[HeadFiend]

That means, you couldn't tell the difference.

 

[Leslie Dorner]

I got my AKG K 702 with the Cardas 10' Fat Pipe cable yesterday. It's got the same essence of my Etymotic ER-4PT with the 4S adapter, but the sound stage dimensions are much larger and the treble does not roll off at 16 kHz. These are extremely high resolution and detail retrieval is second to none in its class. They are extremely neutral and accurate.
 
Regarding my MP3 collection, I have to eat my words and I have to say that I can now tell the clear differences between 320 Kbps MP3s and FLAC when comparing the same bit depth and sampling frequency. FLAC has a crisper sound that is more accurate, but it's not a decisive win for FLAC or other loss less codecs. FLAC and WAV sound crisper and clearer with no audible artifacts or limitations imposed by the ageing MP3 codec. For high resolution 24 bit 192 kHz FLAC files, it's simply no contest: FLAC sounds much better than MP3 at high resolutions period.
 
I'm still keeping my very large MP3 library because it's extremely convenient and it's a universal format. I think that I'm going to restore my huge FLAC library from my CrashPlan+ account this weekend.
 
MP3 is showing its age. LAME 3.9.9 --preset-insane 320 Kbps MP3 does have minor yet audible artifacts and it does roll off the treble at 16 kHz. For the majority of music, this is not so much of a problem with ordinary audio components.
 
I have an exceptional home audio system and desktop PC. I'm also the lead singer in my church choir so I have to understand pitch and accuracy. I think that the MP3 codec is not going away anytime soon, but I now strongly prefer Vorbis especially using the Ogg container. To my ears, -v10 at 500 Kbps bit rate Vorbis is indistinguishable from WAV or FLAC loss less codecs. The file sizes are still huge when encoding using these loss less codecs or at the top bit rate for Vorbis, but that's life. This is why I have 18 TB of USB 3 external hard disk drives.
 
I have a mint pair of Wilson Audio Alexandria XLF floor standing loudspeakers and MP3 sounds compressed and lossy through them. They sound compressed and lossy through my AKG K702 with the Cardas Fat Pipe cable. It's just that they are extremely convenience and the file sizes are bearable especially for portable gear that makes it very difficult for me to switch back to FLAC.
 
C'est la vie!

 

[cel4145]

MP3 is showing its age. LAME 3.9.9 --preset-insane 320 Kbps MP3 does have minor yet audible artifacts and it does roll off the treble at 16 kHz. For the majority of music, this is not so much of a problem with ordinary audio components.

Have you seen this new research: Human hearing beats sound’s uncertainty limit, makes MP3s sound worse. Good chance we'll see some upgrading of encoders in the future to take that into account. It also helps explain why you can hear the difference even though the encoder is doing a very good job of compressing the music.

 

[Leslie Dorner]

This is why I have a full backup of my FLAC library in my CrashPlan+ account. My understanding is that upgrading encoders by replacing linear with non-linear algorithms is not being actively pursued at this time and it will most likely lead to a new audio codec that has larger file sizes if we're talking about loss less compression being implemented. You're re-arranging how the bits of audio are encoded by fundamentally altering the classic linear model with a newer non-linear model which has the possibility of capturing more audible bits of data. It would be a revolution in the industry. Whether it supplants the old fashioned and traditional codecs and formats like MP3 and CDs will remain to be seen.
 
The biggest problem that I have with FLAC is that it is not decisively superior to high quality lossy codecs like Vorbis -v9 320 Kbps which is considered to be the best lossy codec and quality setting next to loss less codecs. MP3 ranks about third behind AAC which is usually second. I'm talking about double blind ABX test results with large international samples from different countries and populations. Most of these trials were conducted years ago and no more recent audio trials have been conducted especially for maximum bit rates allowed by each codec. Over time, we've lost recent data samples to draw inferences or conclusions.
 
MP3 is a universal codec and format that is widely supported. It's not terrible, but it's not state of the art either.
 
I conduced one ABX trial between my HD Tracks FLAC 24 bit 192 kHz album and the LAME 3.9.9 --preset-insane 320 Kbps MP3 version using my AKG K 702 with the Moon Audio Silver Dragon cable and I got 90 percent correct between the test samples. That's not a fair test to cite, but my previous Sennheiser HD-580 headphones with the stock cable made it almost impossible to tell any differences in the same ABX trial.
 
When comparing standard resolution 16 bits 44.1 kHz FLAC and MP3 versions of the same album using the AKG K 702 and the Moon Audio Silver Dragon cable, I'm back to where I started in this thread. I get far less than one tenth of one percent correct between the samples.
 
Note: the test samples are too small to draw anything substantive, but they're sufficient for my purposes. I still can't tell the differences in formal double blind ABX trials. However, I think that I can hear the differences between FLAC and MP3 when I know which version I'm listening to and the FLAC version sounds a little bit crisper and it has slightly better clarity and resolution. It's not much, but I can tell the differences when I'm told what I'm listening to.

 

[Leslie Dorner]

I did some more ABX tests and I found some interesting results for myself:
 
I can't tell the difference between -v0 224 Kbps VBR and 320 Kbps CBR with greater than 2 percent probability
I can't tell the difference between -v0 224 Kbps VBR and FLAC or WAV  with greater than 3 percent probability
Transcoding from 320 Kbps CBR to -v0 224 Kbps VBR leads to more audible artifacts
Transcoding from FLAC to -v0 224 Kbps VBR results in clean sounding tracks that sound nearly identical to each other in double blind ABX test results
LAME 3.9.9 -v0 224 Kbps VBR albums with 12 tracks at 16 bits 44.1 kHz resolution are typically 75 megabytes in total while --preset-insane 320 Kbps CBR are around 105 megabytes.
 
I used LAME 3.9.9 --vbr-new -v0 and --preset-insane as my options when transcoding from FLAC.
 
My new favorite is LAME 3.99 using --vbr-new and -v0 options for speed and quality combined. This helps me to rip and encode CDs or FLAC files quickly with no differences in sound quality for end results.
 
Lovin' it!

 

[cvision123]

For one album, yes.

When doing an ABX test, the percentage of correct answers needs to be greater than or equal 95% to show that you can clearly distinguish between lossy and lossless source. This is based on well established theory in statistical analysis.

In your case the result means that you can not distinguish between the two. Another point that is worth mentioning is that ABX test results have nothing to do with the quality of lossless or lossy file. ABX test just shows you whether you can distinguish the differences, not the quality of the sound. Of course the perceived quality helps test subject to make decisions.