IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
playing 44.1khz content through 48khz sync
wwphil
post Aug 25 2012, 00:46
Post #1





Group: Members
Posts: 2
Joined: 25-August 12
Member No.: 102647



Simply out if curiosity,

What exactly happens when you play 44.1khz content over a 48khz sync (ex: my external dac is currently connected with USB, can only choose 44.1 or 48khz for sync.)

My computer runs Windows 7, which might be relevant to my question: the other day, I listened to Dark Side of the Moon SACD (flac version) and my dac was synced at 44.1khz, the content is 96khz, so I switched it to 48khz thinking "at least I'll get half the resolution".

I have not reverted it yet, and was wondering what exactly happens right now when I play regular 44.1 content over that. Is the resampling lossless ? Are the two formats mathematically compatible ? or am I losing some samples here and there ? Is it win7's sound engine doing the job ? Is it converting it to analog and then sampled at 48khz before playback ?

I know that in WinXP, the resampling is done by the hardware, which is why each company makes very different sound drivers and XP on its own doesn't do much, the sound engine in Win7 has evolved quite a lot and does most of the work, which is why you can now play games on weird asio soundcards, or over hdmi cables, without issues.

Anybody has a clue ?


This post has been edited by wwphil: Aug 25 2012, 00:50
Go to the top of the page
+Quote Post
slks
post Aug 25 2012, 08:52
Post #2





Group: Members
Posts: 426
Joined: 31-March 06
From: Houston, Texas
Member No.: 29046



Resampling is never mathematically lossless, although in most cases it is inaudible, unless you're going to very low sample rates (like 32 kHz and under).

Unfortunately I don't know much about how the Windows sound system works internally. But I do know that you can set the default sample rate in the Windows 7 mixer by right-clicking your playback device, clicking Properties, and going to the Advanced tab.

Setting everything in the signal chain (sound card, playback software, and Windows mixer) to the same sample rate should logically avoid resampling. However, in the past there have been cases of poorly-designed drivers or cards that resample any sound that goes through them, even if it's at the same sample rate! So it would be useful for you to name what hardware you're using; someone here may be familiar with it.

However, if you're trying to play 96 kHz audio through a card that only supports 44.1 or 48, resampling is unavoidable. The consensus seems to be that it's best to do resampling in your audio player, since then you have control over which sampling algorithm gets used.


--------------------
http://www.last.fm/user/sls/
Go to the top of the page
+Quote Post
LithosZA
post Aug 25 2012, 10:09
Post #3





Group: Members
Posts: 202
Joined: 26-February 11
Member No.: 88525



If the player uses WASAPI output in exclusive mode then it doesn't use the default sampling rate of windows. It just uses the sampling rate of the input source and tries to output at that sampling rate to your device. You should get an error if the sampling rate is not supported.
DirectSound output will resample anything to match the default sampling rate setting in windows.

My USB DAC also only supports 48Khz max sampling rate and 16bit. What I usaually do:
Windows Settings:
100% Volume
Disable all enhancements
Default format: 16bit, 48Khz
Speaker setup: Set to 'Full Range - Front Left and Right'

In my audio player. Foobar2000:
100% Volume and hide the volume control. (I only control the volume on my headphone amp itself)
Output: WASAPI, 16 bit, Dither ticked
DSP: Resampler (SOX) set at 48Khz, Best Quality
Go to the top of the page
+Quote Post
[JAZ]
post Aug 25 2012, 12:22
Post #4





Group: Members
Posts: 1797
Joined: 24-June 02
From: Catalunya(Spain)
Member No.: 2383



@wwphil: In Windows 7, it depends on which audio API (mme, directsound, asio, wasapi...) your player uses, and which combination of sample rates there is from input file, player, windows setting and driver setting.

A typical scenario like: input 44.1Khz, directsound at 44Kz, windows setting at 48Khz and driver at 48Khz would use windows resampling to 48Khz and sent to the driver.

Also notes:
- as said, resampling is not lossless. (it is not guaranteed to get back, bit for bit, the same data). it should be inaudible.
- resampling can only happen in analog when connection one soundcard output (in analog) to soundcard input. It is an unreasonable scenario for this operation.



@LithosZA: Mmm.. From those settings, i just wonder about the 100% volume necessity. I guess the reason is because that way you can have a lower noise floor thanks to using the volume in your DAC/amp, is it? (Obviously, it is not about getting bit-perfect playback)
Go to the top of the page
+Quote Post
LithosZA
post Aug 25 2012, 12:43
Post #5





Group: Members
Posts: 202
Joined: 26-February 11
Member No.: 88525



I am no audio expert, but my thinking was that because it is digital audio that 100% would be 0dB increase/decrease to the sound?
So there would be no reason to decrease the volume in software. Unless I am missing something?
Go to the top of the page
+Quote Post
[JAZ]
post Aug 25 2012, 13:02
Post #6





Group: Members
Posts: 1797
Joined: 24-June 02
From: Catalunya(Spain)
Member No.: 2383



It should be 0dB. That is not assured, but tends to be the case (for example, usually in linux, 0dB is more about 70% or so).

Other than that, yes, you're correct than reducing the volume in the digital domain (software), reduces the volume but also the SNR.
When using an external amplifier/volume control, it is better to use that control, so that the amplified signal still has all the SNR, which should generally be better than the amplifier's one.

Anyway, I woudn't worry in excess about this. I used to have an integrated amp at 1/3rd of volume driving a pair of 3way 8'' speakers, and the volume control in windows XP at 4% or 8%. Although I usually increased the windows volume when using the headphone out.
Go to the top of the page
+Quote Post
Nichtswisser
post Aug 25 2012, 15:43
Post #7





Group: Members
Posts: 7
Joined: 21-August 12
Member No.: 102516



QUOTE (LithosZA @ Aug 25 2012, 11:09) *
Output: WASAPI, 16 bit, Dither ticked



Why dither at all? Dithering is only required when you reduce the bit depth of sounds files, 24bit to 16bit for example. It's supposed to be inaudible most of the time yet I would not use in cases it isn't needed.

This post has been edited by Nichtswisser: Aug 25 2012, 15:44
Go to the top of the page
+Quote Post
LithosZA
post Aug 25 2012, 15:49
Post #8





Group: Members
Posts: 202
Joined: 26-February 11
Member No.: 88525



Yes, I playback some 24 bit files. I thought it won't apply dithering when the source is already 16 bit?
Go to the top of the page
+Quote Post
Nichtswisser
post Aug 25 2012, 16:02
Post #9





Group: Members
Posts: 7
Joined: 21-August 12
Member No.: 102516



QUOTE (LithosZA @ Aug 25 2012, 16:49) *
I thought it won't apply dithering when the source is already 16 bit?


Not sure about that. Does anyone know for certain?

This post has been edited by Nichtswisser: Aug 25 2012, 16:02
Go to the top of the page
+Quote Post
pdq
post Aug 25 2012, 16:41
Post #10





Group: Members
Posts: 3450
Joined: 1-September 05
From: SE Pennsylvania
Member No.: 24233



QUOTE (Nichtswisser @ Aug 25 2012, 10:43) *
QUOTE (LithosZA @ Aug 25 2012, 11:09) *
Output: WASAPI, 16 bit, Dither ticked



Why dither at all? Dithering is only required when you reduce the bit depth of sounds files, 24bit to 16bit for example. It's supposed to be inaudible most of the time yet I would not use in cases it isn't needed.

Unless the data are being passed unmodified, if any kind of processing has taken place, even just adjusting the volume, technically dither should be added. It is unlikely though that you would be able to hear the difference.
Go to the top of the page
+Quote Post
skamp
post Aug 25 2012, 17:55
Post #11





Group: Developer
Posts: 1453
Joined: 4-May 04
From: France
Member No.: 13875



QUOTE ([JAZ] @ Aug 25 2012, 14:02) *
usually in linux, 0dB is more about 70% or so


[citation needed]

It's always been 100% for me, and I've used linux for 10 years with maybe a dozen sound cards (with ALSA drivers). I never used Pulse Audio or any software like that though, just straight up ALSA.

This post has been edited by skamp: Aug 25 2012, 17:57


--------------------
See my profile for measurements, tools and recommendations.
Go to the top of the page
+Quote Post
wwphil
post Aug 25 2012, 22:28
Post #12





Group: Members
Posts: 2
Joined: 25-August 12
Member No.: 102647



Thanks all for the responses, so I guess I can agree that the resampling is inaudible, correct me if I'm wrong but if that translation is inaudible, that should also mean the higher sampling rates are useless, as far as listening pleasure is concerned ?

I believe resampling must be done by Win7 as my dac shows up as using a microsoft driver under device manager, and I doubt it recognizes any other format than PCM.
On the other hand, the Mini-i does have an internal clock sync which is normally used for resampling, right ?

my setup:
foobar2000 (stock playback settings)
Matrix Mini-i connected through USB@16/48
Bel Canto S300
B&W 705 - Martin Logan Dynamo 500

Go to the top of the page
+Quote Post
Porcus
post Aug 25 2012, 22:59
Post #13





Group: Members
Posts: 1993
Joined: 30-November 06
Member No.: 38207



QUOTE (wwphil @ Aug 25 2012, 23:28) *
that should also mean the higher sampling rates are useless, as far as listening pleasure is concerned ?


In theory it is useless yes. There may be exceptions in circumstances where the device you send it to, handles 44.1 poorly.


QUOTE (wwphil @ Aug 25 2012, 01:46) *
the content is 96khz, so I switched it to 48khz thinking "at least I'll get half the resolution".


96 kHz just has an extra inaudible octave above the treble. That 'half' you are missing is of no value. (Well it may even be harmful because it can introduce distortion, and that distortion could be audible.)


--------------------
One day in the Year of the Fox came a time remembered well
Go to the top of the page
+Quote Post
[JAZ]
post Aug 25 2012, 23:14
Post #14





Group: Members
Posts: 1797
Joined: 24-June 02
From: Catalunya(Spain)
Member No.: 2383



QUOTE (skamp @ Aug 25 2012, 18:55) *
QUOTE ([JAZ] @ Aug 25 2012, 14:02) *
usually in linux, 0dB is more about 70% or so


[citation needed]

It's always been 100% for me, and I've used linux for 10 years with maybe a dozen sound cards (with ALSA drivers). I never used Pulse Audio or any software like that though, just straight up ALSA.


Mmm... It happened to me, with ALSA and some integrated soundcards (mostly realtek). I haven't used it for years, so I can't comment on the current situation, but maybe it isn't as usual as I thought.
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 19th December 2014 - 17:54