MP3 Quality/Conversion: LAME vs. others

[brandonjp]

When I rip an entire cd to wave it's take just under 2 min. When I rip to MP3 it'll take around 13-15 min. I'm encoding everything at 320kbps. In an old post ( http://forum.dbpoweramp.com/showthread.php?t=2176 ) Spoon mentioned a noticable difference at 128kbps when using the faster but crappy codecs. Could a faster codec still give me quality if everything's at 320?

It's not been a big problem, but there are times when I just wish it was faster. I'm a dedicated dMC user, but I was on a friend's machine using MusicMatch Jukebox to encode at 320kbps and it was about 5 min. per cd and the quality still seemed to be there, but I'm not sure what mp3 codec MusicMatch uses.

Any ideas/suggestions?

[MasterOfPuppets]

Technically it all depends on how sensitive your hearing is. Personally I can't tell the difference between mp3 at 128 and a CD unless I'm playing it really loud and listening close, while at the same time I know someone who can identify the write speeds of a given burned music CD. Go figure.

Athough, if you're encoding at 320 I would assume quality is an issue. Honestly, I would say you've got to just trust your ears; download a fast mp3 codec and just spend a few minutes testing yourself.

Just want to clear up a few things, does you have a generally older/slower computer than your friend? How long does it take you to convert wave files to mp3 at 320? Around the same time? The speed of encoding is directly related to the speed of the computer, so keep that in mind. On my recently purchased computer the difference in ripping time between MPEG Suite and Lame at high bitrates is less than two minutes (I tend to multitask while ripping so I leave Priority on Normal, I'm guessing it would probably go up to about 4 minutes if I have it on High).

Question to anyone: can CLIs encode at the same time as ripping?

[Spoon]

Not the CLI, a Wave file will be created.

[brandonjp]

Well, it's not really too vital to your survival, but if anyone's interested, I did test different MP3 codecs for speed vs. quality. For my purposes, the race was between dMC's standard LAME MP3 codec and the MPEG Suite 2000 MP3 codec (http://www.dbpoweramp.com/codec-central-mpeg2000.htm). Here's my unnecessarily lengthy report...first, though, I'm using Windows XP on a Pentium III 1Ghz with 512 RAM...

For this important test, I chose only the finest musical specimen...none other than the 4 minute and 39 seconds of bliss in Amy Grant & Vince Gill's stirring duet "House of Love" - track 4 from Amy's 1994 release of the same title on A&M Records (as if I really needed to tell you all, since I'm sure the disc is in your collection too!)
I also tested other random tracks from Johnny Cash's American IV: The Man Comes Around, Fountains of Wayne's Welcome Interstate Managers, Vigilantes of Love's Summershine, and Pedro the Lion's Control. All conversions were made in dMC's CDGrab.exe, ripping each time from the CD to the specified file/format. However, time results were practically the same, the disc being ripped made not much of a difference. ANYWAY, here were my results as far as encoding time is concerned: (sorry all the colums start to slant here!!)
RIPPING TO.........CONVERSION TIME.....FILESIZE
WAVE - 16/44.1............0:16..............48,063 KB
LAME - 128..................2:45..............4,361 KB
LAME - 320..................1:38.............10,901 KB
MPEG2000 - 128............0:45.............4,359 KB
MPEG2000 - 320............0:51.............10,898 KB

Now, here are my personal opinions on the sound quality:
Well, needless to say, but the WAVE sounded best - full range of the wonderful sonic spectrum (especially for Amy Grant!). As for the others...

LAME - 128kbps -
missing a bit of low end, sounds like standard MP3 compression that you're used to.

MPEG2000 - 128kbps -
missing more of low end than LAME, and highs and mids were a bit more lossy and un-crisp, overall a bit thinner, but not as bad as I expected.

LAME - 320kbps -
better range of sound (of course) than lower bitrate, but still a bit loose on the low end, for most non-archival purposes it would work fine (however, I still prefer Monkey's Audio).

MPEG2000 - 320kbps -
quite impressed again, still a bit thin, naturally not as much as lower bitrate, but did sound a bit falsely compensated in the the ends / low and high ends a bit muddy.

This was certainly a non-technical investigation into the codecs. It took realistically about 25 minutes. However, for those interested, I would guess that for most ears the MPEG2000 codec wouldn't lose too much. For me, the faster conversion is worth the quality loss, as my current project is simply archiving several older (usually untouched) discs for a friend's DJ service, which will never actually play about 90% of the MP3's I'm archiving. If you're deciding for yourself, it's easy to test, and for quick comparison I'd recommend listening to each track in 3 different ways: at low volume, high volume and through decent headphones. And switch back and forth between bitrates listening to the same part of the track on each file.

Bottom line...for this quick project I'm finishing and for my personal listening collection I might possibly use the MPEG2000 codec if I'm in a hurry (because when converting 1000's of wav's to mp3, the faster codec is pretty sweet), but will more than likely stick with dMC's included LAME. However, MP3's for any other purpose, I'll for sure stay safe with LAME. (but i'm still a monkey's audio fan at heart!)

sorry if you wasted your time reading this! thanks & good day...
--brandon

[Tomb]

Brandon - Can I ask if you are just archiving CD's? At 320kbs You will obviously not be using them for portables! I am just asking as if you are after achieving quality at a high bitrate then I would suggest using musepack (mpc) instead as this has been proven to be superior to lame when encoding at the high bitrates. I will see if I can dig out the links for you. Give musepack a try at either the quality level 5 (standard) or quality level six (extreme) and see what you think.

On another note I certainly wouldn't use the MPEG2000 codec as it is no longer developed and I personally think it sounds awful. If mp's are a must then you could try the gogo mp3 codec which was based on Lame 3.89 and optimised for speed. As for Musicmatch it users a Fraunhofer codec or the mp3pro one. Both are behind Lame and MPC.

Personally I do find the converter slow when compared to encoding waves with another frontend but I do wonder sometimes what the rush is when people are after speed against quality!

[zevele1]

Brandon - Can I ask if you are just archiving CD's? At 320kbs You will obviously not be using them for portables! I am just asking as if you are after achieving quality at a high bitrate then I would suggest using musepack (mpc) instead as this has been proven to be superior to lame when encoding at the high bitrates. I will see if I can dig out the links for you. Give musepack a try at either the quality level 5 (standard) or quality level six (extreme) and see what you think.

On another note I certainly wouldn't use the MPEG2000 codec as it is no longer developed and I personally think it sounds awful. If mp's are a must then you could try the gogo mp3 codec which was based on Lame 3.89 and optimised for speed. As for Musicmatch it users a Fraunhofer codec or the mp3pro one. Both are behind Lame and MPC.

Personally I do find the converter slow when compared to encoding waves with another frontend but I do wonder sometimes what the rush is when people are after speed against quality!


I'am with Tomb on this one

To me use of mp3 320Kps is non sense...
Why?
Because at same size you can get much better quality

I'am using MPC braindead as my backup format
SUPERB sound , FAST encoding and size more or less like mp3 320.

MPC braindead is the lossy format the closest to lossless format.

Converting it to any other formats give exellent results very very close to original cd converted to this format.

Anyway , up to each to find out...

PS: i use Braindead because of the converting to other formats aspect.

If for listening quality only , no need to use Braindead. Extreme will do

[brandonjp]

no, i've not tried MPC...i think mainly becuase I don't know much about it (as in how to play the files). that's why a fast mp3 codec appealed to me for this project because I the mp3's could be played on whatever music player is used...in the event that the files are ever played. My intention is to archive the discs to either a hard drive or to data DVDs. So space is in some ways a concern, but the quality also is important....and so is easy access

As for the speed issue, it is a bit tedious at times, but normally not a problem when I'm doing stuff for my personal collection. For this project, the faster the codec = the faster I get the job done, so that's why I was simply searching for something fast.

I'll try MPC..........
.......okay, i just converted some of the waves from my test to mpc.....
really really does sound great, but dbpoweramp was the only player i had that could play it.
it's speed however (using standard setting), wasn't much faster than LAME....usually around 10 seconds or so less than to LAME

Well...i just attempted to try the Gogo mp3 codec, but it crashed MusicConverter.exe so....no results on it.

?? -- as of now I'm still using LAME...i've got about 30gb already converted and about 20gb of waves waiting to be converted (but I won't do it just yet), then about...well, too many cds still yet to convert. my question is.....what would be optimal? speed is not really biggest concern...compatibility *blooper*1, then quality, then size, then speed. what is there (or is there anything)like LAME mp3 @ 320kbps that would give better quality, and possibly less filesize, yet still be very compatible across players??

by the way, thanks so much for your help thus far!!
--brandon

[Tomb]

Hi Brandon.

Foobar 2000 plays MPC natively. You can download it from here:

http://www.foobar2000.org/download.html

For Winamp 2x, 3x and 5x use the plug-in from:

http://www.saunalahti.fi/cse/mpc/winamp/index.html
Note that are different plug-ins for 2x (5x) and 3x

Gogo is fast but as I stated earlier is based on an old version of Lame. Try downloading the codecs from http://rarewares.hydrogenaudio.org/mp3.html and use this front end http://home.wanadoo.nl/~w.speek/download/Wav2Gogo.exe with them. You will need to rip files to wav first though.

As for compatibility mp3 is still the most popular codec. However if you are going to archive your files with a lossy codec then again I would say mpc but you will not be able to play them on a DVD as a CDRW etc. I doubt mpc will ever be supported on any hardware players to be honest. One final one you could try is MP4/AAC (same thing just different containers) using the archive setting or audiophile setting if using NERO. The codec Spoon uses is Fast Enc and that is not really the best one to use. Try a demo of NERO or use Itunes which is free. Both are hefty downloads though!

MP4 will eventually replace MP3 and is starting to be supported on other non-pc hardware if what I have read seems to be true! Most players support MP4/AAC nowdays. Again Foobar does natively and I would use this plug-in for Winamp 2x and 5x http://rarewares.hydrogenaudio.org/files/in_mp4.zip.

You could stick with Moneys Audio although I prefer FLAC!

Let us know how you get on.

Tom

[alrakis]

LAME - 320kbps -
better range of sound (of course) than lower bitrate, but still a bit loose on the low end, for most non-archival purposes it would work fine (however, I still prefer Monkey's Audio).

MPEG2000 - 320kbps -
quite impressed again, still a bit thin, naturally not as much as lower bitrate, but did sound a bit falsely compensated in the the ends / low and high ends a bit muddy.
--brandon


:cool: curious if you did some double-blind test with that files...
are you sure you're not affected by placebo effect?
difference between 320kbps mp3 and original wav is sometimes heared by 0.01% of population...

[Link500X]

check this site:
http://www.airwindows.com/encoders/index.html

They have a comparison of a few mp3 codecs. Lame 256kbps is actually better than Lame 320kbps. Blade is best at high bitrates, and seems to skip less than Lame in most cd players.
MPC is definitely the best lossy codec. For lossless I use wma9 lossless. The only time that I use mp3 is for use in my cd player.

[alrakis]

Interesting site.
Bur forgive me: the only way to estabilish the "transparency" of a lossy codec is to do blind listening tests (ABX test), not graphs, nor measurements.
This is the way developers tune the codecs for best quality.
If you can't tell the difference between original wav and compressed file in a blind test (avoiding placebo effect) then it is transparent for you.

You can try yourself... http://www.kikeg.arrakis.es/winabx/
give it a try with a lame --preset standard or --preset extreme mp3, that surely it is not a 320kbps mp3.

No doubt instead that mpc and vorbis are better than mp3 at the same bitrate.

Best regards.

[Link500X]

It is hard to tell the difference at high bitrates. But as I said before, Blade skips less than Lame in my cd player. Plus you can still use Blade for free in dbpowerAMP; after 30 days you must pay for Lame compression.

[neilthecellist]

You know what? I can never understand why the Lame encoder has a liscence fee for encoding mp3s but the Blade encoder doesn't need a liscence.......it just confuses me......

[Link500X]

Blade is probabily free because it is no longer under devolopment. One of the reasons that devolopment stopped on Blade is that the developer did not want to have to pay for distributing the MP3 codec. Some other MP3 encoders might be free also, but I haven't tried any them recently.

[neilthecellist]

Oh I see now.

(Link500X, how come you're not online ever on AIM, MSN or Yahoo! Messenger? I've been trying to contact you for days now!)

Well, LAME is under development, but it's opensource.......does that mean anything?

[LtData]

It means anybody can modify the source. It also NORMALLY means that its free. However, some places charge for compiled versions of their open source software.

[Tomb]

Blade is probabily free because it is no longer under devolopment. One of the reasons that devolopment stopped on Blade is that the developer did not want to have to pay for distributing the MP3 codec. Some other MP3 encoders might be free also, but I haven't tried any them recently.

Blade stopped development because the developer considered Ogg Vorbis a better codec in part due to the fact that it is patent free.

Blade was terrible on anything less than 256 k/bits.

[neilthecellist]

So I should use Blade only if I want to go on a bitrate higher than 256? (Higher that that is only 320?)

[alrakis]

My personal opinion is that you should always use lame for mp3 encodings (very sad that we must pay now for use it in dBpoweramp :cry: ).
Very generally speaking 256kbps (using lame) is an overkill for most people, so a wasting of space.
But again I suggest you to make some tests to find at which bitrate (or preferably VBR quality setting) a lame mp3 is transparent for you, using some killer samples and then encode your mp3s at that quality setting.

[Tomb]

So I should use Blade only if I want to go on a bitrate higher than 256? (Higher that that is only 320?)

No. I was just highlighting what it's benefit was, it performed poorly against other mp3 encoders at the lower bitrates.. It stopped being developed in 2002 whilst lame is still developing. Use Lame!

Agree with alrakis that 256 k/bits is overkill. As for paying for dbpoweramp there is not a lot that could have been done about it because of patent issues. You could use lame.exe and the CLI generic codec without registering but it hardly breaks the bank to pay for it.

[Link500X]

It is usually best to avoid using mp3 format. The quality is terrible compared to other codecs, such as mpc or ogg. If you have the harddrive space than lossless is best. I usually use wma9 lossless because it loads the fastest on my slow computer.

If you use mp3 and want good quality(high bitrate), than blade is best. Anything below 192kbps, you should use lame. For bitrates between 192kbps and 256, it depends on what type of music you listen to. Lame makes the drums sound best; Blade makes the guitar sound best.

(Link500X, how come you're not online ever on AIM, MSN or Yahoo! Messenger? I've been trying to contact you for days now!)
I only go on messenger about once a week, and my status is usually set to invisible. I will try to get on later tonight.

[neilthecellist]

So.....LAME makes drums sound right...Blade does guitar right....how about violins and trumpets in a classical orchestral piece?

[Link500X]

Lame should be used for hard beats, such as drums; Blade should be used for strings, such as guitar or violins. For classical music you should definitely use Blade. However, for music with alot of hard beats, such as hard rock, Lame might be better. For using Lame, you shouldn't go above 256kbps because as shown at http://www.airwindows.com/encoders/index.html Lame 256 is actually better than Lame 320.

[Link500X]

Blade stopped development because the developer considered Ogg Vorbis a better codec in part due to the fact that it is patent free.

Blade was terrible on anything less than 256 k/bits.

There are many reasons that Blade stopped development(Check the Blade homepage). Blade is pretty good at 192 kbps as well.

Well, LAME is under development, but it's opensource...

Blade is also opensource. The sourcecode is now the only thing available on their website. It must be compiled to be of any use. There are many sites with the compiled code available(the dll version is in codec cetral).

[neilthecellist]

Lame should be used for hard beats, such as drums; Blade should be used for strings, such as guitar or violins. For classical music you should definitely use Blade. However, for music with alot of hard beats, such as hard rock, Lame might be better. For using Lame, you shouldn't go above 256kbps because as shown at http://www.airwindows.com/encoders/index.html Lame 256 is actually better than Lame 320.


OK! I'm gonna re-rip everything classical that I have to Blade mp3.

[ChristinaS]

OK! I'm gonna re-rip everything classical that I have to Blade mp3.
Don't waste your time with "everything". Do a test only.
Try wma also, you may like it :)

[alrakis]

The quality is terrible compared to other codecs, such as mpc or ogg

Lame should be used for hard beats, such as drums; Blade should be used for strings, such as guitar or violins. For classical music you should definitely use Blade. However, for music with alot of hard beats, such as hard rock, Lame might be better. For using Lame, you shouldn't go above 256kbps because as shown at http://www.airwindows.com/encoders/index.html Lame 256 is actually better than Lame 320.


Oh my God...
I want not to be unpolite... I simply disagree.
Lossy codecs are all based on perceptive encoding.
There is a point from which going up with the bitrate does not mean going up with perceived quality: this is called transparency because from this point you can't ABX the orginal wav from the compressed file, and it is subjective. Saying that 256kbps is worse than 320kbps, and in general something is better/worse than others basing on graphs and similar things in the lossy codecs world is simply nonsense, because what codecs reach is always perceived quality... for 99.99% 256kbps mp3 is transparent so who could tell 256 is better/worse than 320?
ABX tests are usually made on killer samples, known samples that when encoded produce known perceptible (by trained ears) artifacts.
If a mp3 is encoded at such bitrate (transparency), saying some is better in drums and something else in strings is another nonsense... it implies that such mp3 is NOT transparent for you.

Again, I not would say that mp3 "quality is terrible compared to other codecs".
Any lossy codec could be transparent, simply mpc and ogg usually are transparent to most people at lower bitrates, usually 170-190 kbps (VBR) while mp3 is usually transparent at 200-220kbps (VBR).
You could encode at a higher bitrate, mantaining a sort of "security" for future situations or for transcoding (that is anyway a bad idea between lossy codecs) but for listening pourposes transparency is what you ask to a lossy encoder.

[neilthecellist]

I don't understand this. ALL lossy codecs, with the exception of Blade and FhG at low bitrates, hurt my ears. I hear this noisy sizzle in the background, (is this called an artifact?) and after listening to the lossy file, my ears ring and I have a migraine. And I've done blind testing with my brother. He generates a bunch of random bitrate files in different formats and asks me if it's lossy or lossless...I can tell.....

[alrakis]

excuse... what do you mean by random and low bitrate?
usually for not hear differences in general music you should encode at something like 160 kbps (lame... not blade). If you hear differences try higher bitrates.
But keep attention not to have DSP effects or similar active on your sound board: they produce a horrible sound with lossy coded music.

[neilthecellist]

I hate DSPs anyway. My equalizer in Winamp is set to default, too.

[Link500X]

I don't understand this. ALL lossy codecs, with the exception of Blade and FhG at low bitrates, hurt my ears. I hear this noisy sizzle in the background, (is this called an artifact?) and after listening to the lossy file, my ears ring and I have a migraine. And I've done blind testing with my brother. He generates a bunch of random bitrate files in different formats and asks me if it's lossy or lossless...I can tell.....

Yes, the static is an artifact. It is to be expected at low bitrates. Blade's artifact is a bit different than the other mp3 codecs; at anything below 192kbps it will produce a massive reverb effect as well as make the drums sound bad.

Again, I not would say that mp3 "quality is terrible compared to other codecs".
Any lossy codec could be transparent, simply mpc and ogg usually are transparent to most people at lower bitrates, usually 170-190 kbps (VBR) while mp3 is usually transparent at 200-220kbps (VBR).

OGG is good at low bitrates; MPC is not, it is great at high bitrates. Also, how transparant the artifacts are is affected by the quality of the speakers/headphones. If your speakers only go as low as 40hz, then you won't hear the static as much, but you also won't hear the low notes of the song.

[alrakis]

Sure.
If in the original you hear bass notes and in the coded not, then codec is not transparent for you.
Public tests are usually made with quite high-end equipment, and very well trained ear testers.
For my little tests i use mostly headphones, cause I have very good hi-fi headpones and middle-quality pc speakers. Then i burn CDs and listen to them in my mid-fi stereo system (something like 2000$ to give an idea).
For me and my equipment a lame "--preset standard" coded mp3 (~200kbps VBR) is almost always transparent, except in some very rare cases that become transparent if I use a "--preset extreme" setting (~240kbps VBR). mpc and ogg for me (and I'm not the only one) are transparent at lower bitrates.
But for others it may be different I know...

ogg is considered better at low bitrates (~128kbps or less) because its artifacts are often less annoying although it's surely a not transparency bitrate range.
very last third-party optimized versions of vorbis codec, especially aoTuV, can rival with mpc (surely the reference) at higher bitrates and reach transparency at same level.
And vorbis is in continuous development... I look forward with trust ;)

[Tomb]

Christ! I suggest you all head to Hydrogen Audio (http://www.hydrogenaudio.org/) and tell them how great Blade is................

Link500X - you make a lot of comments about quality without backing your statements up with test results or samples. Talking about using Blade and your comments about mp3 in general are misleading. Where is your evidence that Blade is better at high bitrates than lame. No tests have ever been carried out on this to my knowledge (if they have please post the links) and in every mp3 encoder test, on all types of music, Blade has always finished last. here is one example (http://www.rjamorim.com/test/128extension/results.html).

Pretty pictures, as per your links, do not indicate quality. This has been proven many a time. This is also the work of one individual and not the audio community. See above link for more on why graphs do not prove Audio Quality.

if you have stated that these settings/codecs etc are best for you then fair enough as we are all different and as are our audio setups. But generalising as you have done is misleading and could have an adverse effect on those browsing this forum.

Neil - you continually mention quality again without any evidence to back this up. Again you need to ABX. How do you know your audio set up is not the cause of your migraines?

Apologies if this is getting "Hydorgen Audiolike" but if you make such comments then you need to be able to back them up with test results etc.

[Tomb]

It is hard to tell the difference at high bitrates. But as I said before, Blade skips less than Lame in my cd player. Plus you can still use Blade for free in dbpowerAMP; after 30 days you must pay for Lame compression.

So its the fault of the encoder rather than the hardware's actual decoder?

[Link500X]

here is a test that I did myself: http://gwave.beplaced.com/

[Tomb]

here is a test that I did myself: http://gwave.beplaced.com/

Very nice but graphs do not prove audio quality for the simple reason that it cannot prove you can hear differences or provide any statistics (results of what you claim). What does that image prove to someone who has no idea what it means if they saw it?

See here for why ABX testing (http://www.hydrogenaudio.org/forums/index.php?showtopic=16295) is the way to go.

Carry out some tests but make sure you have some decent headphones!

[Link500X]

Very nice but graphs do not prove audio quality for the simple reason that it cannot prove you can hear differences or provide any statistics (results of what you claim). What does that image prove to someone who has no idea what it means if they saw it?

See here for why ABX testing (http://www.hydrogenaudio.org/forums/index.php?showtopic=16295) is the way to go.

Carry out some tests but make sure you have some decent headphones!
I will do these tests when I get a good pair of headphones. For now all I can do is compare it myself on my subwoofer.
Blade has always finished last. here is one example (http://www.rjamorim.com/test/128extension/results.html).
Of course Blade finished last in those tests. They are all 128kbps. Why didn't they do any high bitrate tests?

[alrakis]

Listening tests at more than 192kbps are pretty unuseful because most codecs cannot be ABXed at such bitrates.
Can you? ;)

[Link500X]

Listening tests at more than 192kbps are pretty unuseful because most codecs cannot be ABXed at such bitrates.
Can you? ;)

I can hear the difference on my subwoofer, when it is turned up loud. Lame still adds static at any bitrate. Blade is almost perfect at 256kbps.

[alrakis]

If I would apply a custom DSP to a 500kbps ogg vorbis I'll probably can ABX it, I don't know... but it isn't the point.
I don't normally listen to music applying a DSP, and so many others I think.
If you normally turn up so loud your subwoofer listening to music then blade could be the best choice for you and to "listen to music with a subwoofer turned up loud". So thanks to have specified that. Now we have learned something...

[Link500X]

If I would apply a custom DSP to a 500kbps ogg vorbis I'll probably can ABX it, I don't know... but it isn't the point.
I don't normally listen to music applying a DSP, and so many others I think.
If you normally turn up so loud your subwoofer listening to music then blade could be the best choice for you and for "listening music with a subwoofer turned up loud". So thanks to have specified that. Now we have learned something...
If by DSP you mean an equalizer, then I don't use one either. It makes the music sound worse, and the media player that I use doesn't have an equalizer.

What's wrong with listening to loud music(godsmack, iron maiden, metallica, bad religion, etc.) over a subwoofer?

[Tomb]

Of course Blade finished last in those tests. They are all 128kbps. Why didn't they do any high bitrate tests?

Because they are almost impossible to carry out because differences at the higher bitrates are harder to distinguish than at the lower bitrates.

Unless you have 50 (or more like 100) people prepared to abx you will not get meaningful results!

I can hear the difference on my subwoofer, when it is turned up loud. Lame still adds static at any bitrate. Blade is almost perfect at 256kbps.

Can you do this on another setup? It's just to prove a point that one set up may well give different results to another. Headphones are still a better bet for this though.

My point in this debate is if what your saying about your choice of codec is good for you and your set-up then that's great. But please don't use that as evidence to say generally blade is better than lame at such and such a bitrate or that mp3 is a codec not worth considering as you have done in your earlier posts.

[Link500X]

Can you do this on another setup? It's just to prove a point that one set up may well give different results to another. Headphones are still a better bet for this though.

My point in this debate is if what your saying about your choice of codec is good for you and your set-up then that's great. But please don't use that as evidence to say generally blade is better than lame at such and such a bitrate or that mp3 is a codec not worth considering as you have done in your earlier posts.
I will test it on more equiptment within the next month. I don't have any money to buy a good pair of headphones.
I also used a technical comparison as shown at http://gwave.beplaced.com/ . A listening test was mentioned, that is why i said that Blade sounds the best to me on my subwoofer. The headphones that I do have aren't very good, so a listening test on them would be worthless. I did not say that a mp3 codec is not worth considering, I said that other codecs(OGG and MPC) are better.

[neilthecellist]

"Stop fighting please! Can't we just have a group hug??!!" -VCPR from Grand Theft Auto: Vice City

I have really really really really really really REALLY expensive speakers that my dad bought for his own personal pleasure. Some really nice PSB Platinum speakers, paired with an L70 DVD-Receiver. Upstairs, we have some Vandersteen Tower Speakers paired along with a PS Audio pre-amplifier and a B&K Amplifier. Pretty audiophile/hydrogen audio standard-ish if you ask me...

ANd yes, I can tell you, I think i've decided. FhG would be my pick, I now despise LAME. There's SOOO much static in the background.

Actually, to be honest, I hate all lossy codecs. I only use these Lossy codecs because my hard drive is so tiny. (30 GB, UT2004 and Halo take up lots of space)

Can anyone tell me why some people are able to tell , say, for example, that LAME sucks and others can't hear the artifacts? I've come to notice that Asians at my school seem to be able to hear these static noises, whereas the caucasians and Mexicans can't .....hear them. I'm not trying to sound racist or anything, but I think there may be some different hearing abilities among each ethnicity.

[Tomb]

Whose fighting Neil? it's an open and lively debate! Myself and Alrakis are just questioning genral comments made by Link500X and yourself. It's that's fighting I'm the new pope.

I have given up on explaining why you shouldn't make claims regarding quality etc without evidence as you have done so again in your post. It's pointless.

By the way have you done an ABX test on your Asian friends compared to Caucasians regarding your last comment?

That one was a joke by the way............

[Spoon]

I hope you have other speakers with that subwoofer, some have very low frequency responces.

A few facts about audio:

You cannot judge a lossy audio codec on a frequecy graph, I could write a codec that gave a perfect freqeucy responce graph but would sound rubbish if you listened, lossy audio is all about artifacts and pre-echo (ringing etc).

To rate audio codecs the best way is to compare like with like (ie same bitrates, or filesizes) using a blind abx, if you know what codec you are listening to then you are influencing the test. ABX against the original, not the two compressed.

[Link500X]

Yes, I do have other speakers connected to my subwoofer. But they are not nearly as grand as the speaker systems that everybody else in this forum has. My system is a simple gaming subwoofer and the stereo speakers that came with it. I don't have the money for expensive systems.
I don't think that race has anything to do with one's ability to hear the artifacts. I can hear the static that Lame produces, but not the reverb, overring, pre-echo, or whatever else you might call it(it is all the same) that Blade produces.

[neilthecellist]

Well then, you can't judge on how good lossy audio sounds, because I'm assuming you have Logitech 2.1 set, which comes with the computer. Trust me, Link500X, computer speakers suck. SUCK SUCK SUCK. Like, seriously, they SUCK. (Forgive me if I assumed the wrong speaker set; correct me then)

As for ABX testing, have you ever been to one's house and tried it? If you have the time, go for it.

[neilthecellist]

Oh and also, I've done ABX tests, but I don't have a scanner, I don't know how to take screenshots, so if someone tells me how to do that, I'll be able post "evidence".

[Tomb]

As for ABX testing, have you ever been to one's house and tried it? If you have the time, go for it.

You don't need other people to ABX - just software such as WINABX and samples i.e original wav, mp3 at 128 k/bits then converted to wav and so on.. Oh and a good pair of headphones.

[ChristinaS]

Oh and also, I've done ABX tests, but I don't have a scanner, I don't know how to take screenshots, so if someone tells me how to do that, I'll be able post "evidence".
Whenever you have somehting on your screen you want to take a screenshot of, just hit ALT + Print Screen.

Then open any imaging program like Paint and click on Edit > Paste and it will paste in the screenshot. Save it as jpg or whatever you want (don't forget to see where it goes, usually it's MY Documents > My Pictures) . Remember bmp is very large, it is to images what wav is to audio so save as jpg or gif.

[neilthecellist]

but but but.....ABX testing on the computer is pointless and sounds like crap anyways on even the BEST headphones out there...more accurate test results can be obtained from performing the on kickarse audio systems like the L70 DVD-Receiver.

What I do for testing is I get my brother to take an original WAV file, (uncompressed PCM), and convert it to all these random lossy codecs, then convert all those lossy files back to WAV and then burn them to audio CD and then plays them on the L70 system with the PSB Platinum speakers. From then, i can test with friends(or by myself from now on) to see which format sounds better over the other. And we compare like

1. Listen to lossy
2. Listen to original
3. Compare.
4. Listen to new lossy
5. listen to original
6. listen to the first lossy
7. listen to the 2nd lossy again
8. compare to original
9. Choose a 3rd one and then repeat steps 4-8

[alrakis]

Whenever you have somehting on your screen you want to take a screenshot of, just hit ALT + Print Screen.

Then open any imaging program like Paint and click on Edit > Paste and it will paste in the screenshot. Save it as jpg or whatever you want (don't forget to see where it goes, usually it's MY Documents > My Pictures) . Remember bmp is very large, it is to images what wav is to audio so save as jpg or gif.

It's not necessary: winabx creates automatically a log file in pure text.

[alrakis]

...convert it to all these random lossy codecs...

bitrates?

[neilthecellist]

Whenever you have somehting on your screen you want to take a screenshot of, just hit ALT + Print Screen.

Then open any imaging program like Paint and click on Edit > Paste and it will paste in the screenshot. Save it as jpg or whatever you want (don't forget to see where it goes, usually it's MY Documents > My Pictures) . Remember bmp is very large, it is to images what wav is to audio so save as jpg or gif.
thanks

[neilthecellist]

bitrates?

We do 64, 128, 160, 192, and 320. Obviously, 64 is easy, so my brother and i stopped using it. now it's just 128, 160, (178 for ogg vorbis when comparing to 160) , 192, and 320.

[alrakis]

So FhG outperformed Lame in that test? Cool. I think I'm going to reencode all my collection now.
I suggest however to give a serious try to winabx...

[adaywayne]

You cannot judge a lossy audio codec on a frequecy graph
~~~~~~~~~~
Except to prove a "negative", such as a cutoff at, say, 8-10 MHz by mp3 at 64Kbps.
Arnie

[Link500X]

Well then, you can't judge on how good lossy audio sounds, because I'm assuming you have Logitech 2.1 set, which comes with the computer. Trust me, Link500X, computer speakers suck. SUCK SUCK SUCK. Like, seriously, they SUCK. (Forgive me if I assumed the wrong speaker set; correct me then)

As for ABX testing, have you ever been to one's house and tried it? If you have the time, go for it.

The speaker that I have did not come with my computer. I bought them at EB Games. The brand is Century Concept Digital. They are for gaming, both console and computer. They are much better than the speakers that came with the computer.

[neilthecellist]

Ok sorry for the bad assumption.

Except to prove a "negative", such as a cutoff at, say, 8-10 MHz by mp3 at 64Kbps.
Arnie
Who cares? Who uses 64 kbps anyway? ^_^

[Tomb]

So FhG outperformed Lame in that test? Cool. I think I'm going to reencode all my collection now.

Why?

While I don't doubt Neil's word (and he is lucky to have someone buy expensive stereo equipment for him - most people don't) I still see no evidence of what he says and that other tests carried out by the audio community proves otherwise.

If his comments were that blade or Cosmo MPEG 2000 outperformed Lame would you take his word for it?

[Tomb]

but but but.....ABX testing on the computer is pointless and sounds like crap anyways on even the BEST headphones out there...more accurate test results can be obtained from performing the on kickarse audio systems like the L70 DVD-Receiver.

For the simple reasons that:

Most people have computers and not all have a L70 DVD Receiver :) and;
Most people listen to their digital collection via a computer and not via a stereo.

Again you have made a comment without evidence to back it up. How do you know that ABX testing is pointless and sounds crap on even the best headphones? It's just a comment you have made! I hope you do not write school essays saying for instance that Japan didn't bomb Pearl Harbour or the Olympic games were held on the moon in 2000 etc without backing that statement up. Can you see what I'm driving at here?

[alrakis]

Why?

While I don't doubt Neil's word (and he is lucky to have someone buy expensive stereo equipment for him - most people don't) I still see no evidence of what he says and that other tests carried out by the audio community proves otherwise.

If his comments were that blade or Cosmo MPEG 2000 outperformed Lame would you take his word for it?

Don't worry... It was a joke ;)
At this point of debate I was thinking also to say that for me 64kbps blade mp3 is the best in the world and that all other codecs sucks, because it's so, I hear so, and I needn't ABX or I can't do it or perhaps I'll do it in the future...

[neilthecellist]

Don't trust anything I say. I'm a very naive 15 year old.

[Tomb]

Don't worry... It was a joke ;)
At this point of debate I was thinking also to say that for me 64kbps blade mp3 is the best in the world and that all other codecs sucks, because it's so, I hear so, and I needn't ABX or I can't do it or perhaps I'll do it in the future...

And it wasn't April 1st!

Don't trust anything I say. I'm a very naive 15 year old..

That's not the point mate. It does not matter how old you are. My point which you are fail to comprehend, which may be due to your age, is that you cannot make claims without substansive evidence, proof or facts to back them up. You cannot say for example Lyon is the capital of France without proof. Your word therefore is meaningless. It's a shame that you don't understand what I am trying to say especially as you respond to so many threads, and in a mostly valid way, on this forum.

It states on the front of dbpoweramps main page that " Program quality and ultimately audio quality is paramount, we will not settle for anything but perfection, you shouldn't either!" Your statements, without any facts to back them, up break this ideal.

I'm done with this thread. I may as well hit my head against a brick wall............

[neilthecellist]

Ok ok...so I figured out how to make screenshots.....Even with the ability to create screenshots, how do I get my proof in? Posting mini-sound clips?

[LtData]

Alright, it seems that this thread is getting a little out of control. While there may be claims made, they are opinions until backed up with facts. Also, every persons opinion is different. Graphs or no, each person has to decide for themselves how a codec sounds at different bitrates. Personal listening tests with varied tracks should be how each person decides what codec and bitrate they wish to use for their personal collection.

[neilthecellist]

Ok, I'll bear that in mind. So it's all opinion, it CAN'T be supported by proof?

[alrakis]

ya...

[neilthecellist]

ah ok.

Abandon thread, abandon thread!!

[alrakis]

ya...

this was to tomb: i'm done too...

[Tomb]

Alright, it seems that this thread is getting a little out of control. While there may be claims made, they are opinions until backed up with facts. Also, every persons opinion is different. Graphs or no, each person has to decide for themselves how a codec sounds at different bitrates. Personal listening tests with varied tracks should be how each person decides what codec and bitrate they wish to use for their personal collection.

Exactly hence my statements that one person cannot say that A sounds shit compared to B without proof.

I don't care if Neil said that This Codec sounds better to That One to him. However to state this generally as an accepted statement, without facts, as two of the people have claimed in this thread then i'm afraid that this has to be challenged

Why is this thread out of control Lt Data? Can't anyone have a mature debate these days?

[Tomb]

this was to tomb: i'm done too...

Pass me the white flag............

[Link500X]

Ok ok...so I figured out how to make screenshots.....Even with the ability to create screenshots, how do I get my proof in? Posting mini-sound clips?
You can make a separate webpage like I did for my graphs. http://freewebspace.net/ has a long list of free hosts.

Tomb, why don't you do a test yourself?

I don't care if Neil said that This Codec sounds better to That One to him. However to state this generally as an accepted statement, without facts, as two of the people have claimed in this thread then i'm afraid that this has to be challenged
I already provided my proof, in the form of graphs.

[alrakis]

Pass me the white flag............

we both need a bed sheet...

[neilthecellist]

Ok, if we're going back to Speech and Debate in high school....

Even with the power of screenshots, HOW exactly do I prove that one codec is better over another? Survey at school?

[Tomb]

Tomb, why don't you do a test yourself?

I have - I have taken part in the testing carried out by the Audio Community. Test results here (http://www.rjamorim.com/test/). Do you honestly think that I would raise these issues without carrying out tests myself?

My last tests compared 128 k/bits and preset medium k/bits mp3 against preset standard mp3. I will see if I still have the log files.

I already provided my proof, in the form of graphs.

As stated graphs do not prove Audio Quality. Spoon highlights this earlier in this thread.

we both need a bed sheet...

:)

Lets give it up mate.

[Link500X]

The proof has already been provided. I provided graphs. Neil and myself both say that Lame sounds bad compared to other codecs like Blade and fhg. What other proof is needed(both technical and sound tests were made)? Tomb just won't accept that Blade is better than Lame at high bitrates.

Graphs are good proof. The idea behind lossy encoding is to get the sound as close to the original as possible. My graphs prove that Blade gets closer to the original. Lame adds artifacts that may sound good to some people, but it alters the song by doing this and therefore is not keeping the sound as close to the original as possible.

[Tomb]

Tomb just won't accept that Blade is better than Lame at high bitrates.

Show and provide me with the samples that you used, the bitrates used and the results i.e. the number of tracks where you picked A (blade) was better than B (Lame). Show me the number of samples where the probability was that you were guessing. Show me the different genres you used (Rock/Classical/Indie/Pop etc). Show me the problem samples, freely avaialble on the web, that you used.

I say that you cannot provide proof hence my original statements stand.

I will tell you what how about we run our own tests? I've not run tests for multiple users before but there is always a first.

[alrakis]

Graphs are good proof.

No.

The idea behind lossy encoding is to get the sound as close to the original as possible.

Sound, not signal.

[LtData]

Why is this thread out of control Lt Data? Can't anyone have a mature debate these days?
Yes, you are allowed to debate. However, the debating is starting to get a little personal. And your starting to talk in circles.

And graphs may be evidence of what the sound spectrum looks like, but again, how it sounds is all opinion and is different for everyone. Again, it all depends on opinions. If you think X encoder sounds better than Y encoder, then use X encoder and say why you use it: YOU think it sounds better than Y encoder. Not that X is better at encoding than Y. Not that the graphs for X look prettier than Y's. Not that X is easier to pronounce than Y. The reason you picked X is how YOU think it sounds. This is why we have more than one mp3 encoder. Everyone differs on their prefered sound.

[Tomb]

Yes, you are allowed to debate. However, the debating is starting to get a little personal. And your starting to talk in circles.

And graphs may be evidence of what the sound spectrum looks like, but again, how it sounds is all opinion and is different for everyone. Again, it all depends on opinions. If you think X encoder sounds better than Y encoder, then use X encoder and say why you use it: YOU think it sounds better than Y encoder. Not that X is better at encoding than Y. Not that the graphs for X look prettier than Y's. Not that X is easier to pronounce than Y. The reason you picked X is how YOU think it sounds. This is why we have more than one mp3 encoder. Everyone differs on their prefered sound.

That's it - their own personal sound. As I have said more than once. I HAVE no problem with that. But when Neil and Link500X make their statements that one codec is better than another in more than one thread, without proof but based on their own personal preferences, then I take exception. Is that too hard to understand?

As for circles it's because it's like hitting my head against a wall. My simple point about proof/facts and abx'ing is not being understood. maybe your comment about graphs may solve this.

As for being personal I have just re-read this thread and no one has made a single personal comment against another. If you want to point out where that occurs then fine and we can address it.

If trying to get my point across is offensive then I apologise. I have helped out on this and other forums (including currently moderating forums for another CD Ripper) for many a year. Maybe it's time to stand out of the way of facts and let speculation rule in it's place.

[ChristinaS]

Gentlemen! Time out!

There are enough codecs and settings out there to please everybody :D

Sharp hearing, dull hearing, wax buildup, something for everybody.

For those who prefer to look at graphs, you got those too. Those who hug the subwoofer, be my guests.

Personally I prefer to hear sound rather than watch it or feel it, but that's just me :D

This is like the diffrence between Coke and Pepsi. I really prefer Sprite :p

[neilthecellist]

And I like Mountain Dew. Sue me! I made an opinion

[Tomb]

And I like Mountain Dew. Sue me! I made an opinion

Opinions are different from factual evidence - that's been the point of this thread. That's my last comment on the matter. Full stop.

[Spoon]

I am closing this thread because it is going nowhere (6 pages).

Personal opinions are fine, but if you have made a statement of quality (such as Ogg Vorbis @ 128Kbps is better than WMA @ 128Kbps) you need to point to blind ABX test result (where people, or a person) have listened and rated the audio quality (that they have listened to). Saying that 'I think WMA @ 128 Kbps is better than Ogg Vorbis @ 128 Kbps' is ok as it is a statement of personal preference, but 'WMA @ 128 Kbps is better than Ogg Vorbis @ 128 Kbps' need backing up (if people ask for it to be quantified).

Audio graphs are a no-no (yes they can show a cut off) but it is like describing (by words) the quality of a picture, just wrong a picture should be looked at, audio listened to.