Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[gazjam]

Hi,
Have big Flac library I now want to convert to Wav on my server.
My plan is to convert all my Flacs and keep the tagging info I spent 6 months of my life completing!
Wouldn't want to lose those...

I know Wav can contain Tags in JRiver, but my question is this:

Are the Tags EMBEDDED in the Wav files or "linked" to each file somehow in the MC database?
Just concerned if I ever lose my backup copy of JRiver library database all the "Linked" tags for my wav files will be lost.


If the tags are embedded in the file then I've no worries and can get converting.
Just wanted to double check before I did, thanks.

Also, if its a go-er, is there a "best" way to Batch Convert file format for multiple files?


Thanks for any assistance.

[InflatableMouse]

Playing wave cannot be better than flac, unless something is wrong. Flac is lossless and playback plays the exact same bits as when playing back the wave file directly. Basically, when you play back a flac file, you're playing the wave file.

Tags cannot be stored in Wav, the file doesn't support it. Flac does support tags though.

Before you spend a lot of time and effort converting flacs to wav, make sure you're doing the right thing. If you hear a difference between flac and wave, your mind is probably playing tricks with you.

[craigmcg]

I suggest that before you continue you do a fully level matched double blind test (Hydrogenaudio.org should have info as to how to do this) to scientifically, empirically confirm that you can tell the difference between WAV and FLAC. Most, if not all users on most digital audio sites will say that there is none and that this can be confirmed by converting a CD to FLAC then back to WAV and using bit comparators to compare it against the original WAV version.

As Inflatable Mouse said, WAV does not support tags (although seemingly MC has created workarounds within the library) so unless you have a rock solid back solution to protect both the files and the library, I think you are opening yourself to a world of heartache.

Personally, I'm not sure I can tell the difference between the FLAC and VBR v0 MP3 version of most of my library consistently. I'm certain I couldn't tell the difference between FLAC and WAV but YMMV.

[Vincent Kars]

WAV support tags but the standard contains very few tags of relevance for media players.
A couple of developers use the same convention to write IDV3 tags in WAV.
This convention is supported by
•   dBpoweramp
•   Foobar
•   JRiver Media Center
•   MusiCHI
•   MusicBee (partial)

In case of JRiver you have to install a plugin.


Bit-wise WAV and FLAC are of course identical.
It is thinkable that the difference in processing needed (WAV is almost raw PCM, FLAC has to be expanded first, WAV requires approximately 40% more I/O) does have an impact on sound quality as it can affect the timing.
If this is true it looks like you have a system very sensitive to input jitter.

Although I don’t like the ABX crowd at HA, I do recommend unsighted testing as we are often fooled by our perception.

[JimH]

In case of JRiver you have to install a plugin.

No.  You don't need to install a plug-in for WAV encoding.

[Vincent Kars]

As far as I know TAGGING WAV (not to be mistaken for encoding)  is off by default.
Right?

[Vocalpoint]

Playing wave cannot be better than flac, unless something is wrong. Flac is lossless and playback plays the exact same bits as when playing back the wave file directly. Basically, when you play back a flac file, you're playing the wave file.

+100.

To to OP - I would take caution in your findings (before you start overhauling your library) since it's not mathematically possible for a WAV to sound different from the same file in FLAC.

As a studio owner and sound engineer - I have done all the blind tests, bit tests, waveform tests, analysis and all the rest over the years. There is no difference when that sound comes out of the speakers.

For me personally - the horrendous tagging ability for WAV alone is more than enough for me to stay far away from it. FLAC wins this game every day hands down.

VP

[gazjam]

you misunderstand...
I KNOW Flac is identical to WAV in bit for bit terms, not suggesting its not.
Flac still needs to be uncompressed though.

It sounds better in my system so..to me its better.
This is in my opinion guys, not looking to debate it, thanks.

Storage is not a problem anymore really, so for me its ok to store larger WAV files.
I read the theory and knew it in depth I just wanted to compare it for myself.

A couple of mates heard it and agreed with it too.

Anyways....

The Flac I converted to Wav retained all my tags...I was just cautious if the tags were not stored with the file but were stored in Jrivers database somehow.

My concern is that if I reinstall and lose my jriver library/settings backup I will lose any tags attached to Wav files.

Thanks.

[Vocalpoint]

My concern is that if I reinstall and lose my jriver library/settings backup I will lose any tags attached to

I would think if you lost your MC library OR switched to a different player at some point in the future - your entire tag library will be compromised (or completely lost). Since the tags cannot be embedded in the actual wav file (outside of some real basic stuff) - you are risking your library from the start.

Considering the potential for ANY loss - personally I would never commit to wav no matter what you believe you are hearing.

But - hey if you feel strongly about wav sounding better - I guess you gotta do what you gotta do.

Cheers,

VP

[Matt]

WAV tagging is fully supported by Media Center.

But lossless is a better choice because of size, error detection, and more standard tags.

They sound the same.

[lasker98]

I store all my music files in .wav format. They were originally .flac files, fully tagged. I used dBpoweramp to batch convert the .flac files to .wav files. This was for a collection of 40k+ files. I've found zero issues with any of the tags in any of the converted files when played back using J River.
If you do have issues (which I can't see you having), you can always use dBpoweramp to reconvert your files back to .flac, and all your tags will be there.

Good luck,

Bill

[Vocalpoint]

I've found zero issues with any of the tags in any of the converted files when played back using J River.

Hmm...so I have like -30 tags per FLAC file using MC - and you are telling me that I can take this file, zap it thru dbPowerAmp to WAV and all 30 tags would automatically just "appear" in my wav file? And suddenly just appear if I turned around and played this new wav file in MC? Or in some other player? I seriously doubt it....

EDIT: Just tried it - and looks like about 10 tags made the trip (must be the standard defaults) so there ya go. I do see the stock tags in MC. Played the same file in Windows Media Player and see pretty much nothing.

If I attempted a conversion like this - my library would be seriously bent out of shape. FLAC it is...

VP

[lasker98]

70 tags? No idea. In my case I average 10 to 15 tags per file.

It would be interesting to hear your result after actually trying it.

[Vincent Kars]

Kinda baffled as how even this amount was possible

As stated above, dBpoweramp and JRiver both write ID3 style tags in WAV.
The WAV standard allows for INFO chunks and one can put anything inside it.

[Vocalpoint]

As stated above, dBpoweramp and JRiver both write ID3 style tags in WAV.
The WAV standard allows for INFO chunks and one can put anything inside it.

But ONLY a very small standard subset...and NOTHING custom to MC can be written out to a WAV file. I just tried it.

Since the custom work is the bread and butter of my library...FLAC is my standard for the foreseeable

And I am thrilled that it sounds identical to the actual CD.

VP

[Matt]

WAV uses ID3v2, which supports ALL tags (both standard and custom).

[Vincent Kars]

Using JRiver I converted a FLAC to WAV
Looks like all custom tags carried over

Uncompressed Audio File (wav)
44.1 kHz, 16 bit, 2 ch

ID3v2.3 Tag: (282624 bytes)
  TIT2 (Name): Brandenburgische Konzerte Nr. 3 in G-Dur: Adagio - Allegro
  TPE1 (Artist): Frans Brüggen; Anner Bijlsma; Gustav Leonhardt
  TPE2 (Album Artist): Gustav Leonhardt
  TALB (Album): Bach,Js Brandeburg Cto #3 BWV1048 in G Leonardt
  TRCK (Track #): 9
  TPOS (Disc #): 1
  TYER (Year): 1980
  TCON (Genre): (01.02) Orchestral Music>Concerto Grosso
  TCOM (Composer): Bach, Johann Sebastian (1685-1750)
  APIC (Picture) (Cover): <too large to display>
  COMM (Comment): 01.02.01 Baroque/Sinfonia\n
  TXXX (Album Artist): Gustav Leonhardt
  TXXX (Composition): Bach JS, Brandenburg Concerto #3 BWV1048 in G-M; Leonhardt
  TXXX (Movement): Adagio - Allegro
  TXXX (Opus): BWV 1048
  TXXX (Series): Bach_BWV 1048
  TXXX (Tool Name): Media Center
  TXXX (Tool Version): 18.0.48
  TXXX (Work): Brandenburg Concerto No. 3 in G major

[Vocalpoint]

WAV uses ID3v2, which supports ALL tags (both standard and custom).

But - your assumption is that you are using MC to create the actual wav FROM the FLAC. In my first test - I used EZCD Extractor - and no custom tags came over.

I will almost guess that if I do the same from MC - the custom tags will make the trip.

I have now tested a FLAC taken from MC and then converted to WAV via dbPowerAmp - and yes - all the tags (even the custom ones) came thru. Impressive.

So I do stand corrected on my opinion regarding tags in WAV (but ONLY when dbPowerAmp or MC is involved).

However - This nice tag transfer does NOT occur with other tools in use out there - even if the FLAC file (and it's tagging) originates within MC

Examples with some other tools I use very frequently....

1. Trader's Little Helper - wipes ALL tags (in ALL fields) when converting FLAC to WAV
2. EZ-CD Extractor Pro - only 12 standard tags are retained

So while - it's nice to know the tags can be preserved under specific conditions with the right specific tools - I still would not abandon FLAC at any point.

VP

[Sop1]

Sounds like the OP should use an archived library in flac as his master file to stay current with format tweaks and then convert a copy, as needed, to WAV for daily use for listening.  Kind of backwards..... instead of keeping a backup of your main file which is use everyday and vulnerable to damage, you would be safeguarding the prime data file and using the WAV copy as a daily driver.

[Vocalpoint]

Sounds like the OP should use an archived library in flac as his master file to stay current with format tweaks and then convert a copy, as needed, to WAV for daily use for listening.  Kind of backwards..... instead of keeping a backup of your main file which is use everyday and vulnerable to damage, you would be safeguarding the prime data file and using the WAV copy as a daily driver.

Or just maybe keep his library as is?

He is already set with a perfect sound in a perfect format with perfect tagging. But wants to disassemble it all - and use twice as much space and 40% more I/O (according to Matt)? I don't get the point...but hey - it's not my library. Whatever floats your boat I suppose...

VP

[gazjam]

Sounds like the OP should use an archived library in flac as his master file to stay current with format tweaks and then convert a copy, as needed, to WAV for daily use for listening.  Kind of backwards..... instead of keeping a backup of your main file which is use everyday and vulnerable to damage, you would be safeguarding the prime data file and using the WAV copy as a daily driver.

hi OP here.

Now that storage space isn't really an issue any more, I'm quite happy to "uncompress" my Flacs to Wav,) especially as (yes:)) I can hear a difference.
S'pose I consider myself an audiophile and my system is pretty transparent and open, so if I can get a slight improvement for free I'm happy.

I tried this experiment before in an older version of my system and didn't hear a difference but can hear one now.
As I said, not night and day..subtle but definately there.

I only ever use JRiver for media so it seems using it to convert my Flacs would cause no problems.
Gonna be sure though and create a bunch of tagged Flacs and copy them on to my other PC and install JRiver on that. If the Tags are picked up - problem solved!

I dont really see it as working backwards either really...have uncompressed files for listening, compressed versions for backup storage.
I think that's what Chris Connaker over at ComputerAudiophile suggests, dunno.

[gazjam]

Or just maybe keep his library as is?

He is already set with a perfect sound in a perfect format with perfect tagging. But wants to disassemble it all - and use twice as much space and 40% more I/O (according to Matt)? I don't get the point...but hey - it's not my library. Whatever floats your boat I suppose...

VP

its an audiophile thing I guess
Storage ain't an issue and I've a fast server (3750k cpu @ 4.5Ghz, lots of ram etc)

Perhaps an odd example but I remember from my Squeezebox Touch days that setting Flacs to be processed from the Server rather than the Touch itself was highly recommended as it sounded better, reason being the heavy lifting was being done by the server and not the limited resources of the Touch itself.
I think this highlights that there is a bit of work required in processing Flacs to PCM?
I'm not saying that there is the same knife edge use of resources at play here, but would I be right(ish) in saying there is an extra layer of processing in converting Flac to PCM in the PC domain?

I know the argument about this vs the increased I/O playing WAV, but is this an issue with modern speedy servers?

Perhaps the "heavy lifting" isn't about the cpu/memory resources as with the SB Touch, but an extra layer of software processing AFFECTING the final result somehow?
Genuinely curious about this.

[Vocalpoint]

I dont really see it as working backwards either really...have uncompressed files for listening, compressed versions for backup storage.

Please note that FLAC is uncompressed when listening so the actual end result is identical.

Continued success with your project

Cheers,

VP

[gazjam]

yeah, but its uncompressed on the fly...   
not sure how memory play affects this though?

[struct]

Without commenting on the flac/wav debate....

Is there something "wrong" with the system that is making it so "senstive" to the flac/wav source, that you may be better off finding the offending culprit and this may yeild even further sonic improvements.  

If the system is so sensitive to whether the file was wav or flac, it might suggest that other more beneficial tweaks to the system might be possible.  Maybe change operating systems?  Maybe dac/computer aren't workin well together??

Craig

[Vocalpoint]

yeah, but its uncompressed on the fly...    not sure how memory play affects this though?

With the quality of JRiver - it should not effect anything since MC is sending the same PCM stream as would be sent if a WAV version of the file was playing.

Plus someone mentioned earlier in the thread that WAV requires about 40% more I/O - so if that's true - it's WAV that is taxing the system more. Not sure if anyone would ever be able to "feel" that...but there ya go...

VP

[srwooten]

Have you considered using uncompressed FLAC? Best of both worlds.

[Vincent Kars]

I’m afraid memory playback as implemented in JRiver won’t help us.
It is an input buffer. The track is read and stored in memory.
The decoding is done during playback.

Memory playback as implemented in Foobar/Cplay/ SignalistHQ read the track and do the decoding on the fly. What is stored in memory is an output buffer with “DAC ready” output.
This eliminates both the differences in I/O and processing regardless of the format.
This is the trick to eliminate any possible audible difference between (lossless) formats.

[Phil LD]

When using an USb 2.0 asynchronus Dac with an internal reclocking system, you can't hear a difference, when not doing so, you can maybe.
Any additional process that is lossless will only effect time, no other dimension. If any time issue is corrected at the end, there can't be a difference

[JimH]

Have you considered using uncompressed FLAC? Best of both worlds.

FLAC, uncompressed FLAC, or WAV would give you the same sound.

[gazjam]

Jim,
I absolutely love JRiver, audio wise I feel it is a lot better sounding than any of the others, and I've tried most of them.
Thats why you'll get my continued support.

That said, if I can tell the difference between software players, maybe there's something wrong with my system?  

[InflatableMouse]

That said, if I can tell the difference between software players, maybe there's something wrong with my system?  

No, there's probably something wrong with the way you test which leads to you hearing something that is not there.

[JimH]

A few posts were split to a new thread:
http://yabb.jriver.com/interact/index.php?topic=75123.0

[audioriver]

I’m afraid memory playback as implemented in JRiver won’t help us.
It is an input buffer. The track is read and stored in memory.
The decoding is done during playback.

Memory playback as implemented in Foobar/Cplay/ SignalistHQ read the track and do the decoding on the fly. What is stored in memory is an output buffer with “DAC ready” output.
This eliminates both the differences in I/O and processing regardless of the format.
This is the trick to eliminate any possible audible difference between (lossless) formats.

I wish MC did just that...

[gazjam]

No, there's probably something wrong with the way you test which leads to you hearing something that is not there.

uhm, and with that...I'm out. 
Cheers guys, will follow the thread.

i can confirm that Wav files created from Flacs in JRiver retain their tags, even when used in a different JRiver install.
Impressive work Jim and the rest of the guys.

[InflatableMouse]

uhm, and with that...I'm out. 
Cheers guys, will follow the thread.

i can confirm that Wav files created from Flacs in JRiver retain their tags, even when used in a different JRiver install.
Impressive work Jim and the rest of the guys.

I apologize. I should have given it more consideration before posting something so blunt. Also, I should have read your message better because I actually agree, I read something else and replied without thinking.

[gazjam]

I apologize. I should have given it more consideration before posting something so blunt. Also, I should have read your message better because I actually agree, I read something else and replied without thinking.

No harm done, thanks for that.
Just didn't want to get in to the whole debate.
been there done that. 

[desa]

gazjam, have you considered a client-server setup with the server doing the uncompressing work?

It should be very easy to set up using the client-server capabilities of Media Center.

This way you can leave your library in a more standard format than WAV with embedded tags and, at the same time, play a plain, uncompressed, PCM stream.

[MetalHOE]

Uncompressed flac is the solution.

[JimH]

Ordinary FLAC would be a better one. 

[AndrewFG]

Tags cannot be stored in Wav, the file doesn't support it.

This statement is incorrect.

Wav files consist of "chunks"; the first three chunks - namely the "RIFF" chunk (containing the file format and size), the "FMT " chunk (containing the audio format), and the "DATA" chunk (containing the pcm audio stream data), are obligatory, but the file can include additional "INFO" and/or "ID3 " chunks containing meta data.

[gazjam]

Ordinary FLAC would be a better one. 

Why Jim?
(genuinely interested in your view)

I know you said no sound difference (or bit difference) between Flac and Wav, is this why?

[gazjam]

gazjam, have you considered a client-server setup with the server doing the uncompressing work?

It should be very easy to set up using the client-server capabilities of Media Center.

This way you can leave your library in a more standard format than WAV with embedded tags and, at the same time, play a plain, uncompressed, PCM stream.

I'd like to try that Desa, but unsure how.
Can you point to a Wiki entry or link somewhere?

thanks.

[JimH]

Why Jim?
(genuinely interested in your view)

I know you said no sound difference (or bit difference) between Flac and Wav, is this why?

Zero difference in sound, standard tagging, more widely usable.

[Vocalpoint]

Zero difference in sound, standard tagging, more widely usable.

Not to mention more space, more robust error checking and excellent archiving/backup (again - more space) .

And yes - coupled with an MC client connecting to an MC "server" - one can simply have MC "serve" up the FLAC as an uncompressed wav and then best of both worlds - FLAC on the server to WAV on the client - it's all good.

VP

[gazjam]

Not to mention more space, more robust error checking and excellent archiving/backup (again - more space) .

And yes - coupled with an MC client connecting to an MC "server" - one can simply have MC "serve" up the FLAC as an uncompressed wav and then best of both worlds - FLAC on the server to WAV on the client - it's all good.

VP

I run JRiver from an SSD computer pulling in Flacs from server in another room via ethernet.
Is this the same thing as you are saying?

You say "Wav on the client"...how do you mean?
(If its different from setup I currently have I'd like to look into it)

thanks.

[Vocalpoint]

I run JRiver from an SSD computer pulling in Flacs from server in another room via ethernet. Is this the same thing as you are saying?  You say "Wav on the client"...how do you mean?
(If its different from setup I currently have I'd like to look into it)

First - you would have to clarify if you are accessing the FLAC files directly via the MC library on your computer (you have set a specific folder on the server for MC to use/scan/import the music files for it's library)

In a client server setup - like I use - my main office workstation is the master "server". This machine runs MC so I can use it for listening on the desktop, it has a direct link to the FLAC audio files (which are all located on a Windows 2008 storage server in my machine room) and it runs the all important MC Media Server component - which allows other MC "client" instances around the house (4) to connect to this master server and access the master library.

It's the Media Server component configuration (on my desktop machine) that allows you to choose to have the "server" send out the audio as uncompressed wav to a client machine.

Let us know how you are accessing the FLAC first and we can take it from there...also - if this is the one (and ONLY) installed MC instance. If it is - client/server may or may not be an option for you.

Cheers

VP

[desa]

I'd like to try that Desa, but unsure how.
Can you point to a Wiki entry or link somewhere?

thanks.

Steps:
1) Install Media Center on the server and import all your music in MC library, as usual
2) Go to Options->Media Network and check "Use Media Network to share this library and enable DLNA
3) In the same page, click on "Access Key: Click for access from another Media Center, ...". After clicking you should see an access key. You will need it in step 12
4) Click OK
5) On Client system, go to "Options->Media Network" and click "Client Options (when connected to a Library Server)" to open that section
6) Uncheck: "Play Local file if one that matches Library Server file is found" and "Show playback zones from the server on the client"
7) Click "Audio Conversion" to open that section
8 ) Set "Conversion: Always convert audio" and "Encoder: Uncompressed"
9) Click OK
10) Go to "File->Library->Library Manager" and click on "Add Library..."
11) Choose a name for the new library and check "Library Server"
12) Put key from step 3 in "Access Key or URL" field
13) Click OK

Now your client library should be populated with items from the server library. If you play a track, you should see a fixed bit rate corresponding to the uncompressed bit rate of your track (in kbps). For a CD track: 16 bit x 2 channels x 44.100 samples/sec / 1000 = 1411 (it's 1411.2, but MC rounds it).

This setup works as long as you play only 16 bit tracks. If you have 24 bit tracks you should choose "Uncompressed - 24 bit" at point 8. I wrote you should because if you choose "Uncompressed - 24 bit" everything becomes very very slow. I tried it yesterday, after advising you, and I discovered the problem.

I started a topic on this. I hope to get a feedback (and possibly fix) soon.

Another way may be using Asset UPnP as a server, configure it to uncompress FLACs on the fly and point the client to it. It works without any of the problems described above.

Hope it helps.

Marcello

[desa]

It's the Media Server component configuration (on my desktop machine) that allows you to choose to have the "server" send out the audio as uncompressed wav to a client machine.

When you use native jriver library server, there is no server side setting to force on-the-fly decompressing. The "decompress" flag is set on the client (Tools->Options->Media Network->Client Options) which then sends it to the server when requesting a file, forcing on-the-fly decompressing.

In any case, this setup cannot be used with 24 bit contents, which makes it simply useless.

[gazjam]

Ah, thanks for the info guys.

Unfortunately I have a lot of 24/96 music which I cant do without. Ah well, maybe there will be a solution to 24bit slowing down at some point.

For my own setup, I have JRiver installed on my SSD machine in the listening room, pulling Flac files from a server via ethernet.
Read up on the server/client side of things..interesting!

If I could shift 24bit files *I have 24/192 as well..) then I'd maybe experiment with it a bit.

*Update on Flac/Wav tags*
Just noticed than a custom tag "Styles" I created (to allow multiple genre's for a single song) did not get transferred over when converting to Flac in JRiver.
Would need to manually copy/paste tags for each Artist...bit of extra work, hmnn..