WAV vs FLAC vs AIFF

I stand corrected! I knew there was some issue with transcoding on the Synology server. I've been dabbling with Minim on my Synology over the last few days, just to see how it compares with the Unitiserve........

Simon-in-Suffolk posted:

Hi, assuming the ripper is not corrupting or altering the sound there will should be no difference.

just a point to stop an unnecessary rabbit hole being followed, different rippers can encode a wav file in one of two ways for Redbook PCM, and this results in different header formats being used, but the resultant media stereo sample data is identically formatted. From memory iTunes does encode WAV with a different header construct than the rippers I am aware of that insert metadata.

iTunes on PC used to also some sometimes have offset errors when ripping and this was dependent on CD reader used, so the first few milliseconds of a rip were absent.. but I'd be suprised if this is not resolved now.

Simon

My tests in comparing the payload after the header don't necessarily agree with that unless the discs are very new and pristine. Even for well kept older discs, deviations in rip quality crept in and in counting the number of 4 byte blocks that deviated from a reference rip that matched the Accurate Rip database I found that Mindows Media player rarely produced any 2 identical rips from repetative runs and iTunes also deviated to a lesser percentage. While the number of these deviated blocks is fewer than a tiny fraction of 1%, over the course of 1 second that amounts to a significant number of deviated samples. I found the differnces between a bit perfect CD rip using something like EAC versus a Media Player WAV rip to be audibly larger than the difference between that same bit perfect rip at 16/44.1 and a hi-res FLAC download at 24/192 of the same master. I'd go as far as to say rips of albums only a few years old on a drive os similar age via MediaPlayer sounded flat and lifeless. In fact it was noticing this sound difference that made me go and test this in more detail as opposed to test results impacting my perception of what I heard. While it only takes one bit in the payload to change the checksum of a whole disc (which would not be audible), rips that fail Accurate Rip generally have a high number of deviations.

The drive used for the ripping and its age also seems to impact the results greatly. I've written at length about these test a while back on the forum (over a year ago I am sure). My conclusion was that your could get bit perfect rips from standard software like iTunes or Media Player but that it was by no means guaranteed and depending on the disc and the drive, even moving to unlikely and the audible differences where there.

As a result, I'd always use something that verified the results against Accurate Rip so you can re-rip as necessary. Not every disc is in the database but if you have refined your ripping setup so that in general, a mismatched disc is extremely rare, then you can also be more confident that your rip is likely close to or at bit perfect.

I emailed innuos asking if the Zenith could be set to rip to WAV, their reply can be summed up as "not yet, possibly".

I've noticed that signals have sort of lost the HDX, NS01 and Unitiserve from their web site, do they know something or is their new site design the cause?

I won't be moving the CD player until January so can wait and see.

feeling_zen posted:

Simon-in-Suffolk posted:

Hi, assuming the ripper is not corrupting or altering the sound there will should be no difference.

just a point to stop an unnecessary rabbit hole being followed, different rippers can encode a wav file in one of two ways for Redbook PCM, and this results in different header formats being used, but the resultant media stereo sample data is identically formatted. From memory iTunes does encode WAV with a different header construct than the rippers I am aware of that insert metadata.

iTunes on PC used to also some sometimes have offset errors when ripping and this was dependent on CD reader used, so the first few milliseconds of a rip were absent.. but I'd be suprised if this is not resolved now.

Simon

As a result, I'd always use something that verified the results against Accurate Rip so you can re-rip as necessary. Not every disc is in the database but if you have refined your ripping setup so that in general, a mismatched disc is extremely rare, then you can also be more confident that your rip is likely close to or at bit perfect.

Indeed I always use Accurate Rip - but I have found with drives that are well calibrated that the rips are identical across rippers and platforms. A drive that is losing its alignment - typically at the end of very long disc (remember CDs read from the inside out) will show errors and Accurate Rip will fail its checksum.  I performed some tests with a file byte analysis program across a range of non new discs, but using modern CD drives, and the rippers I used returned identical results... in fact the tests that a few of us undertook on this forum are searchable on the forum.

If the disc is not in the database I get the ripper (and dBpoweramp can do this) to rip using two different methods and then cross check - if they match its usually a good rip.

Of course the other things to note - is that the odd unrecoverable data error - which will fail Accurate Rip -  usually results in a small block of silence or a click. I have yet to find a data error from ripping that subtly changes the media - its usually all or nothing - and the the nature of binary data - a flipped bit can have a significant overall effect that will sound awful, drive a ticking static type sound or drive silence.

ThatsNotMyNaim posted:

This weekend I borrowed and Innuos Zennith to demo in my system.

Once again I was assured there was no difference in sound between WAV / FLAC / AIFF.

Well I've slot loaded a few cds and then imported the same albums as WAVs and there is definitely a positive with WAV.

WAV sounds altogether more musical, natural and less separated. More swing and definitely more emotion which is the give away for both me and the missus.

AIFF makes the voices jump out in a horrible, uneasy way.

FLAC is all bright and lively. Loses smoothness and emotion over WAV.

Tested with tracks by Lene Marlin, Sade, Stacey Kent, Enya and Michael McDonald. 

Surely others can hear this? Innuos claim extensive testing found no difference im quality / sound repro of formats.

Rubbish.

Good system though, the Zennith. Improves sound quality. More assured, composed, fuller than the Mac / Audirvarna over USB & Optical.

Maybe the Micru Rendu would give it a run for it's money. 

And how did you test this - each track in 3 formats that you've purchased / downloaded?
Because any other comparison will be rubbish (as per your statement).

Adam

Adam, you can convert the file between WAVE, FLAC and AIFF and the results are just as valid when using UPnP.

Huge posted:

Adam, you can convert the file between WAVE, FLAC and AIFF and the results are just as valid when using UPnP.

Good to know Huge.

I should have phrased my question a little more precisely...

* Was the comparison made on each of the tracks from the artists listed, and each track then listened to in 3 different formats?

Adam