Better mastered music could be on the way!

[relocated]

Has What HiFi got nothing to say about Steve's findings on the Linn files?

Would seem like a very valid investigation that What HiFi could pursue on behalf of possibly exploited members of the music buying public. Might not be so good for advertising revenue though?

 

[oldric_naubhoff]

Here are todays Linn downloads. The track is called 'The Missing Plutonium'.

As you can see from these pictures the two files are different again today. The FLAC version has several seconds of silence at the beginning and end but the MP3 version has had these silent sections removed. Apart from that the two tracks are the same. They are both at the same volume level and they both have a matching amount of dynamic range.

IMO this is actually a sensible way of editing the original 24bit FLAC studio version before converting it into an MP3. Just cropping the silent sections off the beginning and end but leaving the rest of the sound exactly as it is.

What does seem odd though, is that over the past three days all of the MP3 tracks have been edited but in different ways. The first had the dynamic range compressed, the second had the volume altered and the third had the silent sections cropped. :?

Wednesdays 24 bit FLAC

http://i1239.photobucket.com/albums/ff512/steve__1979/Day4FLAC.jpg

Wednesdays 320kbps MP3

http://i1239.photobucket.com/albums/ff512/steve__1979/Day4MP3.jpg

good job Steve, again. 8) but I have to disagree with you. they are not identical masters. by looking at the pictures it seems like some of the background instrument of the mp3 version is shelved down compared to 24 bit version. this can be seen easiest at the beginning of the song. this trick will make seem hi-res sound more "clear".

 

[The_Lhc]

Here are todays Linn downloads. The track is called 'The Missing Plutonium'.

As you can see from these pictures the two files are different again today. The FLAC version has several seconds of silence at the beginning and end but the MP3 version has had these silent sections removed. Apart from that the two tracks are the same. They are both at the same volume level and they both have a matching amount of dynamic range.

IMO this is actually a sensible way of editing the original 24bit FLAC studio version before converting it into an MP3. Just cropping the silent sections off the beginning and end but leaving the rest of the sound exactly as it is.

What does seem odd though, is that over the past three days all of the MP3 tracks have been edited but in different ways. The first had the dynamic range compressed, the second had the volume altered and the third had the silent sections cropped. :?

Wednesdays 24 bit FLAC

http://i1239.photobucket.com/albums/ff512/steve__1979/Day4FLAC.jpg

Wednesdays 320kbps MP3

http://i1239.photobucket.com/albums/ff512/steve__1979/Day4MP3.jpg

good job Steve, again. 8) but I have to disagree with you. they are not identical masters. by looking at the pictures it seems like some of the background instrument of the mp3 version is shelved down compared to 24 bit version. this can be seen easiest at the beginning of the song. this trick will make seem hi-res sound more "clear".

They're different shades of blue as well, I'm sure that's important! #notaclue

 

[oldric_naubhoff]

there's one more thing I need to draw your attention to with regards masters of the hi-res songs from Linn's web site. despite the fact they are 24 bit - meaning they have the possibility to utilise much wider dynamic range than with mere 16 bit version - they are still compressed! it's most clearly seen on the 3rd download. it may not necessarily be digital compression as the Audacity doesn't mark anything in red but I believe some analog compressors were used during mastering process of the songs, most likely on the percussion, as the peaks don't reach as high over the mean as they should and also the peaks couldn't be as level in real life as they are on the recording. here's a graph of Dire Straights' "Walk of Life" from the 80's. Dire Straights early masters are still pointed out as the best examples of how to do mastering properly, despite such a long time has since passed.

so, this is exactly what I said before; 24 bit does not give you better quality recordings. good quality mastering does. and it's really a shame that having access to such a wide dynamic range 24 bit format ca offer you still get compressed music!!!

of course music mastered with preserving high dynamic range will initially sound quieter compared to low dynamic range recording - that's why you need to pump up the volume considerably to get decent volume levels. but when you do the music sounds so much more realistic and natural.... definitely worth an effort to get yourself a better amp able to handle high dynamic peaks.

 

[manicm]

Talking of challanges.....try mine. Download that track as described and see for yourself.

I'm don't like wasting money but if you're willing to credit my PayPal account with the £3.48 that it costs to download the tracks I will be happy to take you up on your challenge.

But as I've already said I've already tried converting 16 and 24 bit music to 320kbsp MP3s before (so have Sony and Apple!) and it makes no audiable difference to the sound quality. To anyone reading this who doubts me I urge you to prove it to yourselves at home.

Rip a CD track to a lossless WAV file.

Convert the WAV file to a 320kbps MP3 using LAME.

Use Foobar with the ABX plugin to test yourself to see if you can hear any difference between the original WAV file and the MP3 file.

Forget about even comparing to a WAV file, I find the most recent versions of LAME pretty, um, lame. There is clearly audible high-frequency roll-off which robs the music of transparency and life. I've also found some recent MP3s bought from 7Digital to sound lifeless - and they mostly use LAME as well. By and large, MP3s and AACs are consigned to the rubbish bin as far as I'm concerned.

 

[steve_1979]

good job Steve, again. 8) but I have to disagree with you. they are not identical masters. by looking at the pictures it seems like some of the background instrument of the mp3 version is shelved down compared to 24 bit version. this can be seen easiest at the beginning of the song. this trick will make seem hi-res sound more "clear".

I'm not sure what you mean by "the background instrument of the mp3 version is shelved down".

BTW. Have Linn replied to the Email that you sent them yet?

 

[manicm]

....when I listen to the converted MP3, there's absolutely no question that it does not sound as good as the original uncompressed file, no matter what LAME encoder I use. No ifs, no buts, and no ABX test needed.

No ABX test needed?

So I take it from this statement that you haven't actually tried converting a lossless WAV file to a 320kbps MP3 using LAME to do the encoding and then using Foobar with the ABX plugin to test yourself to see if you can hear any difference between the original WAV file and the MP3 file.

Go on give it a go. It's a simple test that only takes about 20 minutes to do. People can write anything on forums but I urge everyone reading this to prove it to themselves at home. See for yourself if you can pass this blind ABX test 10 times in a row.

I don't need any ABX or blind testing to tell me that 320k MP3s are audibly inferior to WAVs. And you're telling me that you need ABX and blind testing to tell the difference between 128k and 320k files? Please! ABX and blind testing are not 'science', they're merely methodologies with inherent advantages and disadvantages like any other. Just don't call it science, cos to do so you and CraigM and anyone else are just plain wrong.

I'm in IT and some people evangelise Waterfall, Agile, Extreme software development methodologies. And we take the best bits which will work according to current requirements and situations. There is no one-size fits all. Sometimes an entire methodology will work for a specific situation, sometimes not, but we don't call it science. We just don't call it science, cos there are better ways of laughing at ourselves.

 

[oldric_naubhoff]

I'm not sure what you mean by "the background instrument of the mp3 version is shelved down".

if you take a look at the graphs they are not the same. secondary peaks differ. but as you say they sound the same that may be the case. I'm only judgeing after the graphs.

BTW. Have Linn replied to the Email that you sent them yet?

nope, not yet. if I don't get any reply till tomorrow afternoon I'm going to send in another one. they say I should be waiting up to 3 days for an answer so I'll give them a chance.

 

[manicm]

24 bits don't give you more discrete levels of volume within 16 bits. they only give you more volume levels on top of 16 bits.

Ho ho ho, ha ha ha, where did you get this information mate that more bits simply give you more volume???? That just sounds plain fishy, please substantiate with expert evidence.

The theory behind a larger bit rat, is to give you a wider dynamic range which is just not possible with 16-bit audio. And if you want to win the loudness war then 24-bit audio is an arguably better platform to begin with. And yes, proper mastering also plays a vital role.

And also, while in theory human hearing tops out at around 15-18khz, some audio engineers believe that a higher frequency rate can have a positive effect on the lower frequencies which ultimately can result in differences being heard.

So someone said that human hearing can't be trusted?? And human thinking and logic cannot be always trusted either - welcome to planet earth mate!

 

[oldric_naubhoff]

24 bits don't give you more discrete levels of volume within 16 bits. they only give you more volume levels on top of 16 bits.

Ho ho ho, ha ha ha, where did you get this information mate that more bits simply give you more volume???? That just sounds plain fishy, please substantiate with expert evidence.

is googling really that difficult? wiki:

By increasing the sampling bit depth, quantization noise is reduced so that the S/N is improved. The 'rule-of-thumb' relationship between bit depth and S/N is, for each 1-bit increase in bit depth, the S/N will increase by 6 dB.[2][3] 24-bit digital audio has a theoretical maximum S/N of 144 dB, compared to 96 dB for 16-bit; however, as of 2007 digital audio converter technology is limited to a S/N of about 124 dB (21-bit)[4] because of real world limitations in integrated circuit design.

And also, while in theory human hearing tops out at around 15-18khz, some audio engineers believe that a higher frequency rate can have a positive effect on the lower frequencies which ultimately can result in differences being heard.

So someone said that human hearing can't be trusted?? And human thinking and logic cannot be always trusted either - welcome to planet earth mate!

you know this is rubbish. "some audio engineers"... besides, if this is true, don't you thing that recorded sound is already modulated by ultrasonic content? so it will come out of speakers as such. no matter if ultrasonic content is replayed or not.

 

[SteveR750]

OK I'm not sure what you are trying to say...because the science says otherwise then it cant be so?

I'm only going by what I can hear, and the difference that I described is pretty obvious to me. There is no argument, just assimilate into the data set. If you are trying to discredit my ears simply because your science doesnt fit, well figure where that will end up.

My post asked if the 24/48 and 16/44 that you mentioned in your post are the same master, and if they are how did you downsample the 24/48? Downsampling needs to be done properly if it is not to introduce audible errors/noise, there is lots of information available about it. And given how unreliable sighted listening tests and human perception are, yes the science is more reliable. Oh, and it's not MY science and I have no idea what "figure where that will end up." is supposed to mean, but I suspect I won't care.

Not sure I follow / understand you then. I haven't downsampled anything, the files are being streamed in their native depth / rates into the DM+ which upsamples AFAIK.

Rip a CD track to a lossless WAV file.

Convert the WAV file to a 320kbps MP3 using LAME.

Use Foobar with the ABX plugin to test yourself to see if you can hear any difference between the original WAV file and the MP3 file.

I've tried this before, but using J River to burn discs in CDA and also MP3. I compared both versions using JRMC randomly some time ago, which is why I'm pretty confident (with the psycho caveat etc etc etc)

However, my personal feeling is that unless a hi-res version is from a remaster, then chances are it's a resampled original CDA master, and sold as HD. I'm pretty sure that most of my HDTracks purchases are in that category. I can believe, indeed not at all surprised that files are "manipulated" to make hi res appear to sound better as a business model.

 

[SteveR750]

I found out for myself that 320k AAC (VBR) sounds as good, to me, as Apple Lossless for my purposes. My 'raw material' is rips from my own CDs (16bit 44.1khz) and I have no interest in 24/96 files (or higher) because there is so little choice and prices are so high.

What the rest of you get up to is of no matter to me.

However, the tone these debates take is toxic to the forum. The same characters 'parachuted in' every time to lay down the law and set tests (where do we hand in our homework sir?) as if they, personally, represent the only truth possible and should automatically hold some natural authority on the subject.

Of course they don't. 'The truth' in these matters is always what makes the individual happy.

Some people want to pay the premium and use higher resolution files in the same way that some have preferred SACDs to CDs or returned to vinyl LPs (or use valves or prefer active speakers or whatever else is on offer). Good luck and good listening to all of you. Keep buying the gear and keep the industry and the choices going.

I am stubbornly sticking to my autodidactic and selfish ways of using whatever I like in any way I choose that suits me. If your 320k MP3s sound better than my 320k AACs and someone else's 24/96 files sound even better than my Apple Lossless files then well done. You are probably as happy with your setup as I am with mine.

Stuff the homework, the links, the downloads the tests and the overbearing "this is for your own good, now put on these electrodes", tub-thumping acolytes of 'The One True Way'.

Post of the thread, and quite possibly of all threads like these. Especially that bit in bold.

I'm inclined to agree. When it gets just that little bit ego / personal it becomes slightly unpleasant. I'm a fan of Steve's posts in this thread, but not so much his disciples.

 

[steve_1979]

I don't need any ABX or blind testing to tell me that 320k MP3s are audibly inferior to WAVs.

I've tried this before, but using J River to burn discs in CDA and also MP3. I compared both versions using JRMC randomly some time ago, which is why I'm pretty confident (with the psycho caveat etc etc etc)

Anyone who thinks that they can hear a difference between a 320kbps MP3 and a lossless WAV, FLAC or ALAC file are mistaken. I can easily prove this to you too. All you have to do is follow these two simple steps.*



1. Convert a lossless WAV, FLAC, or ALAC file into a 320kbps MP3 using LAME.

2. Use Foobar with the ABX add on to compare the lossless file to the MP3 file.

* For detailed instructions on how to do this see post number #8 on page 6 of this thread.

 

[Anonymous]

Analyzed: Bob Marley / Legend -------------------------------------------------------------------------------- DR ----- Peak ----- RMS ---- Duration Track -------------------------------------------------------------------------------- DR13 -2.51 dB -17.82 dB 3:52 01-Is This Love DR14 -0.82 dB -17.39 dB 7:10 02-No Woman, No Cry DR14 -1.29 dB -16.71 dB 3:57 03-Could You Be Loved DR14 -1.20 dB -16.71 dB 3:01 04-Three Little Birds DR13 -1.31 dB -16.00 dB 4:18 05-Buffalo Soldier DR13 -4.42 dB -18.87 dB 3:18 06-Get Up, Stand Up DR13 -1.66 dB -16.28 dB 5:31 07-Stir It Up DR13 -1.60 dB -16.71 dB 2:55 08-Easy Skanking DR14 -3.39 dB -19.21 dB 2:52 09-One Love / People Get Ready DR13 -2.53 dB -17.56 dB 4:42 10-I Shot The Sheriff DR15 -0.70 dB -17.00 dB 4:12 11-Waiting In Vain DR13 -3.08 dB -19.84 dB 3:47 12-Redemption Song DR16 -1.91 dB -20.71 dB 4:32 13-Satisfy My Soul DR14 -1.20 dB -17.48 dB 7:41 14-Exodus DR14 -1.14 dB -16.30 dB 3:32 15-Jamming DR14 -0.92 dB -15.91 dB 6:53 16-Punky Reggae Party -------------------------------------------------------------------------------- Number of tracks: 16 Official DR value: DR14 Samplerate: 96000 Hz Bits per sample: 24 Bitrate: 2911 kbps Codec: FLAC

 

[manicm]

oldric_naubhoff - this IS NOT RUBBISH. Thanks for stating the blatantly obvious regarding the theoretical headroom of 24 bit music. While the apparent volume could be louder, it could also be softer. That is not the point, the point is it allows for a much wider passage between the softer and louder bits of a piece of music, thus increasing dynamic range, and so if you want to get rid of the 'loudness' wars then 24-bit is a much better platform to start off with.Even if apparent human hearing tails off at around 15khz, I'd rather have 48khz (effective rate of 96khz recordings), than a limit of effective 22khz (CD quality).

 

[manicm]

Anyone who thinks that they can hear a difference between a 320kbps MP3 and a lossless WAV, FLAC or ALAC file are mistaken. I can easily prove this to you too. All you have to do is follow these two simple steps.*

We are mistaken? Oh the horror, the horror...

1. Convert a lossless WAV, FLAC, or ALAC file into a 320kbps MP3 using LAME.

2. Use Foobar with the ABX add on to compare the lossless file to the MP3 file.

* For detailed instructions on how to do this see post number #8 on page 6 of this thread.

Don't care to follow your instructions - I rip a WAV, and then a MP3 of the same piece of music, and I prefer the former. Thank you and good night.

 

[manicm]

steve_1979 - WRT your Linn testing, a few years ago I contacted Linn about what they use to rip their MP3s, as many of us are impressed with their quality (just plain listening OK?), and they gave me what I thought to be a very vague answer. I suspect that somehow they don't use LAME or somesuch ripper, but master simultaneously to high-res, CD quality and MP3 directly. Could you get clarification from them on this?

And by the way post your results on the Linn Forums - I strongly suggest you do if all of us are to get the facts - as their engineeers regularly and enthusiastically respond.

Why not go straight to the horse's mouth????

 

[John Duncan]

There doesn't appear to be an awful lot of music being listened to in this thread...

 

[John Duncan]

And by the way post your results on the Linn Forums - I strongly suggest you do if all of us are to get the facts - as their engineeers regularly and enthusiastically respond.

Why not go straight to the horse's mouth????

It seems somebody already did.

 

[manicm]

There doesn't appear to be an awful lot of music being listened to in this thread...

Is that a look of surprise on your face Anyway that's where you're wrong Mr Duncan. I have been listening you see - downloaded the free Linn hi-res file of Berlioz and played it through the USB on my CA-751BD - yes the damn thing played the FLAC file - all 24 bits and 96khz of it.

And good it sounded too (but I don't have a CD to compare). But if anything stood out, is that there's a good case of record companies applying compression or escalating the 'loudness wars'. Had to continuously adjust my volume as this piece is suddenly pin-drop quiet one second, and bomb-exploding in another. But I suppose this is more prevalent in classical music.

 

[Andrew Everard]

But if anything stood out, is that there's a good case of record companies applying compression or escalating the 'loudness wars'. Had to continuously adjust my volume as this piece is suddenly pin-drop quiet one second, and bomb-exploding in another.

Surely that's an example of a record company not "applying compression or escalating the 'loudness wars' " ?

 

[manicm]

And by the way post your results on the Linn Forums - I strongly suggest you do if all of us are to get the facts - as their engineeers regularly and enthusiastically respond.

Why not go straight to the horse's mouth????

It seems somebody already did.

Steve should post his results there as he did here. A reader there read his comments here. Not trying to prove who's right or wrong, it would just be genuinely interesting to see Linn's response.

 

[manicm]

But if anything stood out, is that there's a good case of record companies applying compression or escalating the 'loudness wars'. Had to continuously adjust my volume as this piece is suddenly pin-drop quiet one second, and bomb-exploding in another.

Surely that's an example of a record company not "applying compression or escalating the 'loudness wars' " ?

Precisely, and my mistake - I meant to say there's a good case for compression in this particular example.

 

[Andrew Everard]

Precisely, and my mistake - I meant to say there's a good case for compression in this particular example.

I have the same recording, and there's really no case for compression.

 

[manicm]

Well I was listening at night, and perhaps didn't want to crank the volume up too much in fear of disturbing neighbours etc.