CD Ripping: Sound Quality Comparisons Between File Playback & Optical Disc Playback

[microstrip]

Sure, "some people" can say that, but they'd be wrong.
It's not "digital voodoo", it's just noise. FLACs take processing power to decode, while there's nothing to decode in AIFF/WAV. More processing power employed = more noise. Simple.
Using FLAC for your own files, or any kind of compression, defeats the purpose of having a high-end audio system.
Sure, there might be a day where DACs will be completely immune to this kind of noise. But until then, FLAC is pointless.

This one of the most ambiguous and misused arguments that spread on the net. If the main problem was just noise due to local processing power it would be easily solved technically.

I have often said any good digital designer can easily create a buffer in his DAC that stores and delays music for a few minutes. It would be a great test to servers to listen to music for a few minutes with the computer feeding the buffer or with the computer switched off after feeding the buffer.

Think of digital servers as a series of boxes with data inputs and outputs, connected in a chain arrangement - each output supplies data to next intput. The objective is simply that data in the last box - the DAC - should not have any vestiges of noise server. Or perhaps, just some kind of noise that manages to make the DAC sound subjectively better.

Sorry but in digital audio systems claiming that we get more analog noise in the DAC due to more processing power is voodoo. Competent design should be able to suppress or explain it technically - this means in Hz and V .

Yes, the performance of most servers is poor. The problem is that all those who claim SOTA performance seem to sound subjectively different, even if supplying bit exact data.

 

[asiufy]

There are extensions now to the WAV format to allow embedded artwork and metadata. Besides, Roon makes these issues completely irrelevant, as it'll recognize the album regardless of the format it's in.

The extra processing the Auralic needs to do to support Roon exemplifies the "noise" issue. MSB worked extra hard on their Renderer module, and wouldn't put it out until it performed better than the previous, non-Roon enabled Renderer.

 

[asiufy]

Francisco,

I was being simplistic on purpose. I stand by what I said. The reason a lot of transports still sound better than servers is, again, noise. A transport is a very simple electronic device, while a computer isn't, no matter how many corners you cut.

I didn't say "analog noise", I wrote "noise", BTW. There's a sound signature to that "noise", and it's clearly verifiable by using progressively better servers. I've experimented quite a bit in this area, and it's shocking how even extremely high-end DACs are still susceptible to this "noise" coming from the servers, even through Ethernet.

This is still a new area, and as I said before, at some point this "noise" will be nullified at the DAC end. But for now, minimizing it at the source matters.

 

[microstrip]

(...) MSB worked extra hard on their Renderer module, and wouldn't put it out until it performed better than the previous, non-Roon enabled Renderer.

Did you directly compare the performance of the MSB transport spining CD's with the renderer playing the same file data?

 

[asiufy]

Did you directly compare the performance of the MSB transport spining CD's with the renderer playing the same file data?

Yes I have. Different generations of transport as well.
The performance of the Renderer varied greatly depending on the server used to stream the files to it.

 

[Empirical Audio]

[QUOTE="Empirical Audio, post: 537541, member: 10089"
And just FYI, it's not the same thing to play a non-lossy compressed file and compare this to an uncompressed CD spinning. This is why I rip and play only .wav files. Your system has to be pretty refined to hear the difference though.
Steve N.
Empirical Audio

I was wondering if anyone was going to point this out. General belief is that there is no difference in sound quality between FLAC and WAV, but there certainly is! I discovered this myself several years back after running my own comparisons before ripping my entire CD collection to digital. The differences were actually easily dicernable to me on an average system then, which made me realize that as my system evolved and improved in future years, there would be even more of a difference. Why compromise sound quality for smaller file size... file storage space is cheap compared to everything else in this hobby![/QUOTE]
+1

 

[Empirical Audio]

Some people will tell you that if you notice differences between FLAC and WAV its a playback system fault ... Bit exact systems should be able to sound the same ...

I expect that as digital systems evolve differences that are only explained by digital voodoo vanish, and differences creeated to please listener preference become more deterministic. But it seems we not close to it, so we must pick what pleases us most at this moment.

This is not digital "voodoo". It has to do with the software CODEC used for decompression of compressed files and the effect of the software stack on this and perhaps on the D/A chip implementation. I have heard these differences in multiple systems of different types and using different playback apps.

On my own system I have even been able to hear the difference between AIFF, uncompressed FLAC and .wav.

Steve N.
Empirical Audio

 

[Empirical Audio]

I've lately been using the latest AURALIC Aries G2 feeding a Vega G2 using their recommend Lightning Link (Audioquest HDMI) interface. I was surprised to find a consistent difference between Roon sourced rendering vs Auralic's Lightning sourced rendering. The latter sounding consistently more "beautiful", like the sparkle of light on cascading water. AURALIC chalks it up to a difference in the processing of the signal and says the difference should become moot when they fully integrate the Roon processing into the Aries. Got me. But in the meantime, I use Roon to search, study, select, and casually listen, but will use Lightning for attentive listening.

I have compared Roon to my current app combination, Linn Kinsky/Minimserver/BubbleUPnP and I get better SQ with my apps using the exact same hardware. Some say that Roon improves with HQplayer in conjunction with Roon, but never tried it.

Steve N.
Empirical Audio

 

[Empirical Audio]

Francisco,

I was being simplistic on purpose. I stand by what I said. The reason a lot of transports still sound better than servers is, again, noise. A transport is a very simple electronic device, while a computer isn't, no matter how many corners you cut.

Quite true in some systems, but with Ethernet the computer is a don't-care. The app only ships data packets, just like a file transfer to a network disk. With Ethernet, I still hear a difference between .wav and FLAC.

I think your noise concept might hold water in some systems, but the real culprit I think is the CODEC and software stack. I view it a lot like digital filtering in DAC's or like differential common-mode noise rejection with USB. Works in theory, but no so much in practice.

Steve N.
Empirical Audio

 

[sefischer1]

I have compared Roon to my current app combination, Linn Kinsky/Minimserver/BubbleUPnP and I get better SQ with my apps using the exact same hardware. Some say that Roon improves with HQplayer in conjunction with Roon, but never tried it.

Steve N.
Empirical Audio

I've found the sonic difference I've found between Roon and AURALIC Lightning isn't subtle. I still love Roon though and popped for the lifetime subscription.

 

[Empirical Audio]

If you are using Ethernet/WIFI, here are some things that will make it sound much better:

1) use AQVOX switch to drive the Ethernet interface
2) use 0.5m Ethernet cable from AQVOX to EMO EN-70e isolator to 2.0m Ethernet cable to Ethernet interface
3) use Wireworld Platinum Ethernet cables

do these things and you will not believe the difference.

Steve N.
Empirical Audio

 

[tmallin]

As I mentioned in my initial post, I used FLAC with no compression to encode my CDs into computer files on my USB stick. The dBpoweramp program allows you to choose what kind of file you make and to elect various levels of lossless compression if you are encoding to FLAC. I chose no compression, just to eliminate or at least minimize any processing load on the digital electronics during playback. The file size of the CD as shown by dBpoweramp always matches the file size of the ripped files with this method. My goal was not to minimize the storage space occupied by the music files, just to copy them accurately.

I chose FLAC over WAV encoding because I did want to capture the metadata and associate it properly and easily with the music it pertained to. According the dBpoweramp's instructions, this can be an issue if you copy CDs as WAV files.

As to the "noise" some mentioned, in what way does this noise manifest itself? I certainly do not hear any higher background noise on playback of the digital files from the stick. If the bits of the copy match the bits of the original, how could frequency response or distortion added to the program by FLAC-generated "noise" be there without changing the bits in the copy?

 

[microstrip]

This is not digital "voodoo". It has to do with the software CODEC used for decompression of compressed files and the effect of the software stack on this and perhaps on the D/A chip implementation. I have heard these differences in multiple systems of different types and using different playback apps.

On my own system I have even been able to hear the difference between AIFF, uncompressed FLAC and .wav.

Steve N.
Empirical Audio

Sorry you just describing what I call voodoo - referring to differences without explaining the technical why's. Why exactly running a CODEC software program affect the sound quality? If it needs more power, we could just give it more power. If it adds noise we must suppress it. And as soon as you go towards DAC implementation it is also an unknown - there are so many implementations in DACs that is impossible to generalize.

Just because we are whispering Ohm's law while listening to differences does not make it scientific listening.

Have you been able to design a system that sounds the same with all types of bit exact files, compressed or not?

 

[microstrip]

If you are using Ethernet/WIFI, here are some things that will make it sound much better:

1) use AQVOX switch to drive the Ethernet interface
2) use 0.5m Ethernet cable from AQVOX to EMO EN-70e isolator to 2.0m Ethernet cable to Ethernet interface
3) use Wireworld Platinum Ethernet cables

do these things and you will not believe the difference.

Steve N.
Empirical Audio

Thanks. Have you tried the Netgear GS108 switch? In my system it sounded considerable better than the cheap TPlink it replaced.

BTW, I just ordered the EMO EN-70e isolator.

 

[Empirical Audio]

As I mentioned in my initial post, I used FLAC with no compression to encode my CDs into computer files on my USB stick.

FLAC is a lossless compression. Are you maybe using uncompressed FLAC?

I chose FLAC over WAV encoding because I did want to capture the metadata and associate it properly and easily with the music it pertained to. According the dBpoweramp's instructions, this can be an issue if you copy CDs as WAV files.

True.

As to the "noise" some mentioned, in what way does this noise manifest itself?

I think he is talking about noise in the computer, which somehow manifests itself in the playback data or offset. Offset, or the number of transitions before the start of the music track can vary depending on the ripping and playback software. I have a number of track with identical data but different offsets and they all sound different.

The difference I hear between .wav and FLAC is that .wav has a deeper more focused soundstage/image than FLAC.

I certainly do not hear any higher background noise on playback of the digital files from the stick.

It's not background noise as you would get with vinyl or tape analog sources.

If the bits of the copy match the bits of the original, how could frequency response or distortion added to the program by FLAC-generated "noise" be there without changing the bits in the copy?

As I mentioned, the offset could be different between the .wav track and the FLAC track. Also, if there is any real-time DSP performed by the playback app, such as volume control, this can easily change the bits. iTunes is notorious for mucking with the bits with DSP. This would not be evident if you do a digital compare of the .wav and FLAC tracks. It will only happen on real-time playback.

BTW, If I take any FLAC track and use dbpoweramp to change it back to .wav, it sounds much better, better imaging. This proves that the data is identical, but there are other factors affecting real-time playback with FLAC.

Steve N.'
Empirical Audio

 

[Empirical Audio]

Thanks. Have you tried the Netgear GS108 switch? In my system it sounded considerable better than the cheap TPlink it replaced.

I have not. I modded my own TPlink Router and it sounds almost identical to the AQVOX, head-to-head.

BTW, I just ordered the EMO EN-70e isolator.

Very good. Make sure you use a short cable from switch to isolator and longer one from isolator to DAC. Makes a difference.

Steve N.
Empirical Audio

 

[NorthStar]

I was wondering if anyone was going to point this out. General belief is that there is no difference in sound quality between FLAC and WAV, but there certainly is! I discovered this myself several years back after running my own comparisons before ripping my entire CD collection to digital. The differences were actually easily dicernable to me on an average system then, which made me realize that as my system evolved and improved in future years, there would be even more of a difference. Why compromise sound quality for smaller file size... file storage space is cheap compared to everything else in this hobby!

Steve, which one is best, FLAC or WAV?
Let me guess which is the better quality...FLAC ?

 

[tmallin]

Steve N. seems to prefer to optimize his digital systems by reducing the jitter going into the DAC as much as possible and then using a non-reclocking DAC. Others (most?) seem to have addressed the jitter "problem" over the years by developing/adopting reclocking DACs to get rid of as much jitter as possible which may have accumulated in the signal path, with the attack on jitter occurring right before the digital to analog conversion so as to avoid possible jitter injected by downstream cables, connectors, or electronics. While jitter may once have been a significant problem, according to testers such as Archimago (here) and John Atkinson (here) neither the Oppo UDP-205 nor the Benchmark DAC3 I'm using have significant problems rejecting jitter. Benchmark's own published jitter immunity results show its DAC3 to be quite immune to upstream jitter. See Graphs 15 and 16 on pages 61 and 62 of the manual here.

I do agree, with Steve N., however, that digital playback systems, despite bit-perfect playback and low jitter, seem to change in perceived sonic quality in response to various audiophile tweaks. Certainly sonic differences can be induced quite easily. Whether these differences rise to the level of "better" or "worse" is sometimes quite difficult to determine.

My most recent experience of such a change was moving my Auralic G2 streamer atop a Mapleshade Audio 4" thick solid maple platform sitting on Mapleshade Isoblocks instead of having the G2 sitting on a Bright Star Audio Little Rock. Both positions were centered on the same top shelf of the same Salamander Archetype rack. This mounting change caused a definite difference in sonics of programs heard through the G2. I think it's an improvement, but I'm not quite sure yet. Cause? Since research has shown how sensitive the human ear is to even tiny frequency response effects, my money would be on small changes in frequency response resulting from changing the mechanical resonances in the equipment. These surely are analog domain effects rather than anything to do with digital bits or jitter.

 

[microstrip]

(...) These surely are analog domain effects rather than anything to do with digital bits or jitter.

I would not be so sure. Mechanical vibration can induce harmonic distortions in analog, but will also induce jitter in digital - digital signals are supported by analog electronics, threshold in digital is an analog entity. Any induced changes will result in timing differences. And it seems that jitter effects are subjectively more perverse than analog distortions.

 

[3rdRock]

Steve, which one is best, FLAC or WAV?
Let me guess which is the better quality...FLAC ?

Sorry to disappoint you, but WAV presented the best soundstage and imaging. The FLAC files also added a slight sibilance to vocals with 'S' words. That's when I realized that if FLAC playback was messing that up, it was probably messing other stuff up as well. It was an easy decision for me to Rip all my stuff to WAV. I don't give a flip about Meta Data... I listen to music!