.com/site/qaacpage/

Quicktime's TrueVBR mode allows quality settings similar to Nero, only the quality ranges from 1-127.

For stereo, these are the approximate bitrates:

Q0 - Q4 = ~40 kbps
Q5 - Q13 = ~45 kbps
Q14 - Q22 = ~75 kbps
Q23 - Q31 = ~80 kbps
Q32 - Q40 = ~95 kbps
Q41 - Q49 = ~105 kbps
Q50 - Q58 = ~115 kbps
Q59 - Q68 = ~135 kbps
Q69 - Q77 = ~150 kbps
Q78 - Q86 = ~165 kbps
Q87 - Q95 = ~195 kbps
Q96 - Q104 = ~225 kbps
Q105 - Q113 = ~255 kbps
Q114 - Q122 = ~285 kbps
Q123 - Q127 = ~320 kbps


Of course it depends on the audio being transcoded. I use a higher setting for music than I do movie soundtracks. I've found that TVBR 60 is good for small file size while still retaining good quality. One thing you will notice about Quicktime's TVBR codec is that it is very flexible and the bitrate fluctuates much more than any other AAC codec I've seen.

For multichannel audio I would recommend 64-96kbps per channel. So for 5.1 you would want 384-576 kbps, though for full transparency I would go 96kbps per channel. This equates to TVBR around 100-113 quality setting, though TVBR 70 and up will also give good results.

Experiment with your own audio and see what you think.

[Anakunda]

Thanks, very informative. Now how can I encode in qaac with eac3to? Wuold the pipe eac3to stdout.wav | qaac preserve the correct multichannel mapping?
And moreover, gives Vorbis better quality at q5.0 for 5.1 audio, than QuickTime at TVBR 100 (or Nero at q0.45) ?
And finally are these all better than AC3 at 384kbps, when converted out from lossless DTS or TrueHD?

[nibus]

Yes, piping works great with eac3to:


eac3to.exe "input.dts" stdout.wav -normalize | qaac.exe --ignorelength -V 100 - -o "output.m4a"

The correct channel mapping is preserved, however if your source is 6.1 (7 channels) you need to specify the following:


eac3to.exe "input.dts" stdout.wav -0,1,2,3,5,6,4 -normalize | qaac.exe --ignorelength -V 100 - -o "output.m4a"

To downmix to 5.1:

eac3to.exe "input.dts" stdout.wav -0,1,2,3,5,6,4 -down6 -normalize | qaac.exe --ignorelength -V 100 - -o "output.m4a"

I believe for the bitrate multichannel AAC is better than Vorbis, but it is probably very close. The one advantage of AAC is I believe it has greater compatibility than Vorbis, but YMMV.

Quicktime TVBR 100 is probably more comparable to Nero at q0.60 rather than q0.45.

And yes, Quicktime TVBR / Nero / Vorbis are all much more efficient than AC3. The only advantage to AC3 is that you can bitstream it over SPDIF, while AAC you have to send it via HDMI as uncompressed PCM.

edit: to answer your question from the other qaac thread - qaac.exe by default uses -q 2 which is the highest (and slowest) quality. So you don't have to specify anything if you want the highest quality encoding.

[mindbomb]

i thought this was an aac thread lol, i missed the vorbis option.

yea, i would use aotuv b6.03 for lossy audio for a movie, assuming compatibility isn't an issue.

for music, i actually am currently using qaac.

edit:
now im using fraunhofer for music actually.

[Anakunda]

Thanks for all suggestions, It's said that Vorbis @q5 gives already excellent audio quality. That's less on bitrate than what I've achieved by qaac at TVBR 100 (85 kbps per channel), but because of possible compatibility problems I'll probably stick with Aac. Now qaac looks promissing as a good replacement for Nero, I'll surely give it a try.

[Anakunda]

Yes, piping works great with eac3to:


eac3to.exe "input.dts" stdout.wav -normalize | qaac.exe --ignorelength -V 100 - -o "output.m4a"


Hello there, I'm trying qaac currently and have an additional question. The pipe works very fine but I want to know the importance of -down16 switch with this backend encoder. I know StaxRip adds it by default to eac3to when encoding to AAC of Nero but does it matter when encoding via QuickTime? I observed that not giving that switch and converting from lossless DTS with 24bit/48kHz/6chn lead to almost 20GB big temporary WAV file but the resulting AAC file is approx. the same size as if encoded with -down16 option, thus I think it is still 16bit AAC anyways. So the question, provides removing -down16 an additional quality to AAC file or is it just overkill?

[kypec]

AAC (and other lossy formats like MP3) do not incorporate bit-width per sample, they are based on frequency domain transformation instead. Therefore you can't really say that it's 16-bit AAC as there is no such thing. Consequently, AAC encoded file will be almost identical filesize-wise no matter what bit-width is used for its input. Using -down16 option can probably cause slight loss of quality, especially for highly dynamic sources but I wouldn't care too much about it if I was rather constrained by temporary WAV files disk space requirements ;)

[zys4416]

I used to use nero aac encoder encoded multichannel audio. Since I know there is a delay before the aac (50ms or so), I decided to switch to ogg Vorbis format.
I am using Oggenc2.87 (aoTuVb6.03), it improve the multichannel audio code, under q5 quality mode provides great sound quality (~400kbps @ 6ch).
For music, I always using lame.

[Atak_Snajpera]

for 5.1 aac 320 kbps will give you the same quality as 128 kps for 2.0. center channel and lfe are alot easier to compress and therefore they do not require much bitrate.

[Anakunda]

I used to use nero aac encoder encoded multichannel audio. Since I know there is a delay before the aac (50ms or so), I decided to switch to ogg Vorbis format.
I am using Oggenc2.87 (aoTuVb6.03), it improve the multichannel audio code, under q5 quality mode provides great sound quality (~400kbps @ 6ch).
For music, I always using lame.

That's what I've encountered too. Nero inserts about 50-200ms delay in dependency on quality (the lower q the longer delay) but I feel some delay with qaac too (about 50ms).
I suppose that Vorbis at bitrates around 80k/channel (about Q5) gives better quality than any AAC (Nero or QT) at these bitates but the big handicap is reduced compatibility especially on HW players.

[Atak_Snajpera]

From my tests CT/FHG HE-AAC at 64kbps sounds better than Vorbis at q0 (~64kbps).

I suppose that Vorbis at bitrates around 80k/channel (about Q5) gives better quality than any AAC (Nero or QT)

5.1 AAC is transparent at 320 kbps.
5.1 AC3 is transparent at 640 kbps
5.1 DTS at 1.5 Mbps

Anything above that is just overkill. In your case 80*5=400kbps

[pururin]

In Hydrogenaudio wiki, they state that
Nero AAC "is generally percieved to have the highest quality VBR LC AAC implementation"
http://wiki.hydrogenaudio.org/index.php?title=Nero_AAC

Now that it is 2012 now, is this still true, especially for 5.1 channels AAC?

Or that sentence can be rightfully replace to the new champion Quicktime AAC once and for all ?

[smok3]

pururin: afaik there were no 5.1 tests, there was 96 kbps test where i think qt won.

[pururin]

Guess I'll have to try QAAC then.

[SeeMoreDigital]

Guess I'll have to try QAAC then.
You may be interested to know that LoRd_MuldeR has added QAAC encoding to his beta builds of LameXP... It works very well ;)

[Asmodian]

Using qaac TVBR @ 127 I have had some audio that seemed lower quality, mostly noticeable on weird tonally complex sound effects. It was pretty subtle but I did notice it while just watching a recent encode, not expecting an issue with the audio. Going back to the flac (which was a transcode of a DTS-HD source) sounded better.

Using Nero at quality=0.70 gave a perceptually transparent aac which was also much larger. I didn’t test qaac in normal vbr mode. I assume most of its benefit is the new tvbr mode? I also assume I wouldn't be able to tell the difference with my hardware and ears if the files were the same size (and the bitrates this high). For most music or human voice and/or targeting a small file size I believe TVBR is great but if you want maximum quality and can spare the space (sill about 1/2 the size of flac) I believe a constrained VBR mode can provide better quality.

This was not a double blind test and based on only one track where I did a lot of testing but given that I am not too space conscious (another 30MB won't change anything for me) I am sticking with Nero (via eac3to).

[pururin]

You may be interested to know that LoRd_MuldeR has added QAAC encoding to his beta builds of LameXP... It works very well ;)
I'm interest indeed. Thx for the info!

Using qaac TVBR @ 127 I have had some audio that seemed lower quality, mostly noticeable on weird tonally complex sound effects.
...
This was not a double blind test and based on only one track where I did a lot of testing but given that I am not too space conscious (another 30MB won't change anything for me) I am sticking with Nero (via eac3to).
That sounds scarry, QT-AAC still not stable?
But Nero didn't update for years. I wonder if they still developing (at least a bit)?

[LoRd_MuldeR]

Doesn't look like Nero is still working on their AAC encoder. At least not on the stand-alone CLI encoder that you can download for free :(

There is a long-standing bug with flushing the STDOUT, which exists for years and could have been fixed with a single line of code...

[Asmodian]

That sounds scarry, QT-AAC still not stable?
But Nero didn't update for years. I wonder if they still developing (at least a bit)?

It isn't that I think QT-AAC might not be stable it is just that its TVBR mode may not be suited to all content and I think TVBR is designed to do decent low bit rate encodes rather than higher bit rate audibly transparent encodes. I wanted to point this out as the reviews I have read about TVBR had always stated it to be higher quality than constrained VBR modes. I think this only holds true at lower bit rates.

I will try some normal VBR modes in QT-AAC but this kind of test isn’t very definitive as it depends on the source (and it is subjective). I assume QT-AAC in normal VBR mode would be audibly equivalent to Nero at high bit rates.

[IgorC]

My personal choice varies on bitrate basis.

For stereo:
64 kbps - FhG HE-AAC encoder (from last Winamp).
80 -160 kbps - Apple QuickTime/iTunes AAC encoder
192 kbps and higher - Apple, FhG and Nero. All of them are excelent at such high bitrates.