PDA

View Full Version : FLAC TO WAV Conversion



Audiocom AV
15-06-2008, 17:46
Hello All

What is the optimum way to convert FLAC files back to WAV without losses? At the moment I add the files to FLAC frontend and decode.

Also, how do you create a CUE sheet so the files converted back can be added to EAC and written to CD-R??

Any help would be appreciated.

Regards,
Mark Bartlett

NRG
16-06-2008, 04:36
Don't know about the Que sheet but FLAC front end is the best way to convert back to WAV.

BTW FLAC *is* lossless and the conversion back to WAV will *always* result in the WAV file being an exact bit copy of the original. So all you need to FLAC front end to perform the conversion....

colin.hepburn
07-09-2008, 15:48
Hi All
OK tiring too figure out digital audio files conversion types i.e. from Mp3 through to flac as I may want to download some music or copy my CD collection to hard drive as I don’t want to use MP3 as play back more like Flac I think if that’s the best format so what happens here if I want to convert mp3 or wav /wma files to Flac I’m thinking this cant work or can it:confused: I see Flac down to mp3 but not mp3 up to Flac anyone got advice on this
Cheers
Colin

lurcher
07-09-2008, 17:16
Simple way would be mp3 to wav, then wav to flac, but if you are starting with mp3, there is little (IMO) to gain converting it to flac, the information is already lost.

StanleyB
07-09-2008, 17:42
Simple way would be mp3 to wav, then wav to flac, but if you are starting with mp3, there is little (IMO) to gain converting it to flac, the information is already lost.
Some of it is lost, but not all of it. However, in order to keep any losses to a minimum, it is best to keep the mp3 files as is and use an oversampling DAC to play it back. An 8x oversampling DAC would in theory be reproducing a 128kbs mp3 file at 8X the 128kbs rate.

lurcher
07-09-2008, 17:51
Some of it is lost, but not all of it. However, in order to keep any losses to a minimum, it is best to keep the mp3 files as is and use an oversampling DAC to play it back. An 8x oversampling DAC would in theory be reproducing a 128kbs mp3 file at 8X the 128kbs rate.

No, but the information that is lost by encoding to mp3 has gone forever. I fail to see any way that using a oversampling filter would manage to restore any of the information removed by the perceptial encoding, they are two very different processes.

You seem to be implying that using a 8X OS filter on a 128kbs mp3 file would produce the same results as a 1024kbs mp3 file, and thats just not true in any sense.

purite audio
07-09-2008, 18:12
From a DAC manufacturer!!!!!!

StanleyB
07-09-2008, 18:32
You seem to be implying that using a 8X OS filter on a 128kbs mp3 file would produce the same results as a 1024kbs mp3 file, and thats just not true in any sense.
Have you tried converting various waveforms from WAV to mp3, and then comparing the WAV and mp3 playback, or are you just guessing?

leo
07-09-2008, 18:34
No, but the information that is lost by encoding to mp3 has gone forever.

Agreed! maybe we misunderstood the post:confused:

StanleyB
07-09-2008, 18:35
From a DAC manufacturer!!!!!!
Would you prefer an answer from a street cleaner or a pub landlord? This ain't wigwam where experts are not allowed to comment.

StanleyB
07-09-2008, 18:45
No, but the information that is lost by encoding to mp3 has gone forever.

Agreed! maybe we misunderstood the post:confused:
I am afraid to say that it is incorrect to assume the information is gone forever.

Putting aside the fact that I do a DAC now, I also did the world's 3rd mp3 player ever in 1997. It was called the MPress3, and I did a lot of research then into mp3 encoding and decoding. So I am penning down this info as an mp3 expert with mp3 design and manufacturing experience.

The fact of the matter is that it is not an analogue signal that is converted, but a digital one. The method is similar to the zip or rar compression method. So you get a smaller file on compression, but after decompression, the info is supposed to be completely intact.
However, in the case of mp3, there are additional compression levels and techniques. VBR is the best. The mp3 compression discards data, but some of it is recovered through algorithms.

This is not much different from the way lost data on an audio CD is recovered using the Sallen-Key method. I have test discs with holes drilled in them that play quite OK even though it is quite obvious that large chunks of data is missing. The decoding method on an audio CD allows for recovery of lost data.
A similar method is employed in an mp3 decoder.

leo
07-09-2008, 20:03
You must be better at it than most of us , tbh every MP3 converted piece of music either played as is or decrompressed burnt to a cd sounded shite compared to lossless to me

BTW Stan, I'm just trying to learn something here;) just going by past experience with MP3 , its one of the reasons I don't bother with it and usually use flac if I want compressed music files on the HD

lurcher
07-09-2008, 20:41
However, in the case of mp3, there are additional compression levels and techniques. VBR is the best. The mp3 compression discards data, but some of it is recovered through algorithms.

Well, in that case in the context of the point I believe you tried to make, that is meaningless. I am not suggesting that the mp3 decoding process doesn't create a similar sequence of numeric values (well, at least it's percieved by the human hearing system as similar, numerically it's very different), that's the point of a lossy codec. The point I am suggesting you are talking rubbish about, is the result of the mp3 codec can be converted to a state closer to the original by the use of further digital oversampling filtering.


This is not much different from the way lost data on an audio CD is recovered using the Sallen-Key method. I have test discs with holes drilled in them that play quite OK even though it is quite obvious that large chunks of data is missing. The decoding method on an audio CD allows for recovery of lost data.

No, it's entirely different, the Reed-Solomon error correction works through the process of redundancy, so the process can cope with lost data (in fact it doesn't, as it's not lost). But the introduction of redundancy increases the size of the data stream, so is counter to what mp3 tried to achieve. mp3 assumed that data is transfered without error and is not a correction method. BTW, I think you are confusing the name with the Sallen-Key filter introduced in 1955.


A similar method is employed in an mp3 decoder.

No, it's not, in fact even a simple lossless encoding process like huffman works by REDUCING redundancy.

Lets get this clear, and bipass the blather. Are you suggesting that given the following system:

Start with a sequence of 16bit words created by sampling a band limited audio stream at a rate of 44.1 thousand samples a second. (Lets take this as point A). Then pass this through a lossy perceptial coding algorithum (for example m3p), this produces a data stream, the bit rate of which is dependent of the algorithmic process. (we will call this point B). Then this data stream is processed by the matching decoder algorithum. The output of this process is again a 16 bit 44.1 ks/sec stream. (This we call point C). Then we pass this thround a upsampling digital filter that produces a data stream of a given bit depth, and a given sample rate. (This we call point D)

So, you seem to be saying that ther exists a system where information is present at point D, that was not at point C, but WAS at point A.

If that is your assertion, then I suggest you are wrong.

Beechwoods
08-09-2008, 07:04
:confused:

All MP3 does when it decodes is to approximate the data thrown away at compression. At its worst this means we hear artifacts that were never there in the first place. MP3 compression and therefore nasty sounding artifacts has improved over the years, particularly at high bit-rates but even I can hear the loss of space and definition on lossy-compressed material. But the rounding involved in the quantization of the data on compression means the numbers that are there after decoding are not the ones that went in to the encoding step. Exciters and plugins can add in missing harmonics outside those lossed on compression and might make it sound a bit nicer, but it won't be sound that was there on the original recording. Oversampling on playback of an MP3 digital signal will just reduce the affect of errors later on in the signal path at the point we go from D>A. It won't do anything to improve the original signal.

It is a complete waste of time to decompress MP3 and store it as FLAC. Why? The MP3 file contains everything needed to reproduce it's sound at the best possible realtime playback quality through an MP3 decoder. Bloating it to FLAC just creates a file 5 times larger with loads of empty bits...

I'm surprised we're even having this conversation here. The disadvantages of MP3 versus lossless compression is not some flat-earth thing. I thought it was a generally accepted truth. But at least we can all have our say; pub landlords, street cleaners and me, eventhough I don't have a degree in psycho-acoustical compression theory :)

purite audio
08-09-2008, 10:13
Does anyone have an adress for TEC manufacturing in Taiwan?
http://www.avartix.net/english/TC-7510-DAC.html

Yomanze
08-09-2008, 10:32
Does anyone have an adress for TEC manufacturing in Taiwan?
http://www.avartix.net/english/TC-7510-DAC.html

May I ask why? It confuses me as to why an 'ultra-fi' guy such as yourself would be interested in a re-branded product from Taiwan. IIRC, the DAC is Stanley's design and he outsources production to TEC.

http://www.technolink.com.tw/

purite audio
08-09-2008, 10:38
TEC is the manufacturer I believe.

Yomanze
08-09-2008, 10:46
TEC is the manufacturer I believe.

I believe so too...


This professional looking converter is designed by Stanley Beresford in the UK and manufactured for him in Taiwan with components sourced outside of China, Stanley has a low opinion of mainland quality control.

http://www.avreview.co.uk/news/article/mps/UAN/953/v/3/sp/

lurcher
08-09-2008, 11:27
Though I fail to see how any of the above manufacturing detail has any bearing on the topic of the thread.

Beechwoods
08-09-2008, 16:00
It does seem rather strange all round.

StanleyB
08-09-2008, 16:07
Though I fail to see how any of the above manufacturing detail has any bearing on the topic of the thread.
And so do I:scratch:.
If my memory serves my correctly, I was under the impression that AoS was aimed at attracting experts from different branches in the audio field in order to add a sense of purpose to the discussion, and raise it above the level so frequently seen to be the downfall of chats in the tent. At this rate, AoS is going to end up discouraging anyone in the trade to say nothing more than ' hello guys. What's spinning today?'

I rest my case and shall just pop in for a read of what others have to say. My two cent ain't worth the time it takes me to type it.

Beechwoods
08-09-2008, 18:17
I think there are folks here who want to have a decent discussion about these things. It's frustrating when threads go off-tangent but you gotta stay with it. Strong opinions and fervent debate are what happens when you get passionate people together to discuss these things. Evidence Marco's thread about SL1210's recently. While I may not agree with your view on this particular question, I was still interested to see what your response would be!

As far as I'm concerned I'm keen to hear from anyone with a view on things. Trade are welcome, as is anyone else. I thought that was what Art Of Sound was about...

lurcher
08-09-2008, 18:46
While I may not agree with your view on this particular question, I was still interested to see what your response would be!

Agreed. I think Voltaire's quote would be apropiate here.

colin.hepburn
08-09-2008, 21:44
Hi All
What have I started here a bit of debate gowning on I think most of this stuff is over my head:confused:
Lurcher
Says
Simple way would be mp3 to wav then wav to flac
So why not just Mp3 to Flac

Steve Toy
09-09-2008, 01:10
At this rate, AoS is going to end up discouraging anyone in the trade to say nothing more than ' hello guys. What's spinning today?'

AoS continues to respect experienced expert advice. Others are free to disagree of course, but the informed opinions such as yours are still likely to carry more weight :)

lurcher
09-09-2008, 06:53
So why not just Mp3 to Flac

Because you will find it harder to find a converter that will do that in one jump. I would still leave it as mp3.


AoS continues to respect experienced expert advice. Others are free to disagree of course, but the informed opinions such as yours are still likely to carry more weight

Fine, I won't bother in future.

Beechwoods
09-09-2008, 07:31
Ok. So on this thread, when someone states that MP3 compression loses information forever the 'informed opinion' is that 'it is incorrect to assume the information is gone forever'.

This is just not right.

I would be very keen to be walked through how quantization rounding error is not just one example of MP3's fundamental flaw: that information is lost forever.

I was hoping that Sensimilia would do this, particularly as his opinion is valued more than that of the rest of us.

If anyone is confused about what I'm going on about the best intro to how MP3 works is this one:

http://arstechnica.com/articles/paedia/the-audiofile-understanding-mp3-compression.ars

To answer the earlier question about MP3 to FLAC, FLAC is a great format, but only if the source has not been compressed lossily first. there may be software out there that does MP3 directly to FLAC, but it's not something to be encouraged, because FLAC is designed to compress without losing any of the original musical information, whereas MP3 loses some of that information, and no amount of transcoding it will restore what was lost when the MP3 was originally made. A FLAC made from an MP3 wastes hard drive space and sounds no better than the original MP3. It just doesn't make sense.

Mike
09-09-2008, 08:04
I would be very keen to be walked through how quantization rounding error is not just one example of MP3's fundamental flaw: that information is lost forever.

Beechy,

This is something that happens in the analogue to digital process rather than in the compression stage. All digital samples of an analogue signal suffer this to some degree or other.

For there to be no quantization errors you would need an infinite sample frequency and infinite word length. Is it necessary to got to such lengths? No.

One day someone will 're-invent' analogue; "hey look... a system which records a whole note instead on umpteen little slices of one"!

Cheers...

lurcher
09-09-2008, 08:52
Beechy,

This is something that happens in the analogue to digital process rather than in the compression stage. All digital samples of an analogue signal suffer this to some degree or other.

For there to be no quantization errors you would need an infinite sample frequency and infinite word length. Is it necessary to got to such lengths? No.

One day someone will 're-invent' analogue; "hey look... a system which records a whole note instead on umpteen little slices of one"!

Cheers...

Read the link that was posted, the quantization that is being talked about is part of the mp3 encoding, not the initial A/D. In the same way as with JPEG video compression, the result of the DCT is a vector or values (array in the case of JPEG), a quantization vector is applied to that, causing some of the values to be attenuated more than others, this adjusts what information is lost (forever).

Oh, and if it helps, I worked on a early video codec and transmission system in 1995 that used JPEG compression of images. MPEG wasn't usable in this case as the image needed to be of known quality and usable as stills (it was a surveillance system). But the basic maths was similar.

But what do I know.

purite audio
09-09-2008, 08:59
Everything I have ever read about MP3 states that once compressed the information is lost, gone for good, Stanley suggests otherwise, which is correct?

Mike
09-09-2008, 10:51
Read the link that was posted, the quantization that is being talked about is part of the mp3 encoding

Ah!... gotcha.



But what do I know.

Plenty.

Cotlake
09-09-2008, 21:32
Yep, I'll take a seat with Mike here. Nick (Lurcher) is enormously knowledgable. He doesn't overtly bang this out but thankfully is happy to share his knowledge and also to expose his own research for the benefit of others. He's not in the industry but certainly very much at the top of DIY R&D which in my own personal limited experience exposes the short comings of many profiteering commercial organisations.

Stan, I own your DAC and I'm very pleased with it. This post has no reflection on you or your product. I would however advocate to anyone interested to take notice of what Lurcher posts. He is no mavarik. Frequently at our fests, his kit comes out on top for performance. He is an interlectual man albeit happy to be naturally unconventional. He has time for everyone, regardless of their competance and I feel privaleged to be included in his attention albeit I fit at the lower end of the knowledge scale.

Nick has much to offer. It would be wrong to ostrasize his knowledge simply because you, as a manufacturer feel challenged by what he says. Any assumption that a manufacturer holds an authority over others that are not within the industry is misplaced. Nick is happy to contribute. I suggest we value that and maybe learn a thing or two from him.

Best wishes,

Greg

lurcher
09-09-2008, 21:55
I am touched Greg, I might be forced to buy you a pint some time :-)

Mike
09-09-2008, 21:58
I'd be more than happy to buy you both several pints! :)

purite audio
09-09-2008, 22:04
Delicately put Greg.

Marco
10-09-2008, 06:36
Let's all have a group hug :kiss:

No tongues, Mikey! :eyebrows:

Marco.

Tony Moore
10-09-2008, 08:09
Well said Greg, I agree 100%.

Please don't stop posting, Nick. There are many of us here who value your input and even when we say things that might be inaccurate and get corrected by you :) (I know _I_ have been) I'm sure we're all big enough to realise that it's a gift to have our understanding of something improved.

Cheers,
Tony

Marco
10-09-2008, 08:37
I don't think there's any danger of Nick no longer posting, Tony ;)

He is a most welcome and valued member (as is Stan). 'Experts' will always differ on technical matters - it's the nature of the beast, otherwise everyone would be building the same stuff!

I heartily agree, though. People of Nick (Stan and Anthony TD's) experience make this forum in terms of useful audio content a cut above most others (and I'm not talking about the WD or audio-talk forums).

Marco.

lurcher
10-09-2008, 08:41
<Father Ted>Steady Now</Father Ted>

Thanks folks, bit over the top though. Maybe I should go and check if I can change water into wine, I certainly know I can do it the other way round :-).

Marco
10-09-2008, 08:54
Thanks folks, bit over the top though. Maybe I should go and check if I can change water into wine...

Excellent. Can I place an advance order for 10 cases of Chateau Petrus 1964 Pomerol? :eyebrows:

Marco.

Mike
10-09-2008, 09:55
No tongues, Mikey! :eyebrows:

Oh pleeeease. Yuk! :doh:

Marco
10-09-2008, 10:40
And before you've had your lunch, too!

Marco.

Mike
10-09-2008, 10:53
Yup!

Actually I'm off work today so I'm doing some phono stage stuff before 'the house police' gets home.

Or before the world ends! :sofa:

Marco
10-09-2008, 10:56
Lurvely... Have you send off me 'thingy'? :)

Marco.

Mike
10-09-2008, 10:57
No, I'll do it today while I'm off though....

Marco
10-09-2008, 11:05
Ta, dude :)

We better watch this doesn't wander too far off-topic... ;)

Marco.

nat8808
14-09-2008, 03:10
Hi,

To chip in on the previous discussion, I have read about similar things before in different fields - I occasionally read New Scientist don't you know! (that was in jest rather than a boast, can't see a relevant smiley) - where a lossy compression method is used and the 'file' or whatever can be uncompressed with more of the original detail recovered than would perhaps be implied possible because of what was thrown away in the compression.

I certainly remember this appearing on Tomorrow's World, relating to image compression using fractal-based algorithms ages ago, early 90's perhaps.

The same type of phenomenon is touched upon in an article about the Meridian 618 dither processor in Hifi News in the 90's - Martin Colloms with Tony Faulkner and input from Meridian - where recordings made at 20 bit, dithered down to 16 bit and recorded to CD, when played back through a 20 bit DAC (still a 16 bit signal) portrayed most of original 20 bit resolution (whereas a simple truncation, not dithered, to 16 bit sounded like the usual 16 bit resolution). I think you end up with the equivalent of a 19 bit resolution according to the manual - alas mine doesn't work so i can't find out for myself.

The Meridian 618 part may be a red herring as that could be down to moving the noise floor, but then why can't there be a mathematical equivalent of 'the noise floor' when compressing digital files?

It is all down to mathematics and maths can appear to be illogical with illogical but none-the-less real outcomes. Logic is afterall just your brain judging things by previous experience so if it is out of your realm of experience it can certainly seem illogical and therefore unbelievable (and therefore never ever possible, if you are that way inclined..).

Stan seemed to be saying that you could recover better 'detail' from uncompressing the MP3 with the correct algorithm than you would get from simply playing the thing [through the MP3 standard's decoding algorithm]. I guess the 8 x times oversampling DAC part may be part of that process. If he has designed algorithms himself, I would personally take on board what he was trying to say and forget my own layperson or user's logic.

StanleyB
14-09-2008, 07:59
Stan seemed to be saying that you could recover better 'detail' from uncompressing the MP3 with the correct algorithm than you would get from simply playing the thing. I guess the 8 x times oversampling DAC part may be part of that process. If he has designed algorithms himself, I would personally take on board what he was trying to say and forget my own layperson or user's logic.
The problem is that I did work for various companies who could end up claiming commercial rights on products and ideas I use in any of the products I may have or wish to develop under my own brand name. So I have to be careful with what I say:eyebrows:.

nat8808
17-09-2008, 03:55
Is that the gist though, that certain methods of MP3 encoding can hold better information than the MP3 standard's decoding method can then extract? (This question seems to slightly change what I was saying in my previous post but that post reflected my thinking at that given time..).

Anyway, why do we need any file compression at all? Data storage is so cheap: 750Gb is ~ £65 per drive and holds 960 uncompressed, full 79 minute CDs! Buy two and run as mirrored drives for extra safety, or even three and have 1.5TB with RAID 5 safety.

StanleyB
17-09-2008, 11:58
Anyway, why do we need any file compression at all? Data storage is so cheap: 750Gb is ~ £65 per drive and holds 960 uncompressed, full 79 minute CDs! Buy two and run as mirrored drives for extra safety, or even three and have 1.5TB with RAID 5 safety.
I posted in another topic some details and a picture of my PC-less HD based system. The HD has a coax output and is the best thing since sliced bread. My DVD and CD players are no longer required.

Rick O
26-09-2008, 23:48
As I posted on The AudioWorks (http://theaudioworks.co.uk/forum/viewtopic.php?t=251) forum:


Lossless audio compression has always intrigued me, as I've always wondered whether it is really truly lossless, so what I thought I'd do is perform a little experiment; I would rip a CD to wav format, encode it into FLAC (an open-source lossless audio format), then decode that file into another wave file and compare the two files.

In order to make this experiment as fair as possible I used Exact Audio Copy to rip the CD to wave in the first place, then I would use the original FLAC codec with FLAC front end to encode/decode the files using the most accurate compression setting. Finally I would check the files md5 hash.

An md5 hash is a technique which can be used to check the integrity of a file on a computer. It uses cryptographic techniques to create what's called a "digest" of the file, and that digest consists of 128 bits (represented in 32 hexadecimal digits). If a file has been altered even in the slightest way, the md5 hash will change completely, thus enabling us (in this case) to see if the decoded file is a bit by bit representation of the original. Yes there are only a finite number of combinations with a 128 bit hash, but in order for two files to create the same hash they have to be completely different.

Anyway, here are the results:

Original Wave file: 9673440623ECD5BC65CCD4D960D08E91
Compressed->decompressed file: 9673440623ECD5BC65CCD4D960D08E91

Yes, the two files are completely identical, thus making FLAC lossless encoding truly lossless. Data can be lost in the ripping process, but the actual compression means it loses nothing.

Please discuss,
Rick

Beechwoods
27-09-2008, 07:13
What's to discuss? You're right!

lurcher
27-09-2008, 09:17
Data can be lost in the ripping process

And by the use of EAC you can see if data was lost, or at least if your copy of the data seems to be identical to what others got from the CD. by comparing checksums.

That reminds of a pet rant caused by some of the reviews in the last couple of issues of HFW (DP if you are reading). Reed Solomon error correction is able to PERFECTLY recover from errors, so the result is IDENTICAL to the original, it doesn't involve interpolation or any such guessing process. That gets involved for data loss that can't be recovered by the Reed Solomon encoding.

StanleyB
27-09-2008, 10:24
Reed Solomon error correction is able to PERFECTLY recover from errors, so the result is IDENTICAL to the original, it doesn't involve interpolation or any such guessing process. That gets involved for data loss that can't be recovered by the Reed Solomon encoding.
Tell them bro. The fact is that the data on a CD is riddled with errors. Without Reed Solomon error correction, the CD format as we know it would not have been possible.

Rick O
14-10-2008, 11:39
My next experiment is to stick some FLAC on my Eee PC and see how it sounds through one of the new dCS DACs. This will be interesting as Eee PCs do not contain hard drives, thus drastically reducing the amount of noise that is put on the output signal.

Filterlab
23-10-2008, 11:24
Would you prefer an answer from a street cleaner or a pub landlord? This ain't wigwam where experts are not allowed to comment.

Good call Stan, all input is welcome from whatever source.

The important thing to understand about MP3 is that the 'loss' data that's removed is data that the codec knows to have no or very little audible impact (for its common purpose). From that we can deduce that most (if not all) of the musically important information is retained in the data file, particularly if Variable Bitrate Encoding is used, and consequently when converted back to a lossless waveform which plays with no data constraints then the resultant output will be preferable.

However, moving from lossy compression back to lossless compression does not recreate the data that was removed as this is, for want of a better word, impossible. Still, I bet many ears wouldn't be able to tell the difference between the two, other than the fine audiophile members of this excellent forum who have hi-fi in their blood of course. ;)

Historicus
25-10-2008, 11:08
Flac to Wav and Wav to Flac conversion:
There is a nice Flac- frontend for windows or MAC on line that makes the conversion a job for dummies

http://flac.sourceforge.net/download.html

Flac or APE (Money Audio) are both LOSELESS compressors. The original is as the copy. It's quite truth as I could test it.

To create cuesheets from a cd I suggest to rip the cd with EAC (Exact Copy Audio) http://www.exactaudiocopy.de/ that has a great tool inside to get cd information from the remote freedb. So you don't need to write anything in the cuesheet.

If someone likes filesharing on this russian forum you'll find the links to high quality discs (cdroms, xray disc or DVD audi) http://netlab.e2k.ru/forum/index.php?showforum=175

Filterlab
25-10-2008, 11:27
My next experiment is to stick some FLAC on my Eee PC and see how it sounds through one of the new dCS DACs. This will be interesting as Eee PCs do not contain hard drives, thus drastically reducing the amount of noise that is put on the output signal.

Post your results Rick, the Eee PC interests me somewhat, particularly the solid state version. I wonder if Mac OSX can be installed on it. :) Certainly cheaper than the Macbook Air.

Beechwoods
25-10-2008, 12:37
Or...

Make your own solid state hard drive and remove the spinning one :)

http://news.cnet.com/8301-17938_105-9982978-1.html

Filterlab
25-10-2008, 12:43
Interesting, I'm not quite ready to go at my Mac with a soldering iron just yet though. :)

Beechwoods
25-10-2008, 13:14
Yeah, with an iMac these parts aren't so easy to get at! I have a PowerMac, a whole world of opportunity :)

Filterlab
25-10-2008, 13:16
Indeed, the iMac is a little restrictive to say the least, however I've just upgraded the RAM to 4gb which has changed the computer from 'simply very very fast' to 'immediate'. :) I'll probably change my Mac at the end of next year or the beginning of 2010, I'll go for a MacPro instead I think. I can start saving now. :lol:

nat8808
01-05-2009, 14:20
The important thing to understand about MP3 is that the 'loss' data that's removed is data that the codec knows to have no or very little audible impact (for its common purpose). From that we can deduce that most (if not all) of the musically important information is retained in the data file, particularly if Variable Bitrate Encoding is used, and consequently when converted back to a lossless waveform which plays with no data constraints then the resultant output will be preferable...

Obviously the codec is passive and doesn't know a thing (not meaning to be pedantic, just shifting people's thoughts in a certain direction) it depends on the designer of that codec as to what is removed.

Who are these designers and what do they understand about human perception of sound? Normally it is based on phsyco-acoustic research done many years previously by other people (the time it would take to translate research into a workable algorythm). It should be recognised that firstly no-one fully understands how our ears/brains (and the rest of our senses) work in relation to sound - new ideas pop up all the time - and secondly that the phsyco-acoustic phenomena that are considered to design the codec do not apply to all , they apply to a certain percentile of the people who took part in the research, in those lab conditions.

One such example of this is when two similar tones are playing at the same time but at different volumes. The brain tends to drown out the lower level tone and so normally can't be heard. These are deleted by most compression techniques. However! You can train your ear/brain to listen deeper and pick up the quieter tones and also these tones may combine with another close tone or harmonic of a tone so as to create beating effects that one can pick up etc etc etc. In stereo too, the relative levels of these tones and their phase can provide unconcious queues to their location. Basically, in the end you loose richness in the timbre of sounds and instruments, richness that is lost forever. This richness in stereo also relates to spacial info that will also be lost.

Still, I would agree that most people on most systems would probably not notice. Im not saying that I would either but would much prefer for all the detail to be there just in case else I may as well stop improving my hifi..