Topic: Best Bitrate? VBR extreme or 320bpm (Read 16080 times)

malcolmm · « **on:** January 04, 2009, 07:37:13 am »

I play my mp3s through my hi fi and would love to know what is the absolute bitrate to rip at.

Anyone got the drum on this?

Cheers, Mal from Cessnock (Au)

ADDiCT · « **Reply #1 on:** January 04, 2009, 08:27:11 am »

You could as well ask for the meaning of life. There is no real objective answer to your question, it depends on your hearing, equipment, preference, focus, etc. .

I'm using VBR V0, which is good enough for me. It gives relatively small filesizes with good quality. LAME reference websites claim that V0 is producing artefacts compared to a 320kbps encode, so if diskspace isn't an issue for you, you could go for 320kbps.

I'd recommend to choose one album from your collection, encode it in various bitrates, and listen to the mp3's on the equipment you normally use. This will help you in deciding which encoding options to use.

Matt · « **Reply #2 on:** January 04, 2009, 08:38:37 am »

The best audio quality will be with a lossless format.

Monkey's Audio (APE) or FLAC for example.

mark_h · « **Reply #3 on:** January 04, 2009, 10:06:24 am »

If you *have* to use mp3 with your hi-fi, I'd recommend 320kbps constant bit-rate.

On my Mark Levinson/Revel Salon system I have to listen very closely to spot the difference between 320kbps cbr and the CD. As soon as I engage VBR the differences become much easier to spot...

Ideally though you'd use lossless, such a FLAC.

Mark

malcolmm · « **Reply #4 on:** January 04, 2009, 10:15:50 am »

Thank you Addict and Matt - I've changed my encoder settings and will be interested to listen to Monkey's output and checkout the file sizes.

This forum is excellent - I'm honoured to be able to pick your brains.

Mark - I've listened to 320 and then the original CD and difference is very easy to pick. The Bose and NAD system provide much more presence than with the flattened sound the mp3s provide. - however the mp3s do provide an eaze of use when background listening is all I want.

Have you listened to records? They are so much better again than CDs - are you old enough? (an 80s song - lol)

Is there much difference between FLAC and APE ? ie file sizes (prolly a good rule of thumb)

Cheers, Mal from Cessnock (Au)

mark_h · « **Reply #5 on:** January 04, 2009, 10:26:23 am »

Quote from: malcolmm on January 04, 2009, 10:15:50 am

Mark - I've listened to 320 and then the original CD and difference is very easy to pick.

Perhaps my ears are letting me down as I age. As you can clearly hear the difference you should set up a test CD yourself containing various bit rate conversions of the same track and see which is best for you?

Quote

Have you listened to records? They are so much better again than CDs - are you old enough? (an 80s song - lol)

Vinyl? Of course! I have a diverse collection and a decent turntable/cartridge. But I have to say that, as much as I enjoy listening to vinyl, I do not believe it to be better at capturing the original than CD. Vinyl is certainly different, but it's not accurate. And lately, it's accuracy that I'm after so I generally only listen to a vinyl album if the CD is unavailable.

Quote

Is there much difference between FLAC and APE ? ie file sizes (prolly a good rule of thumb)

Sonically they should be identical (they are both lossless after all). Size will probably be similar. Can APE store tags?

Mark

malcolmm · « **Reply #6 on:** January 04, 2009, 10:36:24 am »

Oh, Tags! They are so important.

They must store tags or I will have to rethink - 320bpm or FLAC

It's late - but I know what I'll be doing tomorrow night. I'll set up the test CD you mentioned.

Thanks Mark. Mal from Cessnock.

Alex B · « **Reply #7 on:** January 04, 2009, 11:02:39 am »

APE and FLAC can both store tags. APE compresses a bit more (for instance, usually APE High produces smaller files than FLAC's best compression setting). On the other hand FLAC decompresses significantly faster than APE.

Correctly encoded LAME MP3 at -V0 (= Extreme in MC) or 320 kbps CBR is practically indistinguishable from the lossless source in a casual listening situation. The comparison setup is probably flawed if the listener can hear obvious differences.

benn600 · « **Reply #8 on:** January 04, 2009, 11:20:13 am »

There are wars started over what format to use. I personally have used FLAC to encode my 17,015 songs...on highest quality (means spends most time encoding for slightly smaller size).

I did a check a few years ago when I started building this library. On a quality setting less than 8, it was either the default of 6 or perhaps 1, and I found that of my library at the time -- probably 12K songs -- it would save me a few gigabytes. I would rather spend a little more time encoding--and I don't really notice any delay at all. Ripping from the CD takes a while too. BTW: I use the separate ripping procedure--it rips a song and THEN encodes it. In the MC advanced settings, it's 1, 1, 2 combined current....just personal preference.

imugli · « **Reply #9 on:** January 04, 2009, 03:36:19 pm »

If I have the physical CD in my collection it's ripped Uncompressed WAV. Storage is so cheap these day to not be an issue any longer. 1TB < AU$200, why would you bother compressing?

JimH · « **Reply #10 on:** January 04, 2009, 03:38:46 pm »

You might try APE. It's lossless, as you probably know. It also supports internal tags, which WAV does not. You can always re-create the exact same WAV file if you change your mind.

benn600 · « **Reply #11 on:** January 04, 2009, 06:55:25 pm »

Quote from: imugli on January 04, 2009, 03:36:19 pm

If I have the physical CD in my collection it's ripped Uncompressed WAV. Storage is so cheap these day to not be an issue any longer. 1TB < AU$200, why would you bother compressing?

WAV is too old. And no tagging. I'm not doing too bad with a 7TB RAID server but I still choose to use FLAC. At 431.4GB of FLAC now. By using FLAC, I have an extra 400GB free...not that I actually do, I'm down to 70GB free on my server -- 1%.

DVDs take the majority. I gotta stop buying complete series of shows....

Matt · « **Reply #12 on:** January 04, 2009, 08:48:26 pm »

Quote from: imugli on January 04, 2009, 03:36:19 pm

why would you bother compressing?

The built in MD5 / checksum system in APE is pretty slick. You can test the bit-perfect validity of your files, which isn't possible with WAV.

I had bad system memory years back and this helped me find the couple files that got goofed. With WAV, I'd have just never known.

malcolmm · « **Reply #13 on:** January 04, 2009, 10:34:25 pm »

Everybody... thanx for sharing your knowledge.

Cheers, Mal from Cessnock (Au)

benn600 · « **Reply #14 on:** January 04, 2009, 10:39:01 pm »

Matt - Does FLAC have a checksum? Is there a way to test my FLAC files for corrupted files? I get paranoid that over time, random files would get modified. Since I own the original CDs, I could then re-rip the ones I needed to.

Speaking of which, there are at least two songs that I hear a partial second of silence in the song. I really need to find the CD and see if it's in there CD audio. I'm wondering if it somehow got added to the FLAC--which wouldn't be good.

xen-uno · « **Reply #15 on:** January 05, 2009, 03:34:22 pm »

Benn ... see Synthetic Soul's post here ...

http://www.hydrogenaudio.org/forums/index.php?showtopic=45380

Malcomm ... lossless is the way to go if you want the absolute best (say for an on/off line archive). Keep in mind though that with the latest lame/vorbis/musepack encoders at bitrate averages starting around 170 or so (VBR), it is dam near (or simply is) impossible to ABX between the lossless and the lossy file.

mark >> "As soon as I engage VBR the differences become much easier to spot..."

... engage? ... this does not sound like an ABX test. Sounds like your relying on memory which is subjective and inaccurate ... at least mine is.

mark_h · « **Reply #16 on:** January 05, 2009, 04:00:29 pm »

Quote from: xen-uno on January 05, 2009, 03:34:22 pm

mark >> "As soon as I engage VBR the differences become much easier to spot..."

... engage? ... this does not sound like an ABX test. Sounds like your relying on memory which is subjective and inaccurate ... at least mine is.

I never rely on memory when testing as I know how faulty it is. In my system I am usually able to set up parallel feeds, eg one FLAC and one MP3, synced together, ensuring the same cables, players (usually Squeezeboxes) and so on so I can then flip live between the two streams, or get somebody to do it for me so I cannot influence things, and decide which is the best. This way memory is not involved and any differences are typically and immediately apparent.

Cheers,

Mark

glynor · « **Reply #17 on:** January 05, 2009, 04:09:16 pm »

Quote from: Alex B on January 04, 2009, 11:02:39 am

Correctly encoded LAME MP3 at -V0 (= Extreme in MC) or 320 kbps CBR is practically indistinguishable from the lossless source in a casual listening situation. The comparison setup is probably flawed if the listener can hear obvious differences.

QFT

benn600 · « **Reply #18 on:** January 05, 2009, 05:34:07 pm »

Yes, someone please point me to a procedure to use to test mp3 vs. flac (or any low to high quality audio comparison). I think that an immediate switch over is about the only way I'll be able to tell. I've, quite frankly, never been able to notice anything noticeable. But, I know I wasn't following a good procedure. I'd like to create a reproducible test that consistently allows me to pick which is which. I've tried taking an mp3 and flac file into Audacity and then splitting them at exactly the same points. If it is exported as wav, both should reproduce fully the same. I can't ever line the match points up. I think the problem is sometimes it's removing quiet space and sometimes it isn't.

bwaldron · « **Reply #19 on:** January 05, 2009, 05:44:45 pm »

Quote from: glynor on January 05, 2009, 04:09:16 pm

QFT

Seconded.

benn600 · « **Reply #20 on:** January 06, 2009, 08:23:01 pm »

http://netforbeginners.about.com/od/forumsandlists/f/whatisQFT.htm

steveklein · « **Reply #21 on:** January 08, 2009, 10:34:07 am »

for what it's worth, i can't really tell a difference between a good rip at VBR with a ~200 average versus 320 constant bitrate or the original CD.

i really think the differences are overblown. i dare say less than 5% of the population could identify the difference between 192 and the original source CD more times than not... unless we are talking about many, many thousands dollars worth of equipment.

just my .02

mark_h · « **Reply #22 on:** January 08, 2009, 10:45:06 am »

Quote from: steveklein on January 08, 2009, 10:34:07 am

i really think the differences are overblown. i dare say less than 5% of the population could identify the difference between 192 and the original source CD more times than not... unless we are talking about many, many thousands dollars worth of equipment.

just my .02

It's not until 320kbps that I really struggle to tell the difference, but even on my iPod I can hear the compression on certain tracks that highlight mp3 weakness (typically instruments such as tambourines, cymbals etc).

But it all depends on what I'm doing - if I'm listening on my computer I more likely to be happy with lower bit rates. On my iPod it's 320kbps minimum. On my hi-fi it's lossless only.

Cheers,

Mark

benn600 · « **Reply #23 on:** January 08, 2009, 08:58:05 pm »

Any suggestions on how to run a listening test myself? How do those who CAN tell a difference run their test?

CadErik · « **Reply #24 on:** January 08, 2009, 10:10:26 pm »

Quote from: xen-uno on January 05, 2009, 03:34:22 pm

with the latest lame/vorbis/musepack encoders at bitrate averages starting around 170 or so (VBR), it is dam near (or simply is) impossible to ABX between the lossless and the lossy file.

No one has mentionned wma but in my own listening tests its VBR mode sounded the best and there is also a lossless setting up to 24 bits (not that this would be of any use for cds...)

Erik.

mark_h · « **Reply #25 on:** January 09, 2009, 03:25:55 am »

Quote from: benn600 on January 08, 2009, 08:58:05 pm

Any suggestions on how to run a listening test myself? How do those who CAN tell a difference run their test?

One simple way is to find a quiet piece of music with high frequencies in it; tambourines, chimes, cymbals etc as these are often the first things to go wrong in mp3. You can then rerip at higher bit rates and see if you can hear the distortion (it sounds like a sort of tinkly glass sound; you'll know it when you hear it).

When I test stuff like this I set up two players (typically squeezeboxes) into my hi-fi and then I can run both streams simultaneously and simply flip between the two for testing. If I cannot use two devices than I create very small samples, eg 3-5 seconds long and put them on loop. I listen to one loop for a good few times to get the sound in my head and then flip. Again, differences become very obvious.

Or on this you can just trust us

If you want the best possible sound, you want lossless.
If it must be mp3 you want 320kbps cbr.

Mark

benn600 · « **Reply #26 on:** January 09, 2009, 09:53:45 am »

I've already trusted this whole idea. My entire collection is FLAC. I would like to notice this on my own.

Now years ago (probably 8 or more!) I had crappy speakers. I didn't think the bitrate was important and had limited HD space. So I re-encoded my entire junkie library to 64Kb. Then, I bought myself some nicer speakers 2-way plus subwoofer. The addition of the tweeter made audible distortion disturbingly obvious, especially in the treble. It took me most of this time to bite the bullet and start brand new. I literally threw out that collection (well, backed to about 10 DVDs) and then ripped our entire ~500 CDs. Since then, I've almost tripled our number of CDs--which allows me to rip in FLAC and scan the cover art at 1200x1200. I look at this library as the one that I will use forever. It is ripped in the best quality format--more importantly, lossless. I am also taking extra care to ensure tagging accuracy. Lastly, I have a thread going in Other Hardware about how I'm trying to finish rating every song in my library by December 22, 2009. Have been getting ahead of my required rating speed already so I'm serious about meeting the deadline.

Magic_Randy · « **Reply #27 on:** January 09, 2009, 10:35:05 am »

I'm in the process of re-ripping my library. I'm re-ripping everything as .mp3 320 kbps CBR. I believe this is not the best audio quality, but gives the best balance between quality and total portability. Everything works with .mp3 files.

If people think I'm nuts, let me know. It's not too late to change strategies.

Randy

mark_h · « **Reply #28 on:** January 09, 2009, 10:53:45 am »

Quote from: benn600 on January 09, 2009, 09:53:45 am

I've already trusted this whole idea. My entire collection is FLAC. I would like to notice this on my own.

Now years ago (probably 8 or more!) I had crappy speakers. I didn't think the bitrate was important and had limited HD space. So I re-encoded my entire junkie library to 64Kb. Then, I bought myself some nicer speakers 2-way plus subwoofer. The addition of the tweeter made audible distortion disturbingly obvious, especially in the treble. It took me most of this time to bite the bullet and start brand new. I literally threw out that collection (well, backed to about 10 DVDs) and then ripped our entire ~500 CDs. Since then, I've almost tripled our number of CDs--which allows me to rip in FLAC and scan the cover art at 1200x1200. I look at this library as the one that I will use forever. It is ripped in the best quality format--more importantly, lossless. I am also taking extra care to ensure tagging accuracy. Lastly, I have a thread going in Other Hardware about how I'm trying to finish rating every song in my library by December 22, 2009. Have been getting ahead of my required rating speed already so I'm serious about meeting the deadline.

Been on a very similar journey (re-ripping, tagging and voting) but am now fully rated on my 22,000+ song collection. Took some doing but I can now generate some very nice playlists using smartlists.

I have a nice view that shows the last song played compared against others in the library (with various versions; same artist;same genre etc) so that I can vote more accurately as well.

Mark

Peter_T · « **Reply #29 on:** January 09, 2009, 11:18:35 am »

Quote from: Magic_Randy on January 09, 2009, 10:35:05 am

I'm in the process of re-ripping my library. I'm re-ripping everything as .mp3 320 kbps CBR. I believe this is not the best audio quality, but gives the best balance between quality and total portability. Everything works with .mp3 files.

If people think I'm nuts, let me know. It's not too late to change strategies.

Randy

I'm doing the same, Randy. I have the space for FLAC and a system worthy of lossless audio, but mp3 sure is widely accepted, from portables to different software, etc. I have started making FLAC backup, but just to store, not to play. God I'm anal.

mark_h · « **Reply #30 on:** January 09, 2009, 11:24:47 am »

Don't forget that MC can do conversions for you... so store FLAC then convert to 320kbps mp3 for ipod on the fly...

Mark

Peter_T · « **Reply #31 on:** January 09, 2009, 12:44:06 pm »

Quote from: mark_h on January 09, 2009, 11:24:47 am

Don't forget that MC can do conversions for you... so store FLAC then convert to 320kbps mp3 for ipod on the fly...

Mark

You're right, of course, but I found it awkward managing two copies of the file, despite stacking and cache folders. So I shove my FLAC files deep onto some redundant NAS storage as a bit of contingency plan for when the aliens come and shatter all my CDs. It will keep me from crying.

benn600 · « **Reply #32 on:** January 09, 2009, 01:23:38 pm »

If 320 makes you happy, then by all means. I chose FLAC because it is basically AS GOOD AS IT GETS. My point is that since it is lossless, you can later re-encode to another lossless format without losing any information. It averages perhaps 900 Kb over an average library with highs in the 1,100s and lows in the 800s...some variation is easily noticed with heavy metal vs. classical or quieter songs.

Either way, I was going for a once and for all rip. I had ripped my entire collection probably two times previous to me deciding on FLAC. For my huge library, it only takes 431GB. You can fit all that on a hard drive priced at $100 or less! Buy two to have a backup. So $200 to store many thousands of dollars worth of CDs. Of course you can spend a lot more if you want RAID, etc., and more than 2 copies (which I recommend). I have my music stored on my RAID server plus three terabyte hard drives, which keep a copy that I typically renew on the 15th of each month. So I have current plus three previous months -- or 4 total copies. I'm still developing my backup strategy and since I have several unused 500GB drives, I may use these for music only. My terabyte drives store everything on my server except video.

INTERACT FORUM

Author Topic: Best Bitrate? VBR extreme or 320bpm (Read 16080 times)