Audiophile Myths Part 1: MP3 VS FLAC, Cables, Sample Rates, Tube Amps, ETC.

[shibashis]

I am neither knowledgeable nor experienced enough to weigh in on this conversation, I do not have such revealing equipment either.
However I shall only share a story, probably it has little relevance to the topic. Back in 2007 I worked at a small company and were surrounded by many in their early-mid twenties, and the new found world of torrents made life beautiful. On our old Intel desktops were some pretty poor Logitech headsets. I somehow always knew to go with 320Kbps rips as opposed to 128 Kbps which were more prevalent back then.

I used to listen to mostly classic Rock and so downloaded exclusively those, the rest of the pack were of course happy to get recent Bollywood releases, I think it was purely by chance one of my colleagues heard 'And I love her' by Beatles and he liked it. I was delighted, felt that he would be my new pupil and I would lead him to a new world of music and adventure (He was using a terminal where I downloaded some music and did not know they were in 128 Kbps). I told him if you listen to better rips of the same song it would sound much better, he being the software development apprentice that he was almost violently opposed me by commenting that I have been smoking more cannabis than is medically permitted. Luckily I had proof, on my terminal I had the same song from the same album in 320 kbps, and made him listen to it. Identical hardware config with identical software and identical crappy Logitech headsets. The difference to him was night and day, and I know for a fact that his understanding of music - quality - equipment was much like his understanding of 128 versus 320 kbps rips.

Everyone here knows the difference between 128 and 320, it is quite a jump. I can distinguish between 256 and 320, now from 320 to a FLAC rip, should not there be a discernible difference?

 

[Thad E Ginathom]

I can distinguish between 256 and 320, now from 320 to a FLAC rip, should not there be a discernible difference?

It doesn't logically follow, no. because A is worse than B, C must be even better? Because the numbers are bigger?

 

[tcpip]

Everyone here knows the difference between 128 and 320, it is quite a jump. I can distinguish between 256 and 320...

Have you tried distinguishing between 256 kbits/sec and 320, in the following conditions:

• the two encoded files must have been created from the same uncompressed source file with all other encoding parameters kept the same, and
• it must be a double-blind comparison

Have you tried identifying files based on bit-rate under these circumstances? If not, two possibilities are very likely

• the files differed in ways much more significant than just their bit rates. Maybe their source files were different, maybe various other encoding parameters were different.
• you were being led to believe that you could spot differences because of your own listener expectancy (a term well discussed in audio literature)

The first point (about sources themselves being different) is a very common reason why a lot of modern audiophiles prefer vinyl over CD. It is not because the vinyl medium has some magic, but because the music publishers get their vinyl masters processed differently from their CD masters. The publishers know that most vinyl lovers in the 21st century are discerning listeners with good systems whereas most CD listeners listen in cars or on computer speakers (statistically speaking). Therefore, they master the two "sources" differently before sending them for manufacture. The audiophile foolishly thinks there's magic in the vinyl medium. And buys more vinyl. The publisher feels very happy. The same has been reported in some cases about SACD versus CD -- the content itself is different, and deliberately so.

 

[Ambio]

I think the whole pursuit of high resolution is the case of missing the wood for trees. 16/44.1 is good enough. I rarely care to check the resolution of the files unless I feel the quality is distorting. In some cases, the oldies recorded in Mp3 sounds much better than audio CD.

I still buy Hi Rez not because they sounded better but just because it is there.

 

[prem]

I agree Ambio. In fact I preferred the original non remastered cds of several English titles to the SACDs of the same. A well mastered 44.1 kHz 16 bit cd is more than enough

 

[prem]

Ambio, when you mention oldies are you referring to Bollywood oldies?

 

[Ambio]

More of Kollywood. MSV stuff and the early Illayaraja. I used to listen to Hindi songs on vinyl but then when cassette player came they were simply thrown away. Few hundreds of them. I still miss Sangam, Mere Nam Joker, Hare Rama Hare Krishna, Chalte Chalte...occasionally I listen to them from YouTube.

It is difficult to get old Hindi songs here. Miss them very much.

 

[prem]

HI Ambio, for old Hindi songs try and source cds which have PMLP xxxx as their catalogue number and are pressed in England. These are the only good sounding cds. There are about 500 of them. All old Hindi songs, either films or compilations. These were mastered without any digital gimmicks by either Disctronics or Nimbus. These CDs sound very close to how vinyl sounds. Anything else in digital is rubbish. All no noised and compressed

 

[asliarun]

HI Ambio, for old Hindi songs try and source cds which have PMLP xxxx as their catalogue number and are pressed in England. These are the only good sounding cds. There are about 500 of them. All old Hindi songs, either films or compilations. These were mastered without any digital gimmicks by either Disctronics or Nimbus. These CDs sound very close to how vinyl sounds. Anything else in digital is rubbish. All no noised and compressed

Sir, you are a wealth of information.

 

[prem]

Hi Asliarun

Thanks. These CDs were pressed between 1988 and 1993. You will have to source them on eBay. Else if someone has it you can take a rip.

 

[cyber_cat]

I agree Ambio. In fact I preferred the original non remastered cds of several English titles to the SACDs of the same. A well mastered 44.1 kHz 16 bit cd is more than enough

This is what people need to realize. The source file needs to be good. A FLAC of non-remastered cd, the original recordings, beats the pants out of vinyl rips at any or bit any HZ.

You make a flac out of bloated, remastered CDs and play them thru dacs it is going to sound bad.

 

[Thad E Ginathom]

Shouldn't that be Compressed, remastered CDs?

But compression for LPs is specific to the limitations and vagaries of that media. Added to which, the compression demon was alive and well, and started in pre-digital days.

So my suggestion is that one should just say well-mastered, whatever the target medium.

I think I have been lucky, or maybe it is the period/genre of music that I listen to, that late-sixties/early-seventies hippy stuff... I think that all my CDs and/or digital-file versions of LPs that I have known and loved have been true to, but better (cleaner, obviously, but sometimes more detailed too) than the LPs themselves.

I do remember, though, a listening comparison at a HFV meet where the CD and the LP bore little resemblance to each other: I never knew they could be that different. It sounded like a different band!

 

[spirovious]

HI Ambio, for old Hindi songs try and source cds which have PMLP xxxx as their catalogue number and are pressed in England. These are the only good sounding cds.

Is this the list you are saying about?
The Gramophone Company of India (HMV) Records Listing: PMLP/PSLP

 

[cyber_cat]

Shouldn't that be Compressed, remastered CDs?

But compression for LPs is specific to the limitations and vagaries of that media. Added to which, the compression demon was alive and well, and started in pre-digital days.

So my suggestion is that one should just say well-mastered, whatever the target medium.

I think I have been lucky, or maybe it is the period/genre of music that I listen to, that late-sixties/early-seventies hippy stuff... I think that all my CDs and/or digital-file versions of LPs that I have known and loved have been true to, but better (cleaner, obviously, but sometimes more detailed too) than the LPs themselves.

I do remember, though, a listening comparison at a HFV meet where the CD and the LP bore little resemblance to each other: I never knew they could be that different. It sounded like a different band!

Forgetting about compression for some time, some LP to CD transfers have their channels swapped or one of channels out of phase with the other; how about that :lol:

 

[Thad E Ginathom]

Good grief!

 

[prem]

Hi Spirovious

Yes. The CDs though are only from the 5000 series. The 1000 to 4000 series are vinyls. Again the 5000 series is not fully represented in that link. It starts from 5001. That title is Beyond Time by Jagjit it's a running series. I used to have almost all the CDs in that series. Actually the first one is PSLP 5001. S stands for stereo. IIRC 5002 was also a PSLP title, Call of the Valley. 5003 and 5004 were PMLP titles. Tribute to Kishore Kumar Vol 1 and Vol 2. M stands for Mono. More than 90 % of the titles were mono in the 5000 series

 

[arj]

This is what people need to realize. The source file needs to be good. A FLAC of non-remastered cd, the original recordings, beats the pants out of vinyl rips at any or bit any HZ.

You make a flac out of bloated, remastered CDs and play them thru dacs it is going to sound bad.

Exactly..spot on. in the end its the mastering and source rather than the format !
I have CDs which sound worse than the LPs (As prem mentioned any Old hindi cd pressed here is an example ? and also those where they sound very similar..only difference being due to the source equipment (CD/TT) being different sounding

But again there is so much music only available on CDs and not on LPs.

Recently visited an FMs place which is all digital and really high end sound. the purity of sound and the emotional connect was something to be believed..as long as the content was good. when we put a track which was poorly mastered for CD the connect just went and the sound was hollow - one could listen to it no doubt but the earlier magic from the sound had gone away !

 

[shibashis]

Have you tried distinguishing between 256 kbits/sec and 320, in the following conditions:

• the two encoded files must have been created from the same uncompressed source file with all other encoding parameters kept the same, and
• it must be a double-blind comparison

Have you tried identifying files based on bit-rate under these circumstances? If not, two possibilities are very likely

• the files differed in ways much more significant than just their bit rates. Maybe their source files were different, maybe various other encoding parameters were different.
• you were being led to believe that you could spot differences because of your own listener expectancy (a term well discussed in audio literature)

I decided to take the suggestions and put myself through the 'Test'.
I have had limited time to play with it so for the time being I have tried with only one album, the self-titled Lynyrd Skynyrd album from 1991, I own the CD.

Here is what I did:
1) Imported the CD using 'Format Factory' (found here: Format Factory Home Page - Free media file format converter). Imported on my Dell Inspiron, first mp3 at 320 and imported it again using Flac with default settings for the software. Did not tinker with any of the settings on the software or the laptop.
The ripped files were put on the same folder. In the folder of course I can see which one is mp3 and which is Flac.

2) I don't have anyone to help me with the test and so keep the test 'Blind' I decided to use my cell phone (HTC Desire 816, quite decent for music, clear and loud enough with my ATH-M50). If I am using PC and Foobar I can see which file is which before I push play, not to be biased with this information the Cell phone is a better choice. I can not see the file type unless I get to 'details' and still can play the two different tracks one after the other. All the tracks were put on a playlist with HTC Sync Manager and then played with the stock HTC player provided on the phone.

Findings:

At least for this album and the hardware at my disposal the difference is negligible. If anything the Flac rips sounded a little more neutral and warmish compared to the 320 mp3s. I did find the low frequencies extended very slightly more on the Flac files and there was little and I mean very very little extra definition on the strings, I had to strain my ears to confirm this.

Perhaps with a better overall setup the findings would be more definite, with more albums the results could be different and I will try to test with more and different kinds of music. Time permitting I would like to do the same test on my Laptop running Daphile with my stereo setup and try to pick my ears for difference.
These are just my findings, one can call this 'placebo' effect, it could well be that but somehow I feel it is not.

 

[Thad E Ginathom]

You probably know the enormously popular free Windows player Foobar2000? It has an ABX addin. This is what the ABXing crowd seem to favour for such tests.

I would like to do the same test on my Laptop running Daphile

But of course, some of us don't use Windows.

Foobar runs in Wine. I'm not sure at all if that is as "pure" as a player running native in Linux. I tried it a few days ago, for the first time, and would really need to get used to the methods and procedures. On that occasion, I gave up when I realised that I couldn't tell the difference "sighted!" (Play A/Play B) so there was no point in going "Blind."

There is a cross-platform ABX tester: Lacinato ABX. I wanted, specifically to compare sample rates, but I can't do that with JACK, as that means fixing the sample rate and doing SRC, and I could not get it to output to ALSA device.

Always happy to see people willing to actually test. Quite apart from challenging our own beliefs, it is much harder work than listening to music, and ...rather boring!

 

[captrajesh]

I gave up when I realised that I couldn't tell the difference "sighted!" (Play A/Play B) so there was no point in going "Blind."

Hmm! It appears that you are immune to "placebo effect" and "psychoaccoustics" too. May be there could be more! What say Thad.