[Rusty Jefferson]

dB Cooper and Undertow,

Thanks for both your inputs. Definitely something for me to think about particularly dB’s advice.....

Some of dB's comments could be valid in general,  but aren't the reason why streaming services universally sound slightly duller and less dimensional than streaming from a good server.  I don't know the reason but can confirm what the OP asked, that the services don't sound quite as good.

WAV vs. FLAC should essentially sound identical......

WAV and "uncompressed" FLAC do sound exactly the same, if taken from the same source, as they are exactly the same file.  WAV and "compressed" FLAC can/do sound different depending on the source.
 

......FLAC is the all around winner and why most hi rez files are sourced from FLAC today vs. 1999 when WAV was a primary simple format from redbook.

No, FLAC is the winner with hi-rez downloads because the file size is smaller than WAV making it easier to store and transfer.  You should still uncompress your hi-rez FLAC files (and even 16 bit files) to full size after downloading/ripping. You don't want the server to "uncompress" files on the fly for best sound quality. 

[Digi-G]

There's some really good information on this thread.  Thanks everyone. 

I don't do streaming (yet) but have recently been converting my WAV library to FLAC.  In the process I'm tagging the files and adding album artwork.  Because of this thread I'm confident that I'm on the right track.

[undertow]

I guess because some are over complicating the reality in differences here...

The fact is END user equipment and storage of files are going to dictate how good, or high end you really can make any of these files sound.

1 - When possible highly suggest not using a full blown PC or Mac running standard operating systems, best to have a dedicated simple music server interface when possible. This also eliminates the noise and processors adding heat, and power draw from already insufficient power supplies for audio.

2 - Definitely try to use SSD drives and Flash drives. Again dumps excess power draw on already non-audio power supplies, gets rid of Mechanical movement, and noise which is similar to eliminating the CD transport in the first place and running static files off a hard drive. No real lag, buffering etc...

3 - Yes it is difficult to accomplish this because most "Audiophile" dedicated server components with a basic visual interface to access files can get ridiculous in price considering you can buy a near top of the line Laptop today for $400 to $700 and do your taxes, run audio, stream movies, and every other business, and gaming function on the planet when these stand alone machines with one piece of software, and basic USB interfaces can cost you in the thousands.

4 - I agree you need to source the FLAC files from a good lossless process, rip with DB poweramp, convert with DB poweramp etc...

[rollo]

Rollo, Do you use IPeng with a computer as a server/front end or with a streamer?
I like IPeng.

  No I use a music server. I stream Qobuz through the server. iPeng controls the library and selections.

charles

[gbaby]

There's some really good information on this thread.  Thanks everyone. 
n the process I'm tagging the files
I don't do streaming (yet) but have recently been converting my WAV library to FLAC. 

Why would you do that? You seem to be working backwards. If you import with dbpoweramp, you get cover art with both .wav and .aif files which are uncompressed. Flac is a lossy file whose compression maybe inaudible, but it is compressed just the same. I use .aif and now .wav.

[Saturn94]

Why would you do that? You seem to be working backwards. If you import with dbpoweramp, you get cover art with both .wav and .aif files which are uncompressed. Flac is a lossy file whose compression maybe inaudible, but it is compressed just the same. I use .aif and now .wav.

FLAC is lossless, not lossy. 

[gbaby]

FLAC is lossless, not lossy. 

Well sort of. Flac compresses and decompresses for space saving. However, as a purest, I say the less conversion, the better, so why bother with Flac when space is so cheap that you can import files in .wav and .aif? Again, with dbpoweramp, you get cover art for both, and to me, a better sound.

[zoom25]

FLAC is lossless is lossless is lossless. Period. There is nothing lossy about flac.

FLAC is compressed by default. Most ripping and transcoding software typically have FLAC set at level 5 as the default. With software like XLD, you can get uncompressed FLAC which will produce the exact size as WAV and AIFF. This way you can reap all the benefits of FLAC's near-perfect compatibility with metadata (whereas AIFF can sometimes have certain data fields appear incorrectly or entirely missing. WAV is obviously a further step down in this regard.).

You can go with FLAC compressed or FLAC uncompressed. I always recommend a master FLAC un/compressed library for the initial download/rip. A secondary WAV library can always be batch transcoded in the future for actual playback if needed. However, never rip or download in WAV from the start.

You can load and compare FLAC compressed, ALAC, FLAC uncompressed, AIFF, and WAV from either a singular device/location or cross compare different formats between different locations (USB vs. NAS vs. internal for those with BDP-2/3). You can use theoretical reasoning whether there is actually any difference or not. If there is difference, what are the overall impacts of decoding and I/O read operations between compressed and uncompressed counterparts. FLAC compressed decoding is more broadband/random. Uncompressed decoding is more predictable and has a clear processing pattern. What is better or worse in different regards? You can always your ears to come up with your own personal choice.

I do find slight audible differences between these different configurations that remain consistently present over years. What sounds 'better' or 'worse' can be subjective. I'm personally more interested in understanding 'what's the most accurate output option?'. However, this can only be conclusively answered with objective data and measurements. I hope one day a user or Bryston is able to produce measurements/graphs to conclusively answer this question. I feel confident that these differences are real and thus should be measurable. I'm not sure what suite of measurements will finally provide some insight to this story.

[CanadianMaestro]

For BDA-1 owners (and perhaps even -2 or -3), Bryston indicated that AIFF files were not optimal for their DAC, something to do with processing complexity.

Flac still sounds the best to my ears in my system. And yes, it is lossless.

 

[rbbert]

Where is the “bottleneck” in your processing chain?  If it your processor (CPU) then you may want as little data compression as possible to ease the load.  If it is data throughput (from whatever storage and bus you use), FLAC can be smaller and will go through faster.  Realistically, though, even hi-res music files are much smaller, and the data rate much less, than the video files that we watch (play or stream) and that gamers use; it seems unlikely that any of the file formats or storage types should matter, compared to the programs and hardware we are using to play the files.

[RandyH]

I remember when power amps measured near zero distortion but didn't sound so good.  Audio in general and music in particular is too complex to be solely analyzed by numbers.  The numbers can convey important information about what the audio information is but not what the music will sound like.  Just as quality musical instruments can have have subtle and nuanced differences in sound it does not make one right and the other wrong.  High end audio hardware and software is in effect a musical instrument as well.  Whether tubes vs solid state, digital vs analog, or flac vs aiff, there is no right or wrong but rather a matter of individual preference.

[gberger]

Amen!

[gbaby]

FLAC is lossless is lossless is lossless. Period. There is nothing lossy about flac.

FLAC is compressed by default. Most ripping and transcoding software typically have FLAC set at level 5 as the default. With software like XLD, you can get uncompressed FLAC which will produce the exact size as WAV and AIFF. This way you can reap all the benefits of FLAC's near-perfect compatibility with metadata (whereas AIFF can sometimes have certain data fields appear incorrectly or entirely missing. WAV is obviously a further step down in this regard.).

You can go with FLAC compressed or FLAC uncompressed. I always recommend a master FLAC un/compressed library for the initial download/rip. A secondary WAV library can always be batch transcoded in the future for actual playback if needed. However, never rip or download in WAV from the start.

You can load and compare FLAC compressed, ALAC, FLAC uncompressed, AIFF, and WAV from either a singular device/location or cross compare different formats between different locations (USB vs. NAS vs. internal for those with BDP-2/3). You can use theoretical reasoning whether there is actually any difference or not. If there is difference, what are the overall impacts of decoding and I/O read operations between compressed and uncompressed counterparts. FLAC compressed decoding is more broadband/random. Uncompressed decoding is more predictable and has a clear processing pattern. What is better or worse in different regards? You can always your ears to come up with your own personal choice.

I do find slight audible differences between these different configurations that remain consistently present over years. What sounds 'better' or 'worse' can be subjective. I'm personally more interested in understanding 'what's the most accurate output option?'. However, this can only be conclusively answered with objective data and measurements. I hope one day a user or Bryston is able to produce measurements/graphs to conclusively answer this question. I feel confident that these differences are real and thus should be measurable. I'm not sure what suite of measurements will finally provide some insight to this story.

You must be using free ripping software. I use to use free ripping. But since I paid for dbpoweramp, I get all metadata in both wav and air with covert art and all.  Space is no longer an issue so why Flac?

[gbaby]

 

[zoom25]

You must be using free ripping software. I use to use free ripping. But since I paid for dbpoweramp, I get all metadata in both wav and air with covert art and all.  Space is no longer an issue so why Flac?

Hi gbaby,

I've used all kinds of ripping software for CDs like EAC, XLD, iTunes, and Roxio Toast. However, I haven't used any CD ripping software in several years as I've switched exclusively to digital downloads. There is no 'ripping' going on.

My WAVs also have cover art and metadata as I choose INFO and id3 chunks. With software like Audirvana Plus, I can all the fields and cover art identical to FLAC counterpart. In Manic Moose, if you take a closer look at the playlist tracks on left, after the timestamp there is simply a dash "-" and then nothing afterwards. This is not the case with FLAC.

Furthermore, WAV with embedded metadata and cover art isn't always recognized correctly in all softwares. FLAC is handled far better universally. AIFF is practically there as well in compatibility on most devices and platforms. WAV is problematic in that regard. My WAV library with metadata and cover art did not do so well with Roon for example as it was creating issues with splitting a single album into multiple albums or having weird order of tracks. The original mixed FLAC/ALAC library on the other hand worked flawlessly with Roon and just about every software.

That is why I prefer to always have a master library of FLAC. As you point out, space and hard drive is cheap so an identical and secondary WAV library can be derived at any moment with batch transcoding in hours for use with playback if one feels they get better sound with WAV. The second library can also serve as a second/third backup. I hope that clears it up?

[zoom25]

Space is no longer an issue so why Flac?

Hi,

I also forgot to include that there are some Bryston BDP users and reviewers that have experimented and compared these formats themselves for sonic quality and found the compressed FLAC version to sound better than the uncompressed option. So for them, they are using FLAC not for saving space but for sound quality. Most BDP users do tend to prefer the uncompressed format sonically.

For example, the very last page of this 'The Absolute Sound' review of BDP-2 has the reviewer compare FLAC and WAV. He explains his reasoning behind it with some data as well as comments on exactly what he's hearing. He also reviewed Bryston BDP-1. Everyone has a different system and listening preference. With Karl's reviews, you can at least understand what he's hearing and where his preferences lie. Reviews that simply state X is better than Y without any explanation are rarely helpful. Karl typically prefers SPDIF, FLAC, and USB hard drives. He isn't fond of MPD with NAS drives or flash drives at all either.

http://old.bryston.com/PDF/reviews/2015_12_Review_BDP-2_Absolute.pdf

(Good job Bryston on fixing the old Bryston links )

On the other hand, there is Naim where both the company and users explicitly and unanimously prefer WAV over FLAC. They all use the on the fly transcode to WAV option.

[CanadianMaestro]

^ In other words, listen for yourself and decide what's best in your system.   

[rbbert]

http://old.bryston.com/PDF/reviews/2015_12_Review_BDP-2_Absolute.pdf

I had not read this before, but Kal makes basically the same points I mentioned in my earlier post (as have many others, I don't claim to have originated the ideas)

[zoom25]

If we're talking about 'true' bottlenecks, then FLAC has shown to be the more problematic one on BDP-1 with high-res material (24/192). I think Krutsch had issues with FLAC 24/192 (in big files?) and stuttering, however the larger WAV 24/192 played perfectly and you could skip along perfectly.

There's an older thread about CPU processing here:

https://www.audiocircle.com/index.php?topic=134583.0

WAV/AIFF do show lower CPU numbers. (I still have no idea how to look up individual Cpu processes in BDP rather than a singular percentage number in Services which isn't all that useful given it looks almost identical through all playback methods MPD, Roon, DLNA etc.).

Interestingly enough, Chris does say: Fat32 will have the least amount of overhead in comparison to NTFS or hfs+.  Using a samba share (NAS), will have even less impact on the CPU.

If I'm interpreting Chris correctly, playing the same file off of the NAS will result in less CPU load (and potentially noise) then playing the file from a USB drive? If I'm wrong, please correct me Chris. I'm not sure how the overhead and CPU processing differs between an internal drive and NAS.

[CanadianMaestro]

I've never had any playback issues with 192/24 Flac on BDP-1. Zip.
Methinks it may depend on library sizes that are attached or internal to the player?
And I was using an old version of MM.