I think what you're saying is pretty reasonable (especially when it comes to listening to what you like).  However, if bit rate errors were a factor, then computer programs would crash far more often than they do - or maybe this is Microsoft's excuse 🙂.  For every "bit error rate", there's an "bit error detection and correction" algorithm.  If not, digital technology just wouldn't work.  I'll accept that bits can get lost in the DAC (who's to know, it's being converted to analog), but not that bits are routinely lost and not corrected in the digital side of the computer.

Stan

If we accept that a computer can open a *.zip or *.rar file without any errors and all data contained in that file is as it was before being compressed why do we consider there to be some special "hoodoo" in a compressed audio file that means that "lossless is not lossless" - I find this incredible!

There are many tests elsewhere on the net that show that the digital data recovered from a lossless audio file (flac, alac, or wma) is bit perfect when compared to the original source. There are also tests where the analogue audio files generated from the original CD and from playback of a lossless codec through the same soundcard/dac are subtracted - the result is, in essence, single-bit noise i.e. the analogue waveforms were identical.

Of course people can choose to believe or disbelieve, this sublect will continue to run and run across audio forums throughout the net however for all the tests I've seen that show lossless is lossless, I can't find any that stand up to scientific scrutiny that show lossless isn't.

Hifi mains cords anyone?

I promised to myself I won’t get drawn into this type of argument but here it is 🙂  Audio or video information is transfered to a digital media at a certain resolution and speed (bit rate per second and sampling rate).  This concept applies to time domain based signals, mostly those intended to be received in real-time by a consumer (such as voice-over IP or digital mobile phones).  A ZIP file opened on a computer does not fit in this category 1. because it is not a time-based source of data and 2. because it can be error-checked and error-checked and error-checked as long as it takes for the hourglass on the computer to disappear (we should all be accustom to this scenario as long as you have used a computer before).  The other misconception around digital data transfer being lossless is prompted by the internet.  Most people think if you send a file from A to B (such as an email) it will be received by the recipient as it was intended by the source (hence, a perfect transfer).  That assumption actually proves to be correct.  There are a series of network protocols those computer geeks invented (such as TCP/IP) which prevent the data from being lost or degraded.  That concept could even be extended to wireless protocols (such as data transfer between two Airport Express routers).  But, back to the key point, in these scenarios data does not have to be transfered in real-time!  You wait for it, it gets re-sent now and then, it gets buffered and sometimes it even disappears in the “cloud” of hardware and software that exists between the two parties in concern.  In the AV world the concept of buffering (due to size limits and time-boxing) faces constraints for real-time data reproduction and it is also affected by the different OSs (Windows, Mac, Linux etc) as adopted by different brands of digital sources.  B&O happens to use Windows XP Embedded for their media player dubbed BS5.  Lossless is lossless at the source, that is, the disk in your hand, the piece of data stored on your hard-drive (FLAC, WMA or AAC whatever it may be).  Once it leaves the media it goes through a process of: decoding for playback (different codecs, despite the fact of being both lossless, have different latency between source encoding and playback and can achieve the status of lossless at different bit rates BTW, which may have an impact on buffering?), transfer via well insulated high conductivity cables or lower quality cables (subject to one’s setup), conversion from digital to analogue (via a low quality or high quality DAC), and so forth.  Over-buffering or interference results in bit / packet data drops (to ensure the signal is still presented in real-time).  That results ina lower resolution output, or color degradation, or contrast loss etc. The transport process is definitively lossless (unless it uses network-based  protocols and the buffer is large enough for the streamed data signal)!  

Hence my earlier point: you could experience a better system throughput with lower resolution media files than you could experience with higher resolution files (as some have pointed out in this thread unless the perception factor plays an additional role which certainly clouds the entire argument).  Hence the preference of AAC 256 Kbps over a full resolution CD / WAV.  It depends on the weakest link in your system.  If the system is capable to consume the full resolution signal in real-time there is no doubt the output will be superior to the lower resolution source.  But that is not that easy to achieve and many home theatre setups out there have not reached that ideal capability.  Hence the reservation of B&O going Full HD / Blu-ray too early (despite all the marketing hype stating you can do that with little money).  The technology is not mature enough to support that "lossless" type of playback unless you pay the big bucks to upgrade or improve it.  Two source files being lossless (and with different encoding algorithms) doesn't translate to identical playback!  

I don't think we're beating a dead horse quite yet so I'm going to respond. 

First, I think you have a typo...  "The transport process is definitely lossless (unless  it uses network-based protocols... )".  I'm pretty sure that you mean "lossy" (the opposite of lossless) since the rest or your statement talks about the effects of losing data (lower resolution, color degradation, etc.).

Secondly, I agree with most of your facts, just not all of your conclusion.  To paraphrase (and please correct me if I'm misinterpreting your words), your point is that because AV is "real-time", you can lose data because the algorithms/transport run out of time or buffer space to do appropriate error correction.  This may be true with high resolution video such as Blu-Ray as rather intensive processing is required, and perhaps this is why you changed the subject more toward video.

However, I disagree that these factors have any impact on audio (at least the 2 channel CD resolution audio within a self contained system like the BS5) because, with today's hardware, the processing and data volume required for audio decoding is miniscule.  For example, my 1.42GHz PPC MacMini can convert a 4 minute audio ALC to AIFF in ~15 secs.  Only a very poorly designed system could not handle 15 secs of processing spread across 4 minutes of audio playback.  Even though this is "real-time", as you can see, there is plenty of time to drain buffers, deal with codec latencies, do error correction... and this is far from a state of the art computer... it was criticized as "slow" even when it was brand new 🙂.  Yes, codecs can have different latencies, but, for CD audio, they're all well within the hardware's tolerance.

Stan

Stan wrote:

First, I think you have a typo...  "The transport process is definitely lossless (unless  it uses network-based protocols... )".  I'm pretty sure that you mean "lossy" (the opposite of lossless) since the rest or your statement talks about the effects of losing data (lower resolution, color degradation, etc.).

Yes, good pick-up (someone's paying some attention to these threads). 🙂

Stan wrote:

Secondly, I agree with most of your facts, just not all of your conclusion.  To paraphrase (and please correct me if I'm misinterpreting your words), your point is that because AV is "real-time", you can lose data because the algorithms/transport run out of time or buffer space to do appropriate error correction.  This may be true with high resolution video such as Blu-Ray as rather intensive processing is required, and perhaps this is why you changed the subject more toward video.

Pretty much.  And yes, the issue (bit error rate) is more evident as the data transfer rates increase for AV-based signals (Blu-ray, SACD, DVD-Audio etc)... I have no doubt, however, that CDs (PCM that is) faces this problem also (especially for a low-end computer -such as the BM5).  Not so much due to the bandwidth aspect of the "transport" but for all the other factors in between (e.g. poor cabling leading to interference or inferior socket-based contact, master clock of the DAC).  Why would audiophiles, even for CD playback, pay $20,000 for a separate master clock unit?  "The master clock controls the accuracy of the sample frequency in a DAC.  While audio interfaces do have an on-board clock, often these clocks either are not as stable as most of us would like, or they don’t have sufficient distribution."

I'm running out of juice on this one but, as I have stated, I encourage you all to do your own tests (being skeptical is not a bad thing, being wrong is a different matter 🙂) 

I've run out of juice too!

I believe and will leave it at that - as in all of these "audiophile" cases, there can be no end to the debate, the believers cannot offer enough proof to convince those that don't - the ABX double blind test format cannot prove a null result, only that there is a statistically significant difference................... those that believe there is a difference tend to disagree with the double blind ABX test format -

it has always been thus!

Goodnight🙂

"Why would audiophiles, even for CD playback, pay $20,000 for a separate master clock unit?"

Because he wants to, and this proves nothing other than some people have a lot of money to spend.

I'm done, too.

Goodnight/Good Morning/Good Afternoon (depending on where you live).

Stan

Stan wrote:

Because he wants to, and this proves nothing other than some people have a lot of money to spend.

I believe you've inadvertently offended half the B&O community 🙂  How about: because they aspire to the best quality... 😕  

As a side note, I have listened to a sound system with a separate master clock unit and there is a quite noticeable improvement.

Good morning for me (down-under).

dkresz wrote:

Stan wrote:

Because he wants to, and this proves nothing other than some people have a lot of money to spend.

I believe you've inadvertently offended half the B&O community 🙂  How about: because they aspire to the best quality... 😕  

I don't believe he has - price, quality and functionality are not necessarily related as some recent products have highlighted.

dkresz wrote:

puncher wrote:

If both 1. & 2. are lossless conversions, to what do you attribute the apparent superioity of the wma lossless file?

Primarily, on a blind test - comparing the same song at the same (lossless) resolution on the BM5.  I encourage you to do the same.

Now to the speculative part: one is (mainly) Mac and the other is Windows? The codecs have diiferent encryption / decryption algorithms that are affected by the platform?  BM5, being Windows XP Embedded, may be fine-tuned for WMA and WAV playback ?  It will be interesting to re-run the same comparison test when the BM5 / BS5 software gets upgraded to officially support AAC lossless...

I have to toss my foot in here.

All that (and the rest of your posts on this subject) is emphatically wrong.

If in doubt, find an uncompressed track, say, a wave-format file. Md5 checksum it.

Convert the file to Aiff and then back to wave. Md5 checksum it.

Convert the file to ALAC (Apple Lossless) and then back to wave . Md5 Checksum it.

Convert the file to flac and then back to wave. Md5 checksum it.

Convert the file to WMA lossless and then back to wave. Md5 checksum it.

You will find that they ALL have the exact same checksum. The reason for this is, naturally, that they ALL are lossless containers, containing the exact same PCM-data.

As I said in the other thread: Any difference you hear is imaginary.

b