48 vs 96 vs 192

[escknx]

It might be discussed billion times here, but I tried to search for similar thread with no luck.

I definitely hear the difference between 48 vs 96khz playback even from Google Music source.

If DAC allows 192khz, is there any reason to keep your output on lower frequency than 192khz?

Um yes, and 16 vs 24 vs 32 bits. What is the purpose of 16 option, if I have 32?

 

[cossix]

If you're running Windows, the options for bit depth and sample rate should ideally be matched with the music file you're playing. If you're playing mostly 24 bit 192khz music files, you should set your options as such. If you have your options set to something like 24/192 but play a file that is 16/44.1 then Windows will upsample the song which may result in a drop in quality because windows is using it's own software to "fill in the gaps" and artificially make the file 24/192.

If you use WASAPI or ASIO drivers, which is recommended if you're looking for the best sound quality, then Windows "DirectSound" as it's called, will be bypassed and the raw file will be sent to your DAC, which will decode it and give you it's native resolution without resampling.

 

[Music Alchemist]

I definitely hear the difference between 48 vs 96khz playback even from Google Music source.

Convert a hi-res file to lower lossless resolutions (so 24-bit / 96 kHz to 16-bit / 44.1 kHz, etc.) with a program like dBpoweramp and I can pretty much guarantee the differences you're hearing will disappear. (In my experience, the difference is caused by different masters of a recording, so you have to isolate the variables.)

And yeah, just use bit-perfect output in your player (if possible) to send an exact copy of the data to your DAC. If you're using a streaming service that does not support bit-perfect output, you can set your computer settings to the settings of the file you're playing at the time.

 

[buke9]

It might be discussed billion times here, but I tried to search for similar thread with no luck.

I definitely hear the difference between 48 vs 96khz playback even from Google Music source.

If DAC allows 192khz, is there any reason to keep your output on lower frequency than 192khz?

Um yes, and 16 vs 24 vs 32 bits. What is the purpose of 16 option, if I have 32?

I think you might be telling yourself that you hear a difference in 48 and 96 as there should be none as Google is putting out 96kbs to a max of 320 kbs and that is MP3 and that is 16 bit lossy audio. You could be hearing the difference in the upsampling but you can't make a MP3 sound better than what it is. Just because a dac can do 32 bit it doesn't mean that it will sound better than a 16 bit only dac. Probably 98% of the music out there is 16 bit as that is what MP3's are . You can not make a audio file what it is not. You have to listen to the file at the resolution that it was made. There are audio tests out there that have high rez and MP3 audio and see if you can pick which one is which that would be a better test than you saying you can hear the difference. There is no 32 bit commercial audio out there or I have not seen it yet but still most of the 24 bit stuff I have doesn't sound that much better to garner the added price as most are just louder and most think they sound better. Unless you have very high end gear I doubt that you would hear the difference and at that no real need to spend more for just a little upgrade IMHO.

 

[Music Alchemist]

I think you might be telling yourself that you hear a difference in 48 and 96 as there should be none as Google is putting out 96kbs to a max of 320 kbs and that is MP3 and that is 16 bit lossy audio. You could be hearing the difference in the upsampling but you can't make a MP3 sound better than what it is.

kHz (sample rate) and kbps (bit rate) are two different things.

I don't use streaming services. I looked up Google Music hi-res (etc.) on Google. Some sources say they only stream lossy, while others say it supports hi-res now. (It could be different Google services involved.) Either way, like I covered, once you get into lossless, it won't matter whether it's 16/44 or something higher; it's the master that can sound different, rather than the file resolution.

 

[buke9]

That is what I was getting at as I know what I said was kbs rather than khz. As far as I know all streaming is lossy even MQA from Tidal is lossy .

 

[Music Alchemist]

That is what I was getting at as I know what I said was kbs rather than khz. As far as I know all streaming is lossy even MQA from Tidal is lossy .

What I'm saying is that the OP was referring to kHz, not kbps. We'll have to wait for an explanation on that, though.

 

[cossix]

I do enjoy when people take the lossy vs lossless test. The one I used to do gave multiple samples of the same song and randomized the order of which format it gave. I think it was 320kbps mp3 and 16/44.1 flac. I tried it a few times on my 650 and consistently got 8/10 correct which I don't think is bad when the 650 is very forgiving of poor recordings.

 

[Brooko]

I do enjoy when people take the lossy vs lossless test. The one I used to do gave multiple samples of the same song and randomized the order of which format it gave. I think it was 320kbps mp3 and 16/44.1 flac. I tried it a few times on my 650 and consistently got 8/10 correct which I don't think is bad when the 650 is very forgiving of poor recordings.

The problem with those tests a lot of the time is that they can be:
- different masters
- slightly different volumes (and we hear louder as better)
- comparison of mp3 at 128 kbps vs lossless, so everyone says lossy is rubbish

In reality, most people can't (statistically consistently) tell aac256 or mp3 320 from lossless if its properly transcoded from the same master, double blind tested, and volume matched. There are a few killer tracks which can be exploited, but these are pretty rare, and exceptions rather than rules. Do a search over on Hydrogen Audio - lots of good stuff there if you are interested.

 

[Brooko]

It might be discussed billion times here, but I tried to search for similar thread with no luck. I definitely hear the difference between 48 vs 96khz playback even from Google Music source.

If OK with you I'll take this with a grain of salt. If you're talking hi-res 24/48 vs 24/96, and the master is the same, the track has been properly transcoded without introducing artifacts, and you're blind testing properly including the tracks being properly volume matched - then with respect, no-one can (or at elast there are no known instances of someone being able to do this). If you can - using above conditions I just stated - and can do it in proper conditions, you're about to become famous Also - Googel Music (to the best of my knowledge) is streamed at max res 320 mp3 or aac256 (please someone correct me if I'm wrong) - both are lossy.

If you want to actually test yourself - this might help setting up a test - you don't need fancy equipment either. Its an eye opener (or rather an ear opener), and really worth it if you want to find out your own personal transparency level.

If DAC allows 192khz, is there any reason to keep your output on lower frequency than 192khz?

Yes - you can actually degrade the quality. Monty describes it better than anyone I know. He knows what he's talking about too - this guy is not a crackpot
https://people.xiph.org/~xiphmont/demo/neil-young.html

Videos : https://xiph.org/video/vid1.shtml and https://xiph.org/video/vid2.shtml
Second one explains sample rates and bit depth

And if you want to blow your mind about what we think we hear - try this video


Worth watching the whole thing. The woman on the panel is Poppy Crumm (Chief Scientist at Dolby Labs!)

Um yes, and 16 vs 24 vs 32 bits. What is the purpose of 16 option, if I have 32?

You're talking about the bit depth - and most people don't understand what it actually is. In reality, 16 bit is all we need for music playback. 24bit is important for recording. Greg explains it better than I can here - https://www.head-fi.org/threads/24bit-vs-16bit-the-myth-exploded.415361/
Again, well worth reading. And again - the guy is an expert. Think recording studio and producer type expert. I can absolutely validate his credentials (I know his real-life persona - but not at liberty to disclose).

 

[escknx]

Thanks a lot, I got the idea.

If its possible to see whats sample rate for particular flac file, its not quite possible to see it from sources like Google Music, Spotify, etc.(((

 

[Music Alchemist]

If its possible to see whats sample rate for particular flac file, its not quite possible to see it from sources like Google Music, Spotify, etc.(((

Nearly all of the time, when you're dealing with online music streaming, it will be 16-bit / 44.1 kHz.

 

[escknx]

Yes, enabled sample rate column in foobar2k and now I see.
50/50 16/44.1 and 24/96
I understand that higher sample rate may harm.

Regarding bits per sample, according to xiph.org, 24bit over 16bit makes no sense because you won't be able to hear the difference, but can 24 bit per sample actually degrade actual sound quality over original recorded 16bit?

And how to determine if streaming device support bit-perfect output?
Foobar2k and Google Music as a players in Win10, using few DACs:
* Realtec ALC1220 w MSI Audio Boost4
* Asus Xonar U7 MK2 CIRRUS LOGIC CS4398
* Sound Blaster X7 Burr-Brown PCM1794
* Burson Conductor Virtuoso Sabre ES9018

Is there any universal way of telling if exact digital signal has been sent to DAC and is encoded natively? Assume all DACs can override manual rates and encode natively.. Cause In Sonic Control Panel of Asus Xonar there is no option for auto detection.

 

[Music Alchemist]

Yes, enabled sample rate column in foobar2k and now I see.
50/50 16/44.1 and 24/96
I understand that higher sample rate may harm.

Regarding bits per sample, according to xiph.org, 24bit over 16bit makes no sense because you won't be able to hear the difference, but can 24 bit per sample actually degrade actual sound quality over original recorded 16bit?

And how to determine if streaming device support bit-perfect output?
Foobar2k and Google Music as a players in Win10, using few DACs:
* Realtec ALC1220 w MSI Audio Boost4
* Asus Xonar U7 MK2 CIRRUS LOGIC CS4398
* Sound Blaster X7 Burr-Brown PCM1794
* Burson Conductor Virtuoso Sabre ES9018

Is there any universal way of telling if exact digital signal has been sent to DAC and is encoded natively? Assume all DACs can override manual rates and encode natively.. Cause In Sonic Control Panel of Asus Xonar there is no option for auto detection.

If you are using a music player to play files (such as foobar2000), just use bit-perfect output (such as ASIO, WASAPI, or KS) to output an exact copy of the file data to your DAC. Then your DAC will do its thing. (According to Rob Watts, the DAC designer for Chord, all DACs oversample, including so-called non-oversampling ones. This is necessary for digital to analog conversion.)

As for music streaming services (nearly all of which do not support bit-perfect output), just set your computer settings to 16-bit / 44.1 kHz and you're good.

Many recordings were recorded in 24-bit, but they are usually released as 16-bit. If you convert a 16-bit file to 24-bit, it should sound the same. That article mentioned that higher file sample rates (not sample size / bit depth) can in some cases degrade the sound quality.

 

[Brooko]

As M.A. suggested - you can go bit perfect using ASIO or WASAPI and let your DAC do the lifting. personally I can't hear a difference anyway (and doubt most people can) - so no harm setting DAC to either 24/96 or 16/44.1 and just let it upsample of downsample. The point of Monty's article is that there is no use going out and specifically looking for highest res recordings. You'll pay more for them for zero benefit.