The shape of things to come - is there life after CD and DVD?

[Scarpia]

      Amazing. Yet I'm not amazed. Maybe S/PDIF is OK. If it works on expensive audiophile equipment maybe it just....works. What do you think? Is jitter at average rates audible to you?

      Over the years I've learned that not all specs represent audible differences. Instead many differences reside on the far side of audibility and this is generally true of the spread of jitter measurements. I remember when the jitter issue came up in the late '80s the engineers who were familiar with jitter from radio and other applications were somewhat befuddled at the misuse of this term. Jitter was not a concern at audio frequencies. They probably didn't understand how the audiophile mind works. They should have imagined something like a dragon labeled "Jitter" at the edge of the world on a medeival map, or Adam and Eve expelled from the GOE (perhaps wearing BOSE^TM headphones).

Ok, here's an issue.  I finally managed to get all of the cables to send P/SPID from my motherboard built in sound card to my AV receiver.  However, when I send CD data (or flac files) through foobar2000, the status indicator of my receiver says 48kHz PCM (not 44.1kHz PCM, as it does when it gets data from the CD player).  So the computer is doing an on-the-fly resampling from 44.1kHz to 48kHz, which is not good.   The question is, where do I find a sound card that will simply send the bits out at 44.1 kHz, rather than messing with them.  I wonder if this is the sound hardware, or some driver somewhere.  (This is using Windows XP Professional, with an Intel D845PEBT2 motherboard and a generic foobar2000 installation.)

[Daverz]

I haven't noticed any differences between playback directly from CD and playback via S/PDIF from my Squeezebox.  Both use the DAC in my Cambridge 840C CD player, which has digital inputs.

[drogulus]

Ok, here's an issue.  I finally managed to get all of the cables to send P/SPID from my motherboard built in sound card to my AV receiver.  However, when I send CD data (or flac files) through foobar2000, the status indicator of my receiver says 48kHz PCM (not 44.1kHz PCM, as it does when it gets data from the CD player).  So the computer is doing an on-the-fly resampling from 44.1kHz to 48kHz, which is not good.   The question is, where do I find a sound card that will simply send the bits out at 44.1 kHz, rather than messing with them.  I wonder if this is the sound hardware, or some driver somewhere.  (This is using Windows XP Professional, with an Intel D845PEBT2 motherboard and a generic foobar2000 installation.)

     Yes, that's right, your sound card is resampling everything to 48 kHz. You might be able to set the default to 44.1 kHz in the Control Panel > Sound (not called that in XP) or your cards software (it should be there in the Start Menu under the sound cards name).

      Does your card have 96 kHz as an option? Use that. Even audiophiles use it so it must be excellent.   Seriously, upsampling to 96 kHz, if not exactly an improvement, probably is the best way to deal with this problem if native rate isn't available. My sound card allows me to set the default at 44.1 kHz so I do. What about all my 48 and 96 kHz music? It resamples on the fly down to 44.1. Does this make me sad? Only just a little. I'm tired of fussing with it.

   

[Scarpia]

     Yes, that's right, your sound card is resampling everything to 48 kHz. You might be able to set the default to 44.1 kHz in the Control Panel > Sound (not called that in XP) or your cards software (it should be there in the Start Menu under the sound cards name).

      Does your card have 96 kHz as an option? Use that. Even audiophiles use it so it must be excellent.   Seriously, upsampling to 96 kHz, if not exactly an improvement, probably is the best way to deal with this problem if native rate isn't available. My sound card allows me to set the default at 44.1 kHz so I do. What about all my 48 and 96 kHz music? It resamples on the fly down to 44.1. Does this make me sad? Only just a little. I'm tired of fussing with it.

   

I looked for a setting for output frequency, to no avail.  What sound card do you have?

[DavidW]

I looked for a setting for output frequency, to no avail.  What sound card do you have?

After you go control panel-> sound you should hopefully see digital output listed under devices, highlight that then click on properties, then click on supported formats tab and then you should see the options available.

[Scarpia]

After you go control panel-> sound you should hopefully see digital output listed under devices, highlight that then click on properties, then click on supported formats tab and then you should see the options available.

On my Windows 7 system, for which I have a little Turtle Beach USB sound card, I found that dialog box and set it for 44.1 kHz output only.  It still outputs 48 kHz (according to my receiver).  The other computer with the built in sound card runs XP, I'll have to see if I have better luck there.

[Valentino]

My sound card allows me to set the default at 44.1 kHz so I do. What about all my 48 and 96 kHz music? It resamples on the fly down to 44.1. Does this make me sad? Only just a little. I'm tired of fussing with it. 

I too downsaple everything but 44.1 to 44.1. That's the frequency of my 4-way crossover. 

[Scarpia]

I too downsaple everything but 44.1 to 44.1. That's the frequency of my 4-way crossover. 

Down-sampling is less of a problem than up-sampling.  In down-sampling there is more information than necessary to generate the output and some of the information is that is available is not used.  In up-sampling, the output carries more information than the source, and the whatever is doing the conversion has to essentially make stuff up to fill in the gaps.   The question is, how is that done, and who wrote the algorithm?  The "genius" hired to write the driver for a $30 sound card?  The sound card on my Intel Motherboard doesn't let me change the output frequency but gives me a slider control which controls the "quality" of the frequency conversion.   What is that supposed to mean?

All I want is a gadget that will send the bits read off a CD or flac file directly to a S/PDIF cable without messing with them.  Does anyone know of such a gadget?

[71 dB]

Down-sampling is less of a problem than up-sampling.

It's actually the opposite.

In down-sampling there is more information than necessary to generate the output and some of the information is that is available is not used.

There is not such thing as "necessary" information. Conversion to higher sampling frequency doesn't mean more information, only more bits. There is simply more redundancy.

Down-sampling is not trivial. The signal must be band-limited (filtered) properly.

In up-sampling, the output carries more information than the source, and the whatever is doing the conversion has to essentially make stuff up to fill in the gaps.

Wrong. There is not more information, only more bits and redundancy.

The question is, how is that done, and who wrote the algorithm?

Theoretically it's done by reconstructing the signal using sinc-function and resampling it. In practice we must limit the amount and length of sinc-functions to make the calculations finite and real-time if needed. This makes the result less than 100 % perfect but extremely good result is easy nevertheless.

The "genius" hired to write the driver for a $30 sound card?  The sound card on my Intel Motherboard doesn't let me change the output frequency but gives me a slider control which controls the "quality" of the frequency conversion.   What is that supposed to mean?

As I said above, the calculations must be limited. The longer sinc-functions (or other interpolating function) are used, the better result but the need for CPU power increases.

[Scarpia]

It's actually the opposite.
There is not such thing as "necessary" information. Conversion to higher sampling frequency doesn't mean more information, only more bits. There is simply more redundancy.

Down-sampling is not trivial. The signal must be band-limited (filtered) properly.
Wrong. There is not more information, only more bits and redundancy.
Theoretically it's done by reconstructing the signal using sinc-function and resampling it. In practice we must limit the amount and length of sinc-functions to make the calculations finite and real-time if needed. This makes the result less than 100 % perfect but extremely good result is easy nevertheless.
As I said above, the calculations must be limited. The longer sinc-functions (or other interpolating function) are used, the better result but the need for CPU power increases.

You seem to have totally misunderstood my comments.  I was not claiming that the signal resampled at 48 kHz has more information, but that the channel has a larger information capacity.  If the original waveform is sampled at 48 kHz and you want to produce a 44.1 kHz waveform, you know the frequency information up to 24 kHz and need to generate a new signal with response to 22 kHz.   Obviously the frequency spectrum above 22 kHz must be suppressed (or frequency aliasing will occur.)  If you have 44.1 kHz data and need to generate 48 kHz data, you know the spectrum up to 22 kHz and you need to represent it in a medium which represents the signal up to 24 kHz data.  The new carrier has a frequency band for which no information is available from the source.   Anything that appears in this band is noise.  Obviously it is possible to devise a resample scheme that keeps this band clean, but I am questioning whether I can trust the driver that comes with a 30 dollar sound card to contain an elegant resampling scheme.

[71 dB]

You seem to have totally misunderstood my comments.  I was not claiming that the signal resampled at 48 kHz has more information, but that the channel has a larger information capacity.  If the original waveform is sampled at 48 kHz and you want to produce a 44.1 kHz waveform, you know the frequency information up to 24 kHz and need to generate a new signal with response to 22 kHz.   Obviously the frequency spectrum above 22 kHz must be suppressed (or frequency aliasing will occur.)  If you have 44.1 kHz data and need to generate 48 kHz data, you know the spectrum up to 22 kHz and you need to represent it in a medium which represents the signal up to 24 kHz data.  The new carrier has a frequency band for which no information is available from the source.   Anything that appears in this band is noise.  Obviously it is possible to devise a resample scheme that keeps this band clean, but I am questioning whether I can trust the driver that comes with a 30 dollar sound card to contain an elegant resampling scheme.

Whether or not I misundertood your comment, this time you are right about what you say.

Quantization noise on bandwidth 22-24 kHz is nothing to worry about. Even if human hearing let us hear sounds this high frequency, the noise level is very low. In fact, it is clevel to use noise-shaping to move all quantization noise to the highest frequencies.

[drogulus]

     

On my Windows 7 system, for which I have a little Turtle Beach USB sound card, I found that dialog box and set it for 44.1 kHz output only.  It still outputs 48 kHz (according to my receiver).  The other computer with the built in sound card runs XP, I'll have to see if I have better luck there.

     That Turtle Beach is 16/48 kHz only. That's a limitation of the hardware. Windows 7 is fairly good about sample rates if you have a card that supports setting them where you want.

Obviously it is possible to devise a resample scheme that keeps this band clean, but I am questioning whether I can trust the driver that comes with a 30 dollar sound card to contain an elegant resampling scheme.

     I don't know the answer to this. Probably a card that resamples everything to 48 kHz does it reasonably well. The main benefit of a USB card like this is that the D/A conversion is done outside the computer case. If you want a USB card with the ability to set bit/sample rates where you want the M-Audio Transit will do it for about $79:

     

     My receiver doesn't report sample rates, so I use a file that tests it for me. It's a 44.1 kHz DTS surround file disguised as a WAV (?). If the output hasn't been resampled you hear this Swedish guy testing your surround speaker setup (works fine for 2 speakers), if not you get silence or a horrible noise. And heeere it is!!

     

[Scarpia]

     
     That Turtle Beach is 16/48 kHz only. That's a limitation of the hardware. Windows 7 is fairly good about sample rates if you have a card that supports setting them where you want.


     I don't know the answer to this. Probably a card that resamples everything to 48 kHz does it reasonably well. The main benefit of a USB card like this is that the D/A conversion is done outside the computer case. If you want a USB card with the ability to set bit/sample rates where you want the M-Audio Transit will do it for about $79:

     

     My receiver doesn't report sample rates, so I use a file that tests it for me. It's a 44.1 kHz DTS surround file disguised as a WAV (?). If the output hasn't been resampled you hear this Swedish guy testing your surround speaker setup (works fine for 2 speakers), if not you get silence or a horrible noise. And heeere it is!!

   

Excellent, I will try the M-audio gadget.

[drogulus]

Excellent, I will try the M-audio gadget.

     I wouldn't expect a big SQ difference, and possibly none at all. You might consider investing in improvements elsewhere, or just saving your money. Once you've moved the D/A process out of the computer case the big problem has been remedied. There may not be anything else that needs to be done.

[Scarpia]

     I wouldn't expect a big SQ difference, and possibly none at all. You might consider investing in improvements elsewhere, or just saving your money. Once you've moved the D/A process out of the computer case the big problem has been remedied. There may not be anything else that needs to be done.

The fact that the re-sampling can be done properly does not prove that it was done properly in this gadget.  The only reason I started investigating the sampling rate issue is that the sound from the computer did not sound identical to the same track streamed from a cd player.

[XB-70 Valkyrie]

After all these years, my preferred source is still the vinyl LP. I have recorded hundreds of my LPs onto CDs using my CD-Recorder (not a computer drive) hooked up to my preamp. I have FLAC-ed most of my CDs (including those made from vinyl), and I have some of them as MP3s on my iPOD (which was given to me--I didn't buy it). It's useful to have these different formats, but I enjoy playing the vinyl itself more than any of them.

I own hundreds of commercially released CDs, and I still continue to buy these. However, I've always felt that the CD is a very cold, uninteresting, utiliatrian object--not a tactile, sensual thing like a vinyl LP with its beautiful cover designs and interesting liner notes and booklets (to say nothing of its sound, which I much prefer).

I have never bought a single thing from iTunes, and I likely never will. I hate the whole Apple music and content distribution enterprise. And I loathe and fear the day when the media giants will have their fondest wet-dream fulfilled--i.e., when no one will actually own anything, but will be in perpetual bondage to these companies, constantly having to pay a monthly service charge for content. Certainly these whores will have no interest in selling anything obscure that does not net them a substantial profit. Maybe I'm an alarmist, but I fear numerous beautiful, unique performances being lost forever, with only the original disks in the hands of private collectors remaining.

[DavidW]

I have never bought a single thing from iTunes, and I likely never will. I hate the whole Apple music and content distribution enterprise. And I loathe and fear the day when the media giants will have their fondest wet-dream fulfilled--i.e., when no one will actually own anything, but will be in perpetual bondage to these companies, constantly having to pay a monthly service charge for content. Certainly these whores will have no interest in selling anything obscure that does not net them a substantial profit. Maybe I'm an alarmist, but I fear numerous beautiful, unique performances being lost forever, with only the original disks in the hands of private collectors remaining.

Okay let me set this straight--
* Jobs is opposed to renting music, only wants people to buy music, and you can NOT rent music on itunes and you never will be able to.
* The rental programs that have been started are not compatible with ipod products, and most likely will never be.
* Apple is not behind any kind of rental program to control content, and in fact their music is drm free, and Jobs was opposed to drm from the start (was only forced to have it for awhile because the record companies required it).
* As long as itunes and amazon mp3 exist (and between them nearly control the entire digital downloads industry) renting will not take over.
* The record companies were the ones that lobbied to make renting (with the computer loophole) illegal, that's why we don't have a netflix for cds.  They want you to buy albums more than even you do, that's how they became rich in the 90s!!  And in fact the discounts on album downloads on amazon, and album only downloads on both are a push to get people into buying whole albums again instead of just songs.
* The rental programs are just a quick fix increase profit revenue during hard times just like the more lucrative business of selling ringtones.

It's okay to be alarmist, but honestly to say it straight, I think that you're talking out of your ass.  No offense though.

[DavidW]

Franco, I've messed around more and I found that

* EAC is not efficient due to lack of too much things that should be automated
* dBpoweramp is a big step up, did everything including the coverart, but the tagging was still spotty and it wasn't quite automated enough (i.e. wouldn't start automatically) and there is a slowdown there
* Itunes is nearly perfect!  Everything can be automated, tagging is excellent, manages my library, and the only annoyance is that I have to add coverart myself. You were right.

Itunes saves so much time that ripping is not a big hassle anymore.   And you had the right way to go with the ripping.

[drogulus]

     

The fact that the re-sampling can be done properly does not prove that it was done properly in this gadget.  The only reason I started investigating the sampling rate issue is that the sound from the computer did not sound identical to the same track streamed from a cd player.

      So CDs going through your optical drive > USB card > receiver sound different than CDs through your player > receiver? I have some questions.

      1) Output voltage from the Turtle Beach?

      2) Output voltage of the CD player?

      3) Have all the effects been turned off? No EQ? And most importantly have all volume controls been disabled or maxed?

      Your CD player is sending a clean signal to your receiver. If you level match your PC output and make sure it's equally clean will it still sound different? Maybe it will, but you won't know until you try it. My guess is that a tiny USB-powered sound card delivers a lower output voltage than a CD player. One of the most common effects in informal comparisons of sound quality is hearing lower volume as poorer sound.

[Scarpia]

     
      So CDs going through your optical drive > USB card > receiver sound different than CDs through your player > receiver? I have some questions.

      1) Output voltage from the Turtle Beach?

      2) Output voltage of the CD player?

      3) Have all the effects been turned off? No EQ? And most importantly have all volume controls been disabled or maxed?

      Your CD player is sending a clean signal to your receiver. If you level match your PC output and make sure it's equally clean will it still sound different? Maybe it will, but you won't know until you try it. My guess is that a tiny USB-powered sound card delivers a lower output voltage than a CD player. One of the most common effects in informal comparisons of sound quality is hearing lower volume as poorer sound.

I was comparing 1) DVD player playing the CD and sending PCM audio data via digital interconnect (S/PDIF on optical interconnect) with 2) PC streaming digital data from a FLAC file burned from the same CD (again S/PIDF on optical interconnect).   All effects, equalization, etc, disabled and volume controls maxed.    My impression was there was a slight difference, then I noticed that the receiver reported 48 kHz rather than 44 kHz when the PC was the source.