I forgot how amazing a well mastered CD sounds!!!

[tybalt]

beanbag said:

SJobson said:

If the 320kbps MP3 sounds *that* different to the CD, it's going to be down to the playback hardware differences more than the format differences, I think.

Playing a CD or a FLAC file or a reasonable bitrate MP3 file through my stereo sounds pretty much the same - because all are fed into the same DAC. Relying on the DAC in your iPhone or in a PC sounds like the limiting factor.

With all due respect, I don't believe you know what you are talking about if this is what you truly think, or have unfortunately never had the opportunity to listen to audio on a high-end audio system.

However, my CDs don't play through my PC but on my Roksan CD transport and audio from my PC is fed to my AV amp through an optical lead.

The difference between the MP3 and CD was so massive, even my wife who has no interest in hifi noticed the difference commenting how clear the sound was (CD player).

I encourage you to take your 320kbps audio on your iPhone to a good HiFi store and ask for a comparison to a well mastered CD. The difference will astonish you I guarantee.

I'm with SJobson on this one. If you can't ABX it, you can't hear it. Try downloading foobar and doing some tests. Come back when you have ABX identified the difference between a wav and a 320kbps MP3 (same source and playback obviously). If you can do it reliably, then the guys at Hydrogen Audio would love to hear from you. Otherwise you are talking through your hat.

ETA - I see that Yogi has already said this, along with a helpful suggestion of another (possibly better) way of going about it.

Edited by tybalt on Thursday 4th February 11:32

[SJobson]

beanbag said:

And with regards to your sarcastic comments on my CD's. I get them from B&W Society of Sound in ALAC format and I burn them onto CD-R's at maximum 2x.

I also recommend you buy Dire Straits - Sultan of Swing. It's a superbly mastered CD, most audiophiles will agree....

I'm not sure why you think I'm being sarcastic - you don't buy CDs in the same way as most people do, from Amazon or HMV, clearly!

[telecat]

As CD has a bit rate of 1141.2Kbps as opposed to your MP3 at 320Kbps which do YOU think will sound better. Sound quality of MP3 remains dire and it lacks "atmosphere" and detail at frequency extremes. CD is not perfect but certainly is way ahead of MP3 especially when the equipment used has greater dynamic range than the headphones used on most i players.

[beanbag]

SJobson said:

beanbag said:

And with regards to your sarcastic comments on my CD's. I get them from B&W Society of Sound in ALAC format and I burn them onto CD-R's at maximum 2x.

I also recommend you buy Dire Straits - Sultan of Swing. It's a superbly mastered CD, most audiophiles will agree....

I'm not sure why you think I'm being sarcastic - you don't buy CDs in the same way as most people do, from Amazon or HMV, clearly!

Let's see then.....

SJobson said:

Oh, where are you getting these well-mastered CDs, by the way? As far as I am aware, the downloads available from iTunes are from the same masters as the CD, so it's not the mastering at fault either. And it's fairly well known that most CDs are not well mastered currently, with plenty of clipping to make them sound lounder and more initially impressive.

I never said I didn't buy my CD's in HMV or any other high street store. I just said to try out my sound system, I used some B&W SoS downloads which are mastered (according to B&W), to the highest standards.

I did mention I had purchased iTunes music but I also mentioned the quality was very poor and I regretted purchasing it.

I also pointed out a particular commercial album which is also mastered very well. Another recommendation is Radiohead - OK Computer. (Purchased from HMV when I lived back in England.....still got the price on it).

Thank you for your previous advice.

[tybalt]

telecat said:

As CD has a bit rate of 1141.2Kbps as opposed to your MP3 at 320Kbps which do YOU think will sound better. Sound quality of MP3 remains dire and it lacks "atmosphere" and detail at frequency extremes. CD is not perfect but certainly is way ahead of MP3 especially when the equipment used has greater dynamic range than the headphones used on most i players.

Higher bitrate does not necessarily mean it sounds better, for at least two reasons:

1 - Compression
2 - Psychoacoustics (i.e. limits of human perception)

[JustinP1]

tybalt said:

telecat said:

As CD has a bit rate of 1141.2Kbps as opposed to your MP3 at 320Kbps which do YOU think will sound better. Sound quality of MP3 remains dire and it lacks "atmosphere" and detail at frequency extremes. CD is not perfect but certainly is way ahead of MP3 especially when the equipment used has greater dynamic range than the headphones used on most i players.

Higher bitrate does not necessarily mean it sounds better, for at least two reasons:

1 - Compression
2 - Psychoacoustics (i.e. limits of human perception)

You are correct, CD is not going to be 3 times better, but MP3 encoding works by cutting out the data which it has been decided is 'least useful' of course as the amount of compression increases and thus the bitrate is further reduced, all this does is cut out more and more information from the sourse.

In essence, telecat is quite correct.

Edited by JustinP1 on Thursday 4th February 17:28

[Plotloss]

I personally still don't understand why people bother with CBR encoding.

[Eggle]

Some of the best music I have is on DVD-A ;-)

[beanbag]

tybalt said:

telecat said:

As CD has a bit rate of 1141.2Kbps as opposed to your MP3 at 320Kbps which do YOU think will sound better. Sound quality of MP3 remains dire and it lacks "atmosphere" and detail at frequency extremes. CD is not perfect but certainly is way ahead of MP3 especially when the equipment used has greater dynamic range than the headphones used on most i players.

Higher bitrate does not necessarily mean it sounds better, for at least two reasons:

1 - Compression
2 - Psychoacoustics (i.e. limits of human perception)

Not entirely true. By all means you are correct with regards to how MP3 removes "the frequencies you can't hear", however by removing these sound frequencies, it affects those you can hear and this is a huge reason for the lack of depth in MP3 music.

Sound waves interact with each other to create depth in sound and once you start removing bits and pieces, it affects this.

However....this is a good time to discuss the sound differences that I found between MP3 and CD audio.

I started off by taking what I believe to be a very well mastered CD. Both of which I've mentioned on here already however I took one song from each:

- Radiohead: Paranoid Android

I then encoded both songs into MP3 in several bitrates. 128, 256 & 320. To encode the music I used iTunes using the default settings.

- Sample rate: Auto
- Channels: Auto
- Stereo Mode: Joint Stereo
- Smart Encoding Adjustments
- Filter Frequencies Below 10Hz.

Seeing as these are the normal settings for all users, it's the most real life way of performing this test.

Once the music was encoded, I burnt the MP3's into CD-Audio format for playback on my Roksan Attessa ATT-DP2P CD player. My amplifier is an old but very worthy Sony STR-DB925 and I use a pair of B&W 684 speakers.

These were my findings:

* Original CD Audio: Paranoid Android starts with it's extremely detailed intro. The guitar strings were sharp, detailed and very realistic. The drums are subtle but again as with the guitar, very detailed with the sort of "punch" you'd expect from a drum sound. Tom Yorke's vocals then kick in and you can hear his breath between vocals. The whole combination sounds brilliant and with my eyes closed, almost live.

* MP3 - 160kbps: As you may imagine, there was a massive loss in detail with the lowest MP3 bitrate. All the instruments sound flat totally lacking depth or any of the realism from the CD-A.
Tom Yorke's voice was dull and had some tinniness to it. I've noticed this in other lower bitrate recordings. It's barely audible but it's there.

* MP3 - 256kbps: The music is better but still flat and lacking in much depth. Where the improvement lies is certainly with the guitar strings. There seems to be a lot more detail here with and the bass of the drums has more of a punch but it's still doesn't sound at all realistic like the CD-audio format. Tom Yorke's voice is certainly more detailed than the 160kbps bitrate, but again it doesn't live up to the original audio.

* MP3 - 320kbps: This is the bitrate I was ranting about and where many people said I couldn't tell the difference. Well, I had to switch between the CD-audio and the 320kbps encoded track, and I concluded that through casual listening (ie. Walking about the house doing your day-to-day thing), it's hard if not impossible to tell the difference.

However. That's certainly not my conclusion. I don't always listen to music like that and I massively enjoy a whisky glass of wine over my favourite music. This is when it goes terribly wrong.

The 320kbps encoded track started very well. The guitar strings sounded detailed and quite lively but there was something wrong with the drums. They sounded almost identical to the original but there was certainly something missing. Flat is the best way to describe the drums, but it was very subtle and if I wasn't familiar with the music, I would have probably have missed it.

The vocals were once again bright and detailed, but again there was something missing. Almost like the atmosphere of the music. Again it's something you do hear right away but it's there and once you identify the missing sound, it's hard to miss it!

From my point of view, this proves that it is possible to tell the difference between CD-A and MP3 320kbps. Perhaps a person without a tuned ear might not tell the difference, or perhaps somebody using a typical Dixons MIDI hifi couldn't hear the detail, but with my equipment, I certainly can.

It would be very interesting to perform a more scientific test as I'm sure the differences would appear very clearly identifying the inferior sound, however I don't have the equipment to do this.

Anyhow...there you go. I hope it helps!

Edited by beanbag on Thursday 4th February 22:16

[tybalt]

The results of sighted listening tests are not reliable. There's a built in bias to perceive a difference even where one does not necessarily exist. Did you try a blind test?

ETA - I meant to say. Thanks for taking the time to actually do a test and present the results.

Edited by tybalt on Thursday 4th February 23:12

[tybalt]

JustinP1 said:

You are correct, CD is not going to be 3 times better, but MP3 encoding works by cutting out the data which it has been decided is 'least useful' of course as the amount of compression increases and thus the bitrate is further reduced, all this does is cut out more and more information from the source.

In essence, telecat is quite correct.

I realise how MP3 compression works and of course your description is accurate. It's also the case that you can (roughly) halve the bitrate of a CD and have absolutely no loss in quality via lossless compression. The question for lossy compression methods is where is the threshold of perception? I would argue that you can only tell with a sensible test methodology and that means an ABX test. Having looked into ABX testing of lossy codecs (Hydrogen Audio forums are a good place to go for this), it generally seems to be the case that 320kbps is transparent (yes - using decent equipment). There are some exceptions - particular tracks and specific sounds can be used to identify compression but it generally seems to work very well.

I think there is far too much subjective waffle in hifi. The idea of a directional speaker cable being a particularly fine example.

Edited by tybalt on Thursday 4th February 23:44

[tybalt]

JustinP1 said:

You are correct, CD is not going to be 3 times better, but MP3 encoding works by cutting out the data which it has been decided is 'least useful' of course as the amount of compression increases and thus the bitrate is further reduced, all this does is cut out more and more information from the source.

In essence, telecat is quite correct.

I realise how MP3 compression works and of course your description is accurate. The question is where is the threshold of perception? I would argue that you can only tell with a sensible test methodology and that means an ABX test. Having looked into ABX testing of lossy codecs (Hydrogen Audio forums are a good place to go for this), it generally seems to be the case that 320kbps is transparent (yes - using decent equipment). There are some exceptions - particular tracks and specific sounds can be used to identify compression but it generally seems to work very well.

I think there is far too much subjective waffle in hifi. The idea of a directional speaker cable being a particularly fine example.

ETA - obviously you can do just as well with around half the bitrate of CD with lossless compression (that was what I was trying to say above)

Edited by tybalt on Thursday 4th February 23:52

[The_Burg]

Well mastered CD, let us into the secret, what was it?
Massively compressed over loud, yes we know that one but well mastered?? must have missed that one.
The HiFi world is dead, at least for mainstream stuff, how i dream of recorded music that sounds good and makes me want to listen!

Decay has set in, progress is gone. Oh to hear music!

[beanbag]

tybalt said:

The results of sighted listening tests are not reliable. There's a built in bias to perceive a difference even where one does not necessarily exist. Did you try a blind test?

ETA - I meant to say. Thanks for taking the time to actually do a test and present the results.

Edited by tybalt on Thursday 4th February 23:12

Absolutely, however I'm bound to the realms of my limits. I could rope my wife into helping witha blind test but she's about as interested in this as watching paint dry.

With regards to burgs comments; Once again we are all bound by the limits we have. As far as CD's go, OK Computer is a very well put together album however I disagree HiFi is a thing of the past.

If you look, it's there. The issue is that 99.9% of the population simply don't care about god quality music. They just want practicality. A great way to prove this is to look at the number of people who don't replace the earbuds that come with their iPods.

For those of us who enjoy HiFi there are many options including the now more commonly used and supported 24bit audio. Its just a shame my amp doesn't support. 24bit input a sampling rate higher than 48kHz.

Edited by beanbag on Thursday 4th February 23:49

[JustinP1]

tybalt said:

... The question is where is the threshold of perception? I would argue that you can only tell with a sensible test methodology and that means an ABX test. Having looked into ABX testing of lossy codecs (Hydrogen Audio forums are a good place to go for this), it generally seems to be the case that 320kbps is transparent (yes - using decent equipment). There are some exceptions - particular tracks and specific sounds can be used to identify compression but it generally seems to work very well.

In my line of work this is quite a contentious subject.

The effect of compression can be almost impossible to hear on some tracks and obvious on others. There is a very good reason why:

Music had changed. Instrumentation has changed. Two reasons:

1) There is less money around to record real instruments well - that's expensive.

2) It could be well argued that MP3 like any other format has a 'sound' which works well on it.


I'll explain - consider a digital piano. When you play a chord it sounds OK, reasonably good.

Then take a real piano, miked up in a real room. Same notes but it sounds 'real'.

That is the difference. In the real piano, you don't just get the simple notes, but the sound waves react with eachother both in the instrument and in the air. The second, third and fourth harmonics also react with the whole blend and also with the room and the result if well recorded gives an unmistakeable sense of reality and space.

All of this 'extra' information is all in the 10kHz and above. Unfortunately this is exactly the wavelengths which take the most data to digitise and the first to be cut in compression.

So how does this relate to MP3 and whether that sounds different to CD?

Imagine a 'classical' sounding vocal and a real piano. MP3 that and you will notice straight away - you lose the dynamics, the 'breath' and the air.

Then imagine a 'modern' processed vocal and a digital piano. MP3 that. You don't lose any dynamics, because there was little there in the first place. Then the 'air' - there is none on non-acoustic instruments. Even the best sampled pianos cannot model what goes on in an acoustic space.

So, if you MP3ed that, I daresay that you would have to go some to be able to tell the difference, because any 'losses' were not there in the first place and could be safely 'lost'.

In conclusion - I am afraid that the art is following the method of consumption.

Edited by JustinP1 on Friday 5th February 00:35

[CRACKIE]

The_Burg said:

The HiFi world is dead, at least for mainstream stuff, how i dream of recorded music that sounds good and makes me want to listen!

Decay has set in, progress is gone. Oh to hear music!

All is not lost. One of the very best traditional Hi-Fi companies has embraced digital with great success.

http://www.avihifi.co.uk/adm9.html

Could be a perfect solution for many No I don't work for them

Edited by CRACKIE on Friday 5th February 08:34

[SJobson]

CRACKIE said:

The_Burg said:

The HiFi world is dead, at least for mainstream stuff, how i dream of recorded music that sounds good and makes me want to listen!

Decay has set in, progress is gone. Oh to hear music!

All is not lost. One of the very best traditional Hi-Fi companies has embraced digital with great success.

http://www.avihifi.co.uk/adm9.html

Could be a perfect solution for many No I don't work for them

Edited by CRACKIE on Friday 5th February 08:34

Ha ha ha ha ha! Are you one of Ashley James' friends?

[CRACKIE]

SJobson said:

CRACKIE said:

The_Burg said:

The HiFi world is dead, at least for mainstream stuff, how i dream of recorded music that sounds good and makes me want to listen!

Decay has set in, progress is gone. Oh to hear music!

All is not lost. One of the very best traditional Hi-Fi companies has embraced digital with great success.

http://www.avihifi.co.uk/adm9.html

Could be a perfect solution for many No I don't work for them

Ha ha ha ha ha! Are you one of Ashley James' friends?

Ha Ha ~ No it was meant to be a generic "all is not lost" in response to the Burgs plea above, not me advocating AVI in paticular. I've never met Ashley James and not even heard adm9 myself to be fair but AVI's Nutrons and Bigatrons are special and I like the simplicity / flexibility of the ADM9 design. Trying to get back to topic about mastering; the reason I mentioned AVI was because of this ADM9 review http://www.avihifi.co.uk/adm9_hifichoice_review.ht... and the products ability to sound half decent when fed with some lower res file formats. To avoid any bias I could have put a link to Cyrus' DACXP instead or to the Benchmark DACPRE which also have good reputations for Hi-Fi results when fed with some of the lower res file types.

Edited by CRACKIE on Friday 5th February 15:30

[SJobson]

There are certainly enough comments about them on hi-fi forums, but quite often from Ash himself or people who appear to be rather close to him. And the inability to demo them does put me off somewhat.

[tybalt]

JustinP1 said:

In my line of work this is quite a contentious subject.

The effect of compression can be almost impossible to hear on some tracks and obvious on others. There is a very good reason why:

Music had changed. Instrumentation has changed. Two reasons:

1) There is less money around to record real instruments well - that's expensive.

2) It could be well argued that MP3 like any other format has a 'sound' which works well on it.


I'll explain - consider a digital piano. When you play a chord it sounds OK, reasonably good.

Then take a real piano, miked up in a real room. Same notes but it sounds 'real'.

That is the difference. In the real piano, you don't just get the simple notes, but the sound waves react with eachother both in the instrument and in the air. The second, third and fourth harmonics also react with the whole blend and also with the room and the result if well recorded gives an unmistakeable sense of reality and space.

All of this 'extra' information is all in the 10kHz and above. Unfortunately this is exactly the wavelengths which take the most data to digitise and the first to be cut in compression.

So how does this relate to MP3 and whether that sounds different to CD?

Imagine a 'classical' sounding vocal and a real piano. MP3 that and you will notice straight away - you lose the dynamics, the 'breath' and the air.

Then imagine a 'modern' processed vocal and a digital piano. MP3 that. You don't lose any dynamics, because there was little there in the first place. Then the 'air' - there is none on non-acoustic instruments. Even the best sampled pianos cannot model what goes on in an acoustic space.

So, if you MP3ed that, I daresay that you would have to go some to be able to tell the difference, because any 'losses' were not there in the first place and could be safely 'lost'.

In conclusion - I am afraid that the art is following the method of consumption.

All plausible stuff, but I'm afraid I'm still not convinced. Have you ever tried an ABX test? I must say I haven't, so I'm only going from what I've read. The only place I've found where people rigorously use ABX testing is at Hydrogen Audio forums, and there are a lot of people there who, when they actually try an ABX, find it far harder than they think to distinguish one from the other.