android
  #21  
Old 01-30-2012, 05:08 PM
Enigmatic Enigmatic is offline
Senior Member
 
Join Date: Mar 2008
Posts: 734
Default

Quote:
Originally Posted by DSperber View Post
Years ago I tried to use this program, but the absence of documentation teaching me how to use it was frustrating.
• You need an uncompressed WAV and an MP3 of this uncompressed WAV. A FLAC from the uncompressed WAV is also OK instead of using an uncompressed WAV.
• foobar2000 --> File --> Add Files --> select the WAV and MP3 that you want to ABX --> Open --> select the two files that you want to ABX in foobar2000 --> right click --> Utils --> ABX Two Tracks... --> select whether you want to use DSP --> OK --> select “Keep playback position when changing track” --> play A and B to get a good idea of how they sound --> select “Play X” or “Play Y” --> compare with A and B --> select one of “X is A, Y is B” or “Y is A, X is B” --> Next Trial --> select “Play X” or “Play Y” --> compare with A and B --> select one of “X is A, Y is B” or “Y is A, X is B” --> etc.
• If “Probability that you were guessing” reaches 5% or lower, you probably can hear a difference.
• If “Probability that you were guessing” does not reach 5% or lower, you were probably guessing and therefore cannot hear a difference.
• The above instructions worked for foobar2000 v0.9.6.3.
• I recommend that you use headphones, Exact Audio Copy or something similar, LAME 3.99, and quality settings –V0 to –V9.
Quote:
If you can point me to REAL DOCUMENTATION on exactly what the user is supposed to do, how the snippets to be compared are to be legitimately selected (since isolated snippets can probably be found to produce any one desired result vs. another), how long snippets should be (since being too short or too long will interfere with the mind's ability to distinguish), etc. etc., I will give it another shot.
There might be no documentation on exactly what the user is supposed to do when using foobar2000’s ABX comparator. So please give it a try using my instructions above. Music can indeed be found to trip up the encoder.

But that is not the point. The point is to give you a general idea of the encoder’s effectiveness or ineffectiveness. The snippets can be of any duration that you feel will help you best to hear a difference. What is the etc.? Please be as specific as possible.
Quote:
Just because LAME does an excellent job of producing high-quality MP3 that is very difficult or impossible for most people to distinguish from the original WAV (or equivalent FLAC) does not mean it is BETTER or that I want to use lossy MP3 given a reasonable lossless alternative just because it takes up less file space.
But then we have to agree on definitions for better and reasonable.
Quote:
MP3 is essentially a "fake counterfeit, albeit a good one" of the original.
Counterfeit? I prefer MP3 to be described as a way to discard information but still end up sounding excellent.
Quote:
But I am sure if you were given a "magic music server" that could hold all of your physical CD's in original physical form including all original or equivalent WAV files, but would fit in the palm of your hand and also supported metadata/tags, that you would immediately accept that magic device.
Yes. But these, say, 1-TB, small, inexpensive, and flash-memory digital-audio players are definitely not going to be available anytime soon.
Quote:
For me, that's what FLAC buys. And I like it.
For individuals that want a perfect copy, FLAC is an excellent choice. For individuals, however, that want excellent sound quality and small size, MP3 is often a better choice than FLAC.
Quote:
I don't want to buy or wear a Rolex knockoff even though nobody would notice the difference.
You should not regard MP3s as counterfeits of uncompressed WAVs.
Quote:
Of course it has an effective average bitrate which is not the maximum 320kbps.
It is actually a quality setting.
Quote:
Perhaps it would have been more precise to call it ">320kbps VBR" or just left it out entirely since I specified -V0.
It is best to just type –V0.
Quote:
And given a free choice with no cost and no downside, you'd choose the original (lossless, by definition) WAV/CD/FLAC vs. anything lossy that was "derived from the original".
But life is not that simple. There are always costs and downsides. I understand what you are saying about The Ideal Digital-Audio Player. But we are still very far from that day.
Reply With Quote

Advertisement [Remove Advertisement]

  #22  
Old 01-30-2012, 05:23 PM
saratoga saratoga is offline
Rockbox Developer / Moderator
 
Join Date: Apr 2007
Posts: 3,628
Default

Quote:
Originally Posted by Enigmatic View Post
• You need an uncompressed WAV and an MP3 of this uncompressed WAV. A FLAC from the uncompressed WAV is also OK instead of using an uncompressed WAV.
FWIW, its even easier if you use foobar. Just give it the FLAC, then use foobar to covert it to MP3. Then select the MP3 and FLAC and choose "ABX Two Tracks". It'll automatically convert them and ensure that everything is done gaplessly so that the file lengths line up exactly.

Once you've done it'll compute the statistics for you. One warning though, you should hit the "hide results" button. Seeing how good or bad you're doing can really mess with your head and throw off your concentration. It'll work much better if you don't see the numbers since you won't bias yourself.
__________________
Interested in Google's Summer of Code ? PM me.
Reply With Quote

  #23  
Old 01-30-2012, 06:58 PM
DSperber DSperber is offline
Senior Member
 
Join Date: Jun 2010
Location: Marina Del Rey, CA
Posts: 698
Default

Quote:
Originally Posted by Enigmatic View Post
It is actually a quality setting.It is best to just type –V0.
I understand I can restrict the lower and upper limits of the bit rates if I wanted to. But why would I be doing that?? I wouldn't.

As you know, if you just code -V0 then you get the full range that LAME comes up with, going up to the maximum possible of 320kbps. If I had already specified -V0 I would obviously have no reason to just downgrade what I'd opted for (which is "best possible") to something less than that. That would be pointless and inconsistent. And I did NOT do that.

In fact, all of my 5500 LAME-produced MP3 files were produced with nothing but -V0.
Reply With Quote

  #24  
Old 01-30-2012, 07:14 PM
Enigmatic Enigmatic is offline
Senior Member
 
Join Date: Mar 2008
Posts: 734
Default

Quote:
Originally Posted by DSperber View Post
I understand I can restrict the lower and upper limits of the bit rates if I wanted to. But why would I be doing that?
No one is saying that is what you are doing.
Reply With Quote

  #25  
Old 01-30-2012, 07:19 PM
DSperber DSperber is offline
Senior Member
 
Join Date: Jun 2010
Location: Marina Del Rey, CA
Posts: 698
Default

Quote:
Originally Posted by saratoga View Post
That said, I realize you're a new user and you don't know any better.
I honestly don't know where you got that impression, and why you continue to use such demeaning and insulting and belittling language when we are trying to have a many-times battled but still objective discussion about this very sensitive subject to MP3 afficionados and bigots. I didn't think name-calling was proper approved forum etiquette.

As it turns out I have been a longtime (almost 2 years, not as long as you obviously) member of this particular forum, and my experience creating MP3 goes back to 1998 when I started using the Fraunhofer command-line (DOS) "professional" encoder under Win98. I used to boot to Win98 even when using WinXP normally, just so I could continue to run the Fraunhofer encoder to build my MP3's. The Fraunhofer DOS encoder would not run under the WinXP command-prompt window and only could produce CBR versions of MP3 files. I created 192kbps CBR.

I finally went to an early 3.97(?) version of LAME when I first discovered it, sometime in the mid-2000's. It was like a miracle, actually, and I had no complaint... finally being able to ditch Win98 completely. LAME was (a) WinXP compliant, (b) invokable as an "external encoder" from Audiograbber, (c) easily 100 or more times faster than Fraunhofer's DOS encoder so that typically only 5-15 seconds were required to encode a single MP3, and (d) produced significantly better sounding results as compared to Fraunhofer's 192kbps CBR using the -V0 VBR encoding parameter I'd selected. It was like a miracle.

I have been making MP3 files for the past 13 years, and FLAC files for the past 2 years. I am quite familiar with EAC vs. Audiograbber (and prefer to continue using Audiograbber because I greatly prefer its user interface and the superior completeness with which it creates tags), and I am certainly familiar with the effects the command-line parameters available for LAME. I only use "-q0 -V0 -ms", for "best possible". Audiograbber can invoke both LAME for MP3 and FLAC for FLAC, and that's what I do.

So in my own self-defense, I am not "a new user, and don't know any better". A discusson of FLAC/WAV/original vs. MP3 transform should remain just that.
Reply With Quote

  #26  
Old 01-30-2012, 07:30 PM
Enigmatic Enigmatic is offline
Senior Member
 
Join Date: Mar 2008
Posts: 734
Default

Quote:
Originally Posted by DSperber View Post
I only use "-q0 -V0 -ms", for "best possible".
It is actually simpler than that now. http://wiki.hydrogenaudio.org/index....te.29_settings.
Reply With Quote

  #27  
Old 01-30-2012, 08:31 PM
Lagoo56's Avatar
Lagoo56 Lagoo56 is offline
Senior Member
 
Join Date: Dec 2008
Location: SI,NY
Posts: 737
Default

Who needs FLAC on a DAP when there is 400kbps AAC? There is a 99.9999999999999999999999999999999999999999% chance that a person would NOT be able to tell the difference,even more than 320CBR VS FLAC.

I don't encode all my music at 400kbps AAC but my favorite albums that I want to sound their best I convert to this. They are half the size of FLAC files and sound identical to me on both my DAPs and also my laptop.
__________________
DAPs I own: Rockboxed Sansa C250 2GB+64GB MicroSDXC
Tablets I own:Rooted 8GB Samsung Galaxy Tab 2 7.0 running Sungsonic 2 HD JB 4.1.2
Rooted 80GB Asus Memopad HD 7

Last edited by Lagoo56; 01-30-2012 at 08:42 PM.
Reply With Quote

  #28  
Old 01-30-2012, 08:52 PM
DSperber DSperber is offline
Senior Member
 
Join Date: Jun 2010
Location: Marina Del Rey, CA
Posts: 698
Default

Quote:
Originally Posted by Enigmatic View Post
It is actually simpler than that now. http://wiki.hydrogenaudio.org/index....te.29_settings.
According to the official LAME documentation, with my other -V0 and -q0 parameters the -ms parameter is actually the default. So I genuinely could remove it although its presence does no harm. I think I've been using it for a long time, and perhaps "joint stereo" may have been an earlier default. Maybe it was just a carryover from my early days with the Fraunhofer DOS command-line encoder, when it needed to be specified for the 192kbps CBR encoding I was doing. Can't remember now. But certainly there is no harm or negative impact on the MP3 sound quality by explicitly requesting stereo via -ms to LAME.

Since -q5 is the default, obviously -q0 is superior... in fact "the best". The extra few seconds it might take vs. -q2 is of zero concern to me. I want "the best sounding MP3", no matter what.

Since -V4 is the default, obviously -V0 is superior... in fact "the best". Again, this is what I want.

So you're right, I could shorten that to -q0 -V0 and get identical results... in fact "the best".
Reply With Quote

  #29  
Old 01-30-2012, 09:07 PM
DSperber DSperber is offline
Senior Member
 
Join Date: Jun 2010
Location: Marina Del Rey, CA
Posts: 698
Default

Quote:
Originally Posted by Lagoo56 View Post
Who needs FLAC on a DAP when there is 400kbps AAC? There is a 99.9999999999999999999999999999999999999999% chance that a person would NOT be able to tell the difference,even more than 320CBR VS FLAC.

I don't encode all my music at 400kbps AAC but my favorite albums that I want to sound their best I convert to this. They are half the size of FLAC files and sound identical to me on both my DAPs and also my laptop.
Unfortunately, this is 100% irrelevant in a J3 discussion since the J3 does not support AAC.

I'm sure you would also be satisfied with a "perfect duplicate fake/counterfeit" of the Mona Lisa. After all, it's "indistinguishable from the original" and came to you at a very good price. Can't tell the difference, so might as well go for the "smaller price".
Reply With Quote

  #30  
Old 01-30-2012, 10:14 PM
Enigmatic Enigmatic is offline
Senior Member
 
Join Date: Mar 2008
Posts: 734
Default

Quote:
Originally Posted by DSperber View Post
So you're right, I could shorten that to -q0 -V0 and get identical results... in fact "the best".
Actually, you can shorten that to –V0. Please see the documentation that came with your copy of LAME 3.99. The default is –q 3. For the default VBR mode since LAME 3.98, -q 0 to –q 4 uses the best algorithm.
Reply With Quote

  #31  
Old 01-30-2012, 10:38 PM
DSperber DSperber is offline
Senior Member
 
Join Date: Jun 2010
Location: Marina Del Rey, CA
Posts: 698
Default

Quote:
Originally Posted by Enigmatic View Post
Actually, you can shorten that to –V0. Please see the documentation that came with your copy of LAME 3.99. The default is –q 3. For the default VBR mode since LAME 3.98, -q 0 to –q 4 uses the best algorithm.
You're right.

You'd think they would update their "official documentation" page on the web site to match what they've now released in its packaged HTML documentation. Oh well.

So... -V0, and you've got "the best". No argument here (not that -q0 -V0 -ms is any different, soundwise or bitwise).
Reply With Quote

  #32  
Old 01-30-2012, 10:42 PM
El C's Avatar
El C El C is offline
Member
 
Join Date: Dec 2011
Location: Canada
Posts: 366
Default

Have any ABX tests actually been done?
__________________
Desktop: Objective2 > Audio Technica A900
Portable: Cowon D2+ > Headstage Arrow 3G > Phonak PFE 112 (grey filters + Comply tips)
Reply With Quote

  #33  
Old 01-30-2012, 11:05 PM
DSperber DSperber is offline
Senior Member
 
Join Date: Jun 2010
Location: Marina Del Rey, CA
Posts: 698
Default

Quote:
Originally Posted by El C View Post
Have any ABX tests actually been done?
Not yet, but I will. Promise.

More important to me right now is the "internal guts swap" I must do to relocate a machine/motherboard that will not support a new dual-DVI ATI PCIe 2.1 video card (HD5770) with a second machine/motherboard which will support this new video card. This swap will take two days (retaining existing CPU/motherboard and primary hard drive in each machine, and reversing all other peripheral cards and components between the two machines).

I need dual-DVI at one location where the incompatible motherboard machine currently lives, and my solution is to use the other motherboard machine instead.

So quite frankly I have a higher priority demand for my time right now, and I want to spend adquate time on the ABX experiment. Of course I know it will not cause me to delete one FLAC file that I have and replace it with one MP3 file, but I will perform the experiment nevertheless.

Again... you can't be better than the original WAV/CD. And FLAC is a duplicate of that. Anything else is not the same, even if it's a very very very good imitation. Even if that's an emotional argument rather than a "facts-based conclusion from trying an ABX test", you will never convince me that opting for a lossy anything can possibly be a better choice than going with the original WAV/FLAC version... if file space is not a consideration.

That is truly the only argument for choosing MP3, is that it is smaller. There is nothing else which can be claimed for it as an advantage over the original WAV/FLAC. It cannot possibly be "better" than the original WAV/FLAC, although it may be "indistinguishable".

And if it's indistinguishable, then what does it matter to anybody whether I just choose FLAC over MP3 if file size is not of concern to me? It's not like I've made the "wrong" choice and FLAC sounds "worse" than MP3.

Anyway, I definitely will give Foobar and the ABX test another shot. I just was unhappy with not knowing how many comparisons I was expected to perform, how to proceed through the whole process, etc. I mean I'm not incapable of figuring it out by trial and error since the button annotations were understandable, but I just didn't know what all of those posted "results" were that I saw from others. How long did they have to do a test before getting those "results"? And what did they mean... all zeros???

How come everybody got all zeros? What does that mean??? What are they trying to say by posting all zeros?? Did they "win or lose"? Does that mean they had achieved the minimum acceptable encoding parameters so that they could have the smallest MP3 file that was indistinguishable from the original WAV? If so, then it's a total waste of time for me anyway because I'm still going with FLAC. I'm not looking for a smallest-possible MP3 that "fools me and most people, as if it were the WAV" as my goal.

I appreciate attempts here to satisfy my request for instructions, but I still think it would be appropriate for the program to provide formal documentation on how this whole functionality is to be set up and used.
Reply With Quote

Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump



All times are GMT -5. The time now is 07:59 AM.