Aurally
Hasn't posted much yet...
- Joined
- Sep 13, 2017
- Messages
- 55
- Reaction score
- 27
- Points
- 18
- Location
- Massachusetts
- Genesis Model Type
- 2G Genesis Sedan (2015-2016)
Hi,
Having read most of the info on encoding music on this forum, and having tried out some of it myself, I am posting my learning for those seeking an up to date summary. I am not an authority on this, just someone who has spent more time than the average person researching and looking into this topic.
This psot is based on the following:
- You have a bunch of high resolution lossless music files that you seek to convert to usb with highest possible fidelity, at the expense of storage capacity. These could be 24/96 files from HDtracks, or surround files from video, and maybe even iso images of SA-CD and DVD-audio if you happened to possess them.
- You want a simply enough workflow to convert the files in one environment, in my case using Foobar 2000.
- The tests are done on the Lexicon 17 speaker system offered in the Ultimate package.
My choice for encoder boiled down to AAC encoding (using the Nero AAC encoder as an add-on in Foobar) for the following reason:
1. AAC is better than mp3 for the following reasons:
- support for higher sampling rates up to 96kHz.
- support for more than 2 stereo channels.
- support for bit rates higher than 320 kb/sec.
2. AAC is easier to use than WMA9: I have had no success in finding a good way to encode using WMA, and from what I have read, I am not sure higher resolution/quality is supported compared to AAC. I have not verified if the system supports WMA lossless.
3. Ogg vorbis is supported but similar to AAC at best.
4. DTS encoders are not easily available.
The main thing that hasn't been explicitly mentioned to date is that since the system can play back MP4V3 video files, it implicitly supports DVD quality audio, which I believe is AAC-LC (low complexity) up to 448kB/sec. I have encoded and played files back successfully setting an average bit rate of 448kB/sec with 96kHz sample rate or 5.1 channels, but >465kB/sec fails consistently.
So, if you have lossless stereo files, I recommend using AAC at 448kB/sec at the highest appropriate sampling frequency, up to 96kHz. At this bit rate, the bass is better defined (not as sloppy), and the dynamics seem to be better compared to lower bit rates. I do believe most people will hear a difference if the music is played at sufficiently loud volumes, and the source music is of sufficiently high quality.
The question is more complicated for multichannel music. Since the 448kB/sec limit is for all channels combined, it is not clear how to use up this amount of data, i.e.:
A. encode individual channels (e.g. 448/5=90kB/sec per channel), or downmix to stereo (224kB/sec per channel) and let the Logic 7 algorithm reproduce the extra channels?
B. encode at highest possible sampling frequency, or decimate to lower sampling frequency?
Regarding A, my experimentation is that if you have a well mastered 5 channel recording, it is best to keep them as discrete channels. Compared to stereo downmixing, I generally hear better localization and better definition of the sources by keeping it multichannel. Though I do believe that some may prefer the downmix as it provides a more coherent and fuller sound.
Regarding B, from what I read about AAC encoding, higher sampling rates benefit time domain fidelity, where as lower sampling rates give more allocation to frequency info (see link below). To me, this implies that high transient music (guitar plucking etc...) will benefit from high sample rate, where as vocal choir might be better at low sample rates (i.e. decimate to 48kHz as opposed to encoding at 96kHz). In the end, one would need to hire the original mastering engineer to encode this for the best results. Therefore, my decision is to pick one setting for all as the additional time spent is not worth the trouble.
The most relevant article I have found related to this is:
https://www.iis.fraunhofer.de/conte...CD-Quality-24-96HighResolutionAAC_AES5476.pdf
Maybe someone better versed in audio/video encoding can uncover other hidden encoding schemes with higher quality that the system supports? Could the system support ALS? https://en.wikipedia.org/wiki/Audio_Lossless_Coding
Having read most of the info on encoding music on this forum, and having tried out some of it myself, I am posting my learning for those seeking an up to date summary. I am not an authority on this, just someone who has spent more time than the average person researching and looking into this topic.
This psot is based on the following:
- You have a bunch of high resolution lossless music files that you seek to convert to usb with highest possible fidelity, at the expense of storage capacity. These could be 24/96 files from HDtracks, or surround files from video, and maybe even iso images of SA-CD and DVD-audio if you happened to possess them.
- You want a simply enough workflow to convert the files in one environment, in my case using Foobar 2000.
- The tests are done on the Lexicon 17 speaker system offered in the Ultimate package.
My choice for encoder boiled down to AAC encoding (using the Nero AAC encoder as an add-on in Foobar) for the following reason:
1. AAC is better than mp3 for the following reasons:
- support for higher sampling rates up to 96kHz.
- support for more than 2 stereo channels.
- support for bit rates higher than 320 kb/sec.
2. AAC is easier to use than WMA9: I have had no success in finding a good way to encode using WMA, and from what I have read, I am not sure higher resolution/quality is supported compared to AAC. I have not verified if the system supports WMA lossless.
3. Ogg vorbis is supported but similar to AAC at best.
4. DTS encoders are not easily available.
The main thing that hasn't been explicitly mentioned to date is that since the system can play back MP4V3 video files, it implicitly supports DVD quality audio, which I believe is AAC-LC (low complexity) up to 448kB/sec. I have encoded and played files back successfully setting an average bit rate of 448kB/sec with 96kHz sample rate or 5.1 channels, but >465kB/sec fails consistently.
So, if you have lossless stereo files, I recommend using AAC at 448kB/sec at the highest appropriate sampling frequency, up to 96kHz. At this bit rate, the bass is better defined (not as sloppy), and the dynamics seem to be better compared to lower bit rates. I do believe most people will hear a difference if the music is played at sufficiently loud volumes, and the source music is of sufficiently high quality.
The question is more complicated for multichannel music. Since the 448kB/sec limit is for all channels combined, it is not clear how to use up this amount of data, i.e.:
A. encode individual channels (e.g. 448/5=90kB/sec per channel), or downmix to stereo (224kB/sec per channel) and let the Logic 7 algorithm reproduce the extra channels?
B. encode at highest possible sampling frequency, or decimate to lower sampling frequency?
Regarding A, my experimentation is that if you have a well mastered 5 channel recording, it is best to keep them as discrete channels. Compared to stereo downmixing, I generally hear better localization and better definition of the sources by keeping it multichannel. Though I do believe that some may prefer the downmix as it provides a more coherent and fuller sound.
Regarding B, from what I read about AAC encoding, higher sampling rates benefit time domain fidelity, where as lower sampling rates give more allocation to frequency info (see link below). To me, this implies that high transient music (guitar plucking etc...) will benefit from high sample rate, where as vocal choir might be better at low sample rates (i.e. decimate to 48kHz as opposed to encoding at 96kHz). In the end, one would need to hire the original mastering engineer to encode this for the best results. Therefore, my decision is to pick one setting for all as the additional time spent is not worth the trouble.
The most relevant article I have found related to this is:
https://www.iis.fraunhofer.de/conte...CD-Quality-24-96HighResolutionAAC_AES5476.pdf
Maybe someone better versed in audio/video encoding can uncover other hidden encoding schemes with higher quality that the system supports? Could the system support ALS? https://en.wikipedia.org/wiki/Audio_Lossless_Coding
