[zoom25]

I've had my BDP-1 for sometime now. It has put my Macbook Pro and iMac running Audirvana Plus and Amarra to the side (both software optimized for playback). Better sound and ease of use with the BDP-1.

In the past I have compared WAV, FLAC, AIFF, and ALAC of the same tracks in Audirvana Plus through my Macbook Pro. I have 16 GB of memory on my laptop, i7 processors, and 500GB SSD. It's definitely not lacking. Audirvana Plus is supposed to load tracks into its memory, so theoretically there should be no differences between different lossless formats of the same tracks, whether compressed or not, since there is no on-the-fly conversion and decoding. Yet, I found WAV to sound the best. There was a noticeable difference between WAV vs. ALAC and FLAC. WAV and AIFF were very close though.

Today, I decided to do the same experiment with FLAC and WAV on the BDP-1. I found both to be quite good, but ultimately find WAV better. I was hoping this wouldn't be the case as it makes me spend time converting my FLAC library.

I posted my methodology and results on Head-fi when I did the test on Audirvana Plus. The sound impressions are described in more detail there. A few people agreed, many disagreed, some curious. Someone on a Naim system found the same preference as me.

I've even done this test, although in a less scientific manner, on old Windows machine. So across Windows, Mac, and Linux (BDP-1), I seem to not only notice a difference in sound between these formats, but always prefer WAV.

Gear used: Bryston BDP-1 -> Dangerous Music Source -> Sennheiser HD 800

[Mag]

I've been using wav for years!

I can only guess, the difference is small timing errors. For example when I ripped wav files I used jitter correction, it may be called error correction with other software. This takes significantly more time than a fast rip, but the difference is noticeable on a revealing system albeit marginal. What it does is butt byte sectors rather than overlap them slightly. The overlap creates these timing errors. Perhaps other codecs lack precision with overlap vs wav.

I use the BDP-1 as well.

[doveman]

I also use a BDP-1. I currently rip cd's via my Macbook Pro with XLD into ALAC as I use regular iTunes as my main music library and to manage the metadata, then sync with BDP-1 via samba

I figured that ALAC was the best compromise and shouldn't be any different from FLAC, my understanding is that ALAC simply has less compression applied compared to FLAC resulting in a bigger file but had made my songs compatible with Apple devices - i.e. for iPod playing in the car.

When you say that Audirvana Plus loads the tracks into memory, I would assume it simply loads the raw file into there but then when it's time to play it then reads from memory and converts via the CPU as it plays. Although I could be wrong.

Coming from Laptop (iTunes) > DAC via optical and USB to BDP-1 > DAC via AES/EBU I also believe there is a significant difference in sound.

[Ern Dog]

I've had my BDP-1 for sometime now. It has put my Macbook Pro and iMac running Audirvana Plus and Amarra to the side (both software optimized for playback). Better sound and ease of use with the BDP-1.

In the past I have compared WAV, FLAC, AIFF, and ALAC of the same tracks in Audirvana Plus through my Macbook Pro. I have 16 GB of memory on my laptop, i7 processors, and 500GB SSD. It's definitely not lacking. Audirvana Plus is supposed to load tracks into its memory, so theoretically there should be no differences between different lossless formats of the same tracks, whether compressed or not, since there is no on-the-fly conversion and decoding. Yet, I found WAV to sound the best. There was a noticeable difference between WAV vs. ALAC and FLAC. WAV and AIFF were very close though.

Today, I decided to do the same experiment with FLAC and WAV on the BDP-1. I found both to be quite good, but ultimately find WAV better. I was hoping this wouldn't be the case as it makes me spend time converting my FLAC library.

I posted my methodology and results on Head-fi when I did the test on Audirvana Plus. The sound impressions are described in more detail there. A few people agreed, many disagreed, some curious. Someone on a Naim system found the same preference as me.

I've even done this test, although in a less scientific manner, on old Windows machine. So across Windows, Mac, and Linux (BDP-1), I seem to not only notice a difference in sound between these formats, but always prefer WAV.

Gear used: Bryston BDP-1 -> Dangerous Music Source -> Sennheiser HD 800

I came to the same conclusion using my Salk Stream Player.  WAV sounds best and to rip CDs I use XLD .  I avoid using iTunes at all costs.

[gustavog]

Interesting, I think The Absolute Sound found FLAC better sounding on the BDP-2

[Wim J]

I also have the BDP-1  and am using FLAC Uncompressed which is more or less WAV with a FLAC-disguise. The FLAC Uncompressed files are even slightly bigger than WAV. Bitrates are the same. I can't hear a difference on the BDP-1. I could hear slight difference in a/b comparison between WAV and FLAC Lossless.
The main reasons for all sorts of compression were, expensive storage space and transport issues. Since both these factors are non-issues nowadays with multi-TB-HDD's, and Gbit-networks I don't see any reason to apply any kind of compression on locally stored music files.

[Grit]

A few years ago, there was a multi-part article from one of the audio magazines where two reviews compared WAV to FLAC and even files that had been ripped to WAV, converted to FLAC, and converted back to WAV. I searched for a bit but couldn't find it. Maybe if someone else has the link saved, they could post it?

I wish WAV had metadata. I specifically used FLAC because the file format does have metadata. Which brings up Wim J's post - do FLAC files without compression sound the same as WAV?

I tried comparing a few songs myself and could not hear a difference. I now have better quality components (all Bryston electronics now) and maybe should try again.

- Garrett

[Wim J]

A few years ago, there was a multi-part article from one of the audio magazines where two reviews compared WAV to FLAC and even files that had been ripped to WAV, converted to FLAC, and converted back to WAV. I searched for a bit but couldn't find it. Maybe if someone else has the link saved, they could post it?

I wish WAV had metadata. I specifically used FLAC because the file format does have metadata. Which brings up Wim J's post - do FLAC files without compression sound the same as WAV?

I tried comparing a few songs myself and could not hear a difference. I now have better quality components (all Bryston electronics now) and maybe should try again.

- Garrett

If I did understand it correctly FLAC Uncompressed = WAV + metadata header,... Please correct me if I’m wrong.
Big plus for uncompressed files is the lower processor load when encoding en decoding. Since this en/decoding is most of the time done by different machines (eg.: pc for encoding, streamer for decoding) each one can introduce errors. When the whole en/decoding step is skipped it eliminates a potential source of errors. YES, lossless formats are bitperfect but I have the impression they tend to mess up timing and phase of the original signal.

[zoom25]

I have been converting everything with XLD on Mac. It's perfect for batch conversion and maintaining correct folder structures and file names, and reading cue sheets from a single track. I also use it to attach metadata and tags to WAV. Only a few programs and players can recognize this. As a result, Audirvana Plus loads the exact same tags. album art, and metadata for WAV and FLAC, i.e. their functionality is identical. Still trying to replicate the experience on BDP through MPAD. Maniac Moose through external browser also lets you see the live variable bitrate of compressed lossless songs throughout the track.

Regarding which format is more memory and CPU intensive, I conducted an indirect test via Audirvana Plus. I combined audio tracks to be a single 2 hour track, in FLAC, AIFF, WAV, and ALAC. The test is to see how fast Audirvana Plus processes and buffers the track. This is seen by the track bar line getter darker as it buffers. ALAC was the slowest by far. FLAC came third, and AIFF was faster at second. WAV was near instantaneous. Literally no time spent on WAV. I could instantly skip along the track at any moment without a gap. So Audirvana Plus is indeed capable of doing pre-processing and playing from memory.

FLAC uncompressed vs. WAV is interesting as well. Haven't looked into it yet. Maybe I'll give that a try as well. However, at the same time I wonder why is AIFF still sounding different than WAV, as they are both uncompressed. WAV, AIFF, and uncompressed FLAC.

From last night comparing WAV and FLAC:
- FLAC had the same macrodynamic as WAV
- WAV had better microdynamics: you could follow trails, reverbs better
- Vocals are somewhat glossed over in FLAC
- WAV has a bit more soundstage depth than FLAC (only conducted on HD 800 last night)

On same tracks the differences are easier to pick out, on others its much closer. It definitely varies, but nonetheless it still exists.

[audioengr]

I have also done this comparison on multiple PC and Mac platforms, with Amarra, Jriver, Foobar, iTunes and others.  The wav always wins.

The only platform that I have found that delivers the goods with flac and probably ALAC is Antipodes servers.  I use them at shows whenever possible.

[R. Daneel]

There cannot be a difference between a WAV file and a FLAC file created from that very WAV. Numeric equations used in algorhythms will not allow this and there is no other state of being. I won't go into details of numerical mathematcis here obviously.

In short, it is either identical or something is out of order with the software handling the encoding and/or decoding processes. In addition, many things could and do go wrong during CD ripping.

[JimJ]

I've never heard a difference between a properly-ripped FLAC and the WAV.

[zoom25]

I've ripped CDs and converted through XLD, iTunes, Toast Titanium, EAC on windows, and all of them have been bit-perfect copies. The program always notifies you if there was a problem in conversion. Because of the checking, it does not matter if you start with WAV or FLAC or any other lossless format. You can convert back and forth as much as you like and the files will be identical. There is absolutely zero debate about that for me.

The issue for me is only playback where there should be absolutely no reason for differences in sound, i.e. files are buffered and converted into stream ahead of time. This should avoid noise and power consumption due to on-the-fly decoding. Yet, something is going on here that I am fairly certain is not my imagination.

Also, I did take into account various DACs (Dangerous Source, Emotiva DC-1, Macbook Pro's built in sound card) and connections (AES, USB, coaxial). Results hold up for me.

[Grit]

I have also done this comparison on multiple PC and Mac platforms, with Amarra, Jriver, Foobar, iTunes and others.  The wav always wins.

The only platform that I have found that delivers the goods with flac and probably ALAC is Antipodes servers.  I use them at shows whenever possible.

Did you find that to be true with WAV vs uncompressed FLAC also?

[audioengr]

Yes, I tried uncompressed FLAC also.  wav is better sounding.  Makes no sense whatsoever, but there you have it.  Did not try it on the Antipodes server.

Systems where this is not audible are just not to the level of my system.  Everyone is at a different level due to component and speaker selection, system design (preamp or not, transformer isolation or not etc.),  room acoustic tuning, cables and equalization etc..  A LOT of variables there.  My system has minimized digital and analog signal paths, room tuning as well as equalization using DSP.

[CanadianMaestro]

Very glad I don't hear the slightest diff between FLAC and WAV. Saves me time obsessing over it. 

[Krutsch]

Systems where this is not audible are just not to the level of my system.

Ahhh... there it is... .

[Anonamemouse]

Systems where this is not audible are just not to the level of my system.  Everyone is at a different level due to component and speaker selection, system design (preamp or not, transformer isolation or not etc.),  room acoustic tuning, cables and equalization etc..  A LOT of variables there.  My system has minimized digital and analog signal paths, room tuning as well as equalization using DSP.

So when was the last time you actually *enjoyed* listening to *music*?

[Johnny2Bad]

I am surprised by your findings, and can only surmise there is some issue with how the software handles each format on playback. There is no difference between any of the four formats as far as the audio data they incorporate ... they are all bit-for-bit identical.

AIFF and WAV differ only in the file structure ... they place the same information *about* the audio data differently. The audio data itself is, as I said, bit for bit identical. (Put a CD into a MacOS disk drive, and the files will be identified as AIFF files. Do the same on a WindowsOS machine, and those files are identified as WAV files. This is because the OS "wraps" the audio data with a few bytes describing the file type, so that other applications know what to do with the file).

Microsoft, as was it's habit, created an incompatible file format, and swapped the area of the file where this housekeeping data is kept, placing it at the front of the file preceding the audio data rather than at the end as in the Audio Interchange File Format (not an Apple format, as it is often but incorrectly described, but a cross-platform format originating in the early days of computing and used by IBM, Sun-Solaris, Amiga, Silicon Graphics, UNIX, the SONY digital recorders of the 1970's, dozens of permutations of DOS, etc).

Similarly, a non-lossy compressed format only describes the audio data differently, the data itself is identical to the non-lossy audio data.

(A crude description of a lossless compressed format is they describe a digital data section consisting of 111000011 as 3 1's, 4 0's, 2 1's ... and the descriptive language is more compact than the actual data written "longhand"). FLAC and ALAC are under the hood not much different than zip but have some optimizations designed to work better with real-time audio compression and decompression.

Technically there is no reason why you couldn't compress your music as zip format or your data as FLAC, but it just works better the other way around. That is also why there is very little difference in file sizes between the dozens of compression formats and why files don't get appreciably smaller if you compress them twice.

I have no explanation for your findings, as I said, there is zero difference in the actual audio data between any of these formats.

I also have no explanation as to why a decompression operation should have any difficulty with any of the above formats; FLAC and AIFF are open source, WAV is fully documented by Microsoft and although they do demand a royalty if used on a non-MicrosoftOS system or with a non-Windows application, it's not generally a roadblock. ALAC has been open-sourced by Apple so it's fully documented and free to use by anyone.

[zoom25]

Yeah, no arguments from me on the bit-perfect side of things. I was just listing my first hand experience.

I don't really care nowadays about comparing formats. I gave up on it. I use FLAC and WAV interchangeably. XLD on my Mac can convert an entire FLAC album to WAV in about 3 seconds. So I bulk convert with ease and use WAV for my reference hard drive. I still get all the info and art cover that I get from FLAC, without having to worry about any potential SQ differences in how a particular system handles it. It's honestly not a big thing to worry about. I did my experiments and found what worked best for me. That was that. Now it's all about the music. No more experiments (hopefully!).