Windows audio sampling rate setting and mixed content

C

connta

Audioholic Intern
Lately i stumbled on this "issue", video content is standardized at sampling rate of 48kHz and audio is standardized to 44.1 kHz, Windows has to be set to either so if you do not switch constantly theres bound to be some conversion in the output chain, either cutting or dithering (i guess). My question is, for someone who uses PC to play both video and music, whats the lesser evil of the two if you want to "set it and forget it"? Does 44.1 -> 48 just leaves a blank on the difference (less intrusive) and 48 -> 44.1 cuts some of the signal to fit the format (worse, i guess) or does it work some other way? Are there any solutions to this, does it warrant one?
 
TLS Guy

TLS Guy

Seriously, I have no life.
Lately i stumbled on this "issue", video content is standardized at sampling rate of 48kHz and audio is standardized to 44.1 kHz, Windows has to be set to either so if you do not switch constantly theres bound to be some conversion in the output chain, either cutting or dithering (i guess). My question is, for someone who uses PC to play both video and music, whats the lesser evil of the two if you want to "set it and forget it"? Does 44.1 -> 48 just leaves a blank on the difference (less intrusive) and 48 -> 44.1 cuts some of the signal to fit the format (worse, i guess) or does it work some other way? Are there any solutions to this, does it warrant one?
The answer is ha 44.1 gets you to 20 KHz and is perfectly fine. That is the CD standard. The BPO Digital Concert Hall is 44.1 FLAC and it sounds fantastic. The dynamic range is also adequate for almost all music. I stream AV with my HTPC all the time, and it is actually the best streaming source. 44.1 is just fine, the rest is overkill. Don't worry about. If it was a problem then CDs would not achieve the excellent quality they do. This is not a problem in need of a solution. There is more audiophoolery about this than actual fact.
 
C

connta

Audioholic Intern
Im not sure you understood the angle of the problem im going for. I get everything you said and have no remarks but from a technical standpoint when you play 48kHz encoded video and your Windows is set to 44.1kHz output the signal that Windows got (48) will be modified to fit the new output format (44.1) and in that process of conversion usually there is signal degradation and i would like to avoid unnecessary degradation where i can. Or pick a less intrusive form of conversion at least, to fit 48 into 44.1 something has to be cut which will always be "bad" but the other way around might not be as bad since there is no lack of bandwidth but surplus of it which can just be populated with zeros to fit the new format and no information will be lost.
 
TLS Guy

TLS Guy

Seriously, I have no life.
Im not sure you understood the angle of the problem im going for. I get everything you said and have no remarks but from a technical standpoint when you play 48kHz encoded video and your Windows is set to 44.1kHz output the signal that Windows got (48) will be modified to fit the new output format (44.1) and in that process of conversion usually there is signal degradation and i would like to avoid unnecessary degradation where i can. Or pick a less intrusive form of conversion at least, to fit 48 into 44.1 something has to be cut which will always be "bad" but the other way around might not be as bad since there is no lack of bandwidth but surplus of it which can just be populated with zeros to fit the new format and no information will be lost.
There is no significant degradation. You have to move out of the analog world. The answer to your question lies in understanding dither, and its purpose.

As I explained, the frequency response is adequate. You are right there must be a downside, but in this case it is insignificant.

As you know digital signals are binary, that means two choices, 1 or 0. Now as you reduce the modulation, which in this case we can simply call loudness. Its corollary is dynamic range. Now as the audio signal is reduced to silence, you get to a point where the choice is a 1 or a zero, and getting it wrong results in 100% error. That translates to very high distortion on low level signals. The way this is handled is to add a low level of white noise, to the signal. This is known as dithering. So enough low level white noise is added to digital signals in inverse proportion to the sampling frequency. So what this means is that the signal noise ratio (SNR) slightly degrades as you reduce the sampling frequency.

Now a 44.1 sampling frequency gives you a 90 db SNR at 44.1 KHz. To put this in perspective an LP has an SNR of 60 db, roughly. Almost all environments have an SNR below 90 db and usually by a big margin. So in going from 48 to 44.1 KHz sampling frequency requires very slightly increasing the dither signal, causing an imperceptible reduction in the SNR. Now if you record a very high dynamic range work, like say Mahler's symphony 2 or 3, it just might, and I emphasize might, when the Flugel horns play from off stage or ideally a high balcony , notice an increase in background. However this decrease in SNR will almost certainly be dwarfed by the air handlers in the venue.
'No music in the pop culture requires anywhere near the dynamic range and SNR of a CD quality recording. So for that genre 44.1 has dynamic range to spare and plenty of it.

I have made digital recordings and mastering, in fact still do master the odd CD. Judging the dither used to be a skilled business, but now the mastering software does a better job than I can.

I suspect you have been looking at a lot of ignorant verbiage on the Net's blogosphere, of which there is no shortage. Here we aim to give you the correct information and usually we have our ways of ridding ourselves of the willfully ignorant.
 
Bucknekked

Bucknekked

Audioholic Samurai
There is no significant degradation. You have to move out of the analog world. The answer to your question lies in understanding dither, and its purpose.

As I explained, the frequency response is adequate. You are right there must be a downside, but in this case it is insignificant.

As you know digital signals are binary, that means two choices, 1 or 0. Now as you reduce the modulation, which in this case we can simply call loudness. Its corollary is dynamic range. Now as the audio signal is reduced to silence, you get to a point where the choice is a 1 or a zero, and getting it wrong results in 100% error. That translates to very high distortion on low level signals. The way this is handled is to add a low level of white noise, to the signal. This is known as dithering. So enough low level white noise is added to digital signals in inverse proportion to the sampling frequency. So what this means is that the signal noise ratio (SNR) slightly degrades as you reduce the sampling frequency.

Now a 44.1 sampling frequency gives you a 90 db SNR at 44.1 KHz. To put this in perspective an LP has an SNR of 60 db, roughly. Almost all environments have an SNR below 90 db and usually by a big margin. So in going from 48 to 44.1 KHz sampling frequency requires very slightly increasing the dither signal, causing an imperceptible reduction in the SNR. Now if you record a very high dynamic range work, like say Mahler's symphony 2 or 3, it just might, and I emphasize might, when the Flugel horns play from off stage or ideally a high balcony , notice an increase in background. However this decrease in SNR will almost certainly be dwarfed by the air handlers in the venue.
'No music in the pop culture requires anywhere near the dynamic range and SNR of a CD quality recording. So for that genre 44.1 has dynamic range to spare and plenty of it.

I have made digital recordings and mastering, in fact still do master the odd CD. Judging the dither used to be a skilled business, but now the mastering software does a better job than I can.

I suspect you have been looking at a lot of ignorant verbiage on the Net's blogosphere, of which there is no shortage. Here we aim to give you the correct information and usually we have our ways of ridding ourselves of the willfully ignorant.
I always enjoy reading your explanations for things. I can and do learn something almost every time.
For me, the digital sampling rates and the audio quality of the CD are settled issues. No issue.
But not settled for everyone and I thought the OP's question was a fair one. As usual, you killed it on the answer
 
C

connta

Audioholic Intern
There is no significant degradation. You have to move out of the analog world. The answer to your question lies in understanding dither, and its purpose.

As I explained, the frequency response is adequate. You are right there must be a downside, but in this case it is insignificant.

As you know digital signals are binary, that means two choices, 1 or 0. Now as you reduce the modulation, which in this case we can simply call loudness. Its corollary is dynamic range. Now as the audio signal is reduced to silence, you get to a point where the choice is a 1 or a zero, and getting it wrong results in 100% error. That translates to very high distortion on low level signals. The way this is handled is to add a low level of white noise, to the signal. This is known as dithering. So enough low level white noise is added to digital signals in inverse proportion to the sampling frequency. So what this means is that the signal noise ratio (SNR) slightly degrades as you reduce the sampling frequency.

Now a 44.1 sampling frequency gives you a 90 db SNR at 44.1 KHz. To put this in perspective an LP has an SNR of 60 db, roughly. Almost all environments have an SNR below 90 db and usually by a big margin. So in going from 48 to 44.1 KHz sampling frequency requires very slightly increasing the dither signal, causing an imperceptible reduction in the SNR. Now if you record a very high dynamic range work, like say Mahler's symphony 2 or 3, it just might, and I emphasize might, when the Flugel horns play from off stage or ideally a high balcony , notice an increase in background. However this decrease in SNR will almost certainly be dwarfed by the air handlers in the venue.
'No music in the pop culture requires anywhere near the dynamic range and SNR of a CD quality recording. So for that genre 44.1 has dynamic range to spare and plenty of it.

I have made digital recordings and mastering, in fact still do master the odd CD. Judging the dither used to be a skilled business, but now the mastering software does a better job than I can.

I suspect you have been looking at a lot of ignorant verbiage on the Net's blogosphere, of which there is no shortage. Here we aim to give you the correct information and usually we have our ways of ridding ourselves of the willfully ignorant.
I appreciate the explanation but this is all the information i already know.

Is 44.1 absolutely enough for music reproduction? Yes.
Is 90dB of dynamic range enough for 98% of scenarios. Yes
Will there be degradation if you play 48kHz source on a 44.1kHz output and vice versa. Technically, yes.
Does it matter, is it audible? Mostly no.

Im hearing it loud and clear. :)

But none of this answered what i asked. Let me simplify as much as i can.

44.1 source on a Windows 48 output = source upsampling
48 source on a Windows 44.1 output = source downsampling

Which one is the "lesser evil", Windows upsampling or downsampling algorithm? Or are they identical?

I get it that from audible standpoint the answer is mostly "identical" since either is not going to cause audible degradation but from a technical standpoint, which one would be less intrusive on the source signal?
 
TLS Guy

TLS Guy

Seriously, I have no life.
I appreciate the explanation but this is all the information i already know.

Is 44.1 absolutely enough for music reproduction? Yes.
Is 90dB of dynamic range enough for 98% of scenarios. Yes
Will there be degradation if you play 48kHz source on a 44.1kHz output and vice versa. Technically, yes.
Does it matter, is it audible? Mostly no.

Im hearing it loud and clear. :)

But none of this answered what i asked. Let me simplify as much as i can.

44.1 source on a Windows 48 output = source upsampling
48 source on a Windows 44.1 output = source downsampling

Which one is the "lesser evil", Windows upsampling or downsampling algorithm? Or are they identical?

I get it that from audible standpoint the answer is mostly "identical" since either is not going to cause audible degradation but from a technical standpoint, which one would be less intrusive on the source signal?
As I told you the only thing that will change is the SNR will be downgraded by an amount you pretty much for certain won't notice. One point you raise is up sampling.
There is NEVER an advantage to up sampling. A Wav. file will always carry the characteristics of the lowest sample rate it ever existed in. If you think about it, this has to be the logical conclusion. So up sampling does not harm, but brings ZERO benefit.
I do think you are stating to get the hang of it now.

All this assumes that you are working with non lossy codecs. All this is not to be confused by the ubiquitous use of lossy codecs in order to save bandwidth, especially for streaming.

These are codecs we are talking about. There are I'm sure others I have never heard of
  • MP3
  • AAC.
  • AIFF.
  • WMA.
  • Ogg Vorbis.
These codecs use psychological research to determine which bits can be thrown out, without people noticing. Although this sounds an awful idea, in practice the good codecs actually work very well. MP3 is not that good, and I can usually detect it especially at the lower bit rates. AAC is a particularly good codec.

The there is the WAV, file which is the protocol in which the digital form of music is stored, edited, converted etc. Wav. files can be lossy or not, and encoded in pretty much any codec.

Then there are so called lossless codecs like like FLAC where there can be sent in code, but the information is there to fully restore the file to non lossy. ALAC for instance is what Apple uses. So they use a code/encode system. FLAC is totally seamless. It reduces the transmission bit rate by just under 50%. It is totally seamless and undetectable to the listener.
The BPO have started using FLAC in their streams which are excellent.

Since this is all very complicated it is not surprising there is misunderstanding with false information on the NET abounding. Hopefully not here, but no guarantee. But there is usually a member to set us straight in our errors. What we pretty much do not like is the willfully ignorant, especially when peppered with any degree of arrogance or abuse. You on the other hand seem ready to get the hang of this. We recognize this is not easy subject matter and not always intuitive. We have to constantly remind ourselves to leave our old analog thinking at the door. For an old duffer like me this can be easier said then done at times.
 
Last edited:
T

Trebdp83

Audioholic Ninja
Lately i stumbled on this "issue", video content is standardized at sampling rate of 48kHz and audio is standardized to 44.1 kHz, Windows has to be set to either so if you do not switch constantly theres bound to be some conversion in the output chain, either cutting or dithering (i guess). My question is, for someone who uses PC to play both video and music, whats the lesser evil of the two if you want to "set it and forget it"? Does 44.1 -> 48 just leaves a blank on the difference (less intrusive) and 48 -> 44.1 cuts some of the signal to fit the format (worse, i guess) or does it work some other way? Are there any solutions to this, does it warrant one?
 
Bucknekked

Bucknekked

Audioholic Samurai
As I told you the only thing that will change is the SNR will be downgraded by an amount you pretty much for certain won't notice. One point you raise is up sampling.
There is NEVER an advantage to up sampling. A Wav. file will always carry the characteristics of the lowest sample rate it ever existed in. If you think about it, this has to be the logical conclusion. So up sampling does not harm, but brings ZERO benefit.
I do think you are stating to get the hang of it now.

All this assumes that you are working with non lossy codecs. All this is not to be confused by the ubiquitous use of lossy codecs in order to save bandwidth, especially for streaming.

These are codecs we are talking about. There are I'm sure others I have never heard of
  • MP3
  • AAC.
  • ALAC.
  • AIFF.
  • WMA.
  • Ogg Vorbis.
These codecs use psychological research to determine which bits can be thrown out, without people noticing. Although this sounds an awful idea, in practice the good codecs actually work very well. ALAC for instance is what Apple uses. MP3 is not that good, and I can usually detect it especially at the lower bit rates. AAC is a particularly good codec.

The there is the WAV, file which is the protocol in which the digital form of music is stored, edited, converted etc. Wav. files can be lossy or not, and encoded in pretty much any codec.

Then there are so called lossless codecs like like FLAC where there can be sent in code, but the information is there to fully restore the file to non lossy. So they use a code/encode system. FLAC is totally seamless. It reduces the transmission bit rate by just under 50%. It is totally seamless and undetectable to the listener.
The BPO have started using FLAC in their streams which are excellent.

Since this is all very complicated it is not surprising there is misunderstanding with false information on the NET abounding. Hopefully not here, but no guarantee. But there is usually a member to set us straight in our errors. What we pretty much do not like is the willfully ignorant, especially when peppered with any degree of arrogance or abuse. You on the other hand seem ready to get the hang of this. We recognize this is not easy subject matter and not always intuitive. We have to constantly remind ourselves to leave our old analog thinking at the door. For an old duffer like me this can be easier said then done at times.
Doc
I hate it when I think I have a difference of opinion with one of your excellent posts. Hopefully it's a typo or just one of those things that got jumbled in your reply. Maybe I misunderstood. Here is my difference of opinion, respectfully given.

ALAC, listed in your list of problematic codecs, is not inferior to FLAC. They are essentially identical wrappers or containers. Both contain within them the identical file information, if that information was created from the same source file. The only difference, and it is a difference, is that ALAC is supported on Apple devices and is Apple's choice for lossless. FLAC is the open standard. FLAC is not supported directly on iTunes and it's Apple's boneheaded choice to make it that way. FLAC will usually play anywhere else.

One can take a file in either format, ALAC or FLAC, and create a new file in the other format and lose nothing.
The difference is in the wrapper/container formats not the lossless file within it.

flac-vs-alac-quality.jpg


Anyway, that's the size of my opinion. Because of the Apple file format favoritism on their own stuff, I rip all my music to both ALAC and FLAC and keep them organized separately. FLAC goes to a PLEX server and my ALAC stuff hits my Apple machines. Then its all housed on a network based Synology disk farm.
 
T

Trebdp83

Audioholic Ninja
Also, ALAC and FLAC are compressed, though lossless. FLAC can refer to the file format or container itself. WAV and AIFF are lossless and uncompressed and take up double or more the space of ALAC and FLAC. AAC and MP3 are both compressed and lossy. I rip in AIFF and stream FLAC from services using the Onkyo/DTS-Play-Fi app in my Mac to my receiver over my network. One has to choose the output for HDMI and the choices are 16/48, 20/48 and 24/48. I chose 24/48 and everything is output that way.
 
TLS Guy

TLS Guy

Seriously, I have no life.
Doc
I hate it when I think I have a difference of opinion with one of your excellent posts. Hopefully it's a typo or just one of those things that got jumbled in your reply. Maybe I misunderstood. Here is my difference of opinion, respectfully given.

ALAC, listed in your list of problematic codecs, is not inferior to FLAC. They are essentially identical wrappers or containers. Both contain within them the identical file information, if that information was created from the same source file. The only difference, and it is a difference, is that ALAC is supported on Apple devices and is Apple's choice for lossless. FLAC is the open standard. FLAC is not supported directly on iTunes and it's Apple's boneheaded choice to make it that way. FLAC will usually play anywhere else.

One can take a file in either format, ALAC or FLAC, and create a new file in the other format and lose nothing.
The difference is in the wrapper/container formats not the lossless file within it.

View attachment 58373

Anyway, that's the size of my opinion. Because of the Apple file format favoritism on their own stuff, I rip all my music to both ALAC and FLAC and keep them organized separately. FLAC goes to a PLEX server and my ALAC stuff hits my Apple machines. Then its all housed on a network based Synology disk farm.
Good catch, it should have been in the next paragraph. I cut and pasted it from a list and forgot to move it. Too early in the morning!
 
Bucknekked

Bucknekked

Audioholic Samurai
Also, ALAC and FLAC are compressed, though lossless. FLAC can refer to the file format or container itself. WAV and AIFF are lossless and uncompressed and take up double or more the space of ALAC and FLAC. AAC and MP3 are both compressed and lossy. I rip in AIFF and stream FLAC from services using the Onkyo/DTS-Play-Fi app in my Mac to my receiver over my network. One has to choose the output for HDMI and the choices are 16/48, 20/48 and 24/48. I chose 24/48 and everything is output that way.
@Trebdp83
Great minds think alike. When referring to ALAC, what I really meant to be exact is that I rip in to AIFF for my Apple devices and lossless FLAC for my other stuff. I know whether or not its compressed or uncompressed is audibly moot, but, since disk space is beyond cheap I chose AIFF, uncompressed and lossless.

I have a similar choice to make when using my AVR to drive my main system. Whatever it chooses is simply perfect for my ears.

In fact, for my humble and somewhat road worn ears, I have experimented with all the ripped file formats just to make sure I had it straight. I ripped various CDs in to everything from MP3 to AIFF to FLAC and many of those others that Doc referred to. On most rock music and pop music, since I ripped from an original CD, I was unable to tell any real difference between the file formats. But being slightly OCD, I knew that if there were any difference at all that others could complain about or point out, I ripped in to the highest rate, lossless formats and did my entire library over from scratch. Now I just add new ones as they come along and its a piece of cake.

I have a batch encoder somewhere so if a new/improved format shows up someday I should be able to migrate pretty easy if that ever becomes a need. Quite frankly, now that its all done, it is something I rarely ever think about anymore. It simply is an awesome way to listen and transport music. Effortless.

The only part that isn't "effortless" is the file naming conventions to keep media players and meta data straight.
PLEX is a pain in the ass#@ when it comes to naming. iTunes gets it right but uses different convention than PLEX and others. Sigh.
 
lovinthehd

lovinthehd

Audioholic Jedi
Doc
I hate it when I think I have a difference of opinion with one of your excellent posts. Hopefully it's a typo or just one of those things that got jumbled in your reply. Maybe I misunderstood. Here is my difference of opinion, respectfully given.

ALAC, listed in your list of problematic codecs, is not inferior to FLAC. They are essentially identical wrappers or containers. Both contain within them the identical file information, if that information was created from the same source file. The only difference, and it is a difference, is that ALAC is supported on Apple devices and is Apple's choice for lossless. FLAC is the open standard. FLAC is not supported directly on iTunes and it's Apple's boneheaded choice to make it that way. FLAC will usually play anywhere else.

One can take a file in either format, ALAC or FLAC, and create a new file in the other format and lose nothing.
The difference is in the wrapper/container formats not the lossless file within it.

View attachment 58373

Anyway, that's the size of my opinion. Because of the Apple file format favoritism on their own stuff, I rip all my music to both ALAC and FLAC and keep them organized separately. FLAC goes to a PLEX server and my ALAC stuff hits my Apple machines. Then its all housed on a network based Synology disk farm.
Isn't AIFF also lossless? I believe AIFF-C is compressed, but regular AIFF is just more apple lossless not compressed?
 
Bucknekked

Bucknekked

Audioholic Samurai
Isn't AIFF also lossless? I believe AIFF-C is compressed, but regular AIFF is just more apple lossless not compressed?
Yes. AIFF is lossless. I hope I didn’t say otherwise. I use the uncompressed setting.
yes. I wish Apple hadn’t made this pigheaded decision but they did.
 
lovinthehd

lovinthehd

Audioholic Jedi
Yes. AIFF is lossless. I hope I didn’t say otherwise. I use the uncompressed setting.
yes. I wish Apple hadn’t made this pigheaded decision but they did.
No, but it was in @TLS Guy 's list and not commented on. Then again I didn't see ALAC in the list either. Maybe edited by the time I did.....

I just wish Apple and the rest of the world used the same :)
 
lovinthehd

lovinthehd

Audioholic Jedi
Yes. AIFF is lossless. I hope I didn’t say otherwise. I use the uncompressed setting.
yes. I wish Apple hadn’t made this pigheaded decision but they did.
Well it seems AIFF had game origins from a little research but didn't dig deeper.....
 
T

Trebdp83

Audioholic Ninja
AIFF is always uncompressed. Here is an interesting read.
Here are the latest import options on MacOS Ventura 13.
 
newsletter

  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top