HydrogenAudio

Lossy Audio Compression => MP3 => MP3 - General => Topic started by: BFG on 2013-06-01 02:37:32

Title: LAME -V0 -Y
Post by: BFG on 2013-06-01 02:37:32
When encoding MP3s, I usually use these command line options with LAME.  My goal is to ensure transparency in (virtually) all cases for both myself, and anyone else who might be listening, but to also keep a decent amount of compression so as to make nearly all of my library portable on a single memory card.

Without coloring the conversation with my own thoughts (at least at first), I'd like to hear comments from others on this forum - many of whom clearly have better hearing than I - on using this combination of options.
Title: LAME -V0 -Y
Post by: LedHed8 on 2013-06-01 03:35:26
BFG- I am a new member of this forum, but have been visiting for several years.  I appreciate the hard work and lively discussions from all the members.  As far as V0- Y switch, I have not been able to ABX the difference between V0 and V0- Y Switch.  I have found myself using this setting more frequently for portable music.  I've also found that I like Halb27's v3+ variant for portables.
Title: LAME -V0 -Y
Post by: BFG on 2013-06-03 17:34:51
Thanks for the reply.  I'm pretty satisfied with the results so likely will keep these settings (though I am using halb27's variant of LAME 3.100a2).  I just wanted to see if anyone with better hearing had ever tried this, and if so, if they had disliked the results.
Title: LAME -V0 -Y
Post by: Makaki on 2013-06-03 18:26:34
As far as the frequencies above 16 kHz go. Personally, I've tried on different equipment and either I can't hear them or the equipment can't reproduce those frequencies properly. Thus in my case, it seems the -Y switch fits like a glove.

I found a sweep test (in FLAC) going from 10 Hz to 20 kHz. It has individual files for each tone, so unlike continuous sweep tests, you know the exact frequency of the tone that is playing at any given time. Which is useful when the device being tested doesn't give you much information on it's own.

http://www.wavecor.co.uk/testdisk.html (http://www.wavecor.co.uk/testdisk.html)

I suggest keeping the test in a lossless format (FLAC or WAVE) to avoid any high-pass/low-pass filtering or artifacts done by encoding. If you must encode, take note of this.
Title: LAME -V0 -Y
Post by: LedHed8 on 2013-06-05 06:02:04
As far as the frequencies above 16 kHz go. Personally, I've tried on different equipment and either I can't hear them or the equipment can't reproduce those frequencies properly. Thus in my case, it seems the -Y switch fits like a glove.

I agree with you 100% Makaki.  My experiences have been exactly the same.  I like the idea of knowing that I am using the highest quality lame vbr psych-acoustic algorithim with the added benefit of a little space savings.  I would suggest that you try out halb27's latest functional extension if you are either looking for more space-savings and/or quality headroom.  We have 4 Rockbox'd Sansa portable players in our household of various capacities and dap qualities.  As a result, I have files for the purpose of space/quality balance, and "overkill" files for my obsessive/compulsive paranoid side.  Not that I could ever really abx a flaw at those high quality settings, but the thought that I might actually "hear" one someplace/sometime in my best "lossy" portable files compels me to try the latest "upgrades". 

To date, I've used most of the common lossy codecs: mp3 (lame, iTunes/Fraunhofer) iTunes aac, Nero aac, wma std and pro, and Vorbis.  After thousands of hours of mixed casual and intense listening, my personal preferences are Vorbis aoTuVb6.03 for low-to-mid bitrates and lame for mid-to-high bitrates. 

Many happy encodes.   

LedHed8
Title: LAME -V0 -Y
Post by: slks on 2013-06-05 07:44:07
The -Y switch is great, if you can't hear tones above 16 kHz. You won't notice the high-frequency degradation, if you can't hear that range to begin with. It sounds like you've done some testing and determined it works for you. (If you haven't done testing yet, I'd advise doing that).

Of course, individuals with better hearing listening to your files might notice a degradation. Whether you care about them is up to you!

Unfortunately I've never messed with -Y myself, and I don't have great equipment to test on, so I can't give any personal experience.
Title: LAME -V0 -Y
Post by: Dynamic on 2013-06-05 14:49:31
...and for those casual readers, do remember:
• a 16kHz low-pass filter is usually considered very hard to ABX in real music - only test tones make it easy if you can hear those last few semitones of human hearing range.
• the -Y switch does NOT filter out frequencies over 16 kHz anyway. It just permits LAME to represent them with slightly lower precision on those occasions where the psychoacoustic model's desired precision would require all lower frequencies to get excessively high precision and thus waste a large amount of bits encoding inaudible data. (This is called sfb21 bitrate bloat)

For more info, check the Wiki (HydrogenAudio Knowledgebase) - see bar at top of page - which has a good article about -Y and sfb21.

The latest versions of LAME have built in some tricks to greatly reduce the bitrate bloat when -Y is not used.