[Grimbles]

I find Divide Album sounded great on my Mojo, Bifrost Multibit DACs and Yamaha HS7 speakers and Campfire Audio Vega IEM. Maybe your IEM isn't a good pairing with the dragonfly.

Thanks theveterans, quite possibly right and i am currently in the process of upgrading my kit. Bifrost is on my list so hopefully ill find a more listenable option as i enjoy that album! I think this further exemplifies my point though - there is a lot of subjectivity to this and "better" by numbers may not always be more enjoyable to the ear.

 

[theveterans]

Of course we perceive sound differently so what sounds great to me might sound harsh and too peaky for you. Then again, I’m used to listening around 90 decibels (C curve) so high levels of upper mids and treble doesn’t bother me unless they sound unnatural of course

 

[Brooko]

Then again, I’m used to listening around 90 decibels (C curve) so high levels of upper mids and treble doesn’t bother me unless they sound unnatural of course

Wow - that is seriously quite loud. My usual listening level is 65-75 dB. Anything above 80-85dB tends to exacerbate my tinnitus.

 

[Currawong]

Read the link I left - there is simply no point.
https://www.head-fi.org/threads/24bit-vs-16bit-the-myth-exploded.415361/

You're confusing computer settings with arguments about the benefits of high-res music. Setting a computer to 24-bit or 32-bit will have no effect. The extra bits will just be padded with zeros. The only exception is software that specifically dithers 16-bit data up to 24-bit, but that's not what the OP is talking about when it comes to computer settings.

Setting Windows to output at a higher sample rate, eg: 96k, will cause Windows to perform up-sampling, which could be sonically detrimental, depending on the gear used.

 

[Brooko]

You're confusing computer settings with arguments about the benefits of high-res music. Setting a computer to 24-bit or 32-bit will have no effect. The extra bits will just be padded with zeros. The only exception is software that specifically dithers 16-bit data up to 24-bit, but that's not what the OP is talking about when it comes to computer settings.

But Amos - what is the point? All you are doing is padding it with zeros. So there is no gain. So why would anyone suggest this "Set the bit depth to the highest option. The reason is that converting 16 bit audio up to 24 or 32 bit has no negative impact on sound quality, so there's no reason not to set it to the highest."

And by setting the output in Windows mixer to 24/192 - Windows IS resampling and dithering to that bit-depth and sample rate. So again - what is the point?

 

[Currawong]

As per my post, in case you missed the edit.

 

[theveterans]

Wow - that is seriously quite loud. My usual listening level is 65-75 dB. Anything above 80-85dB tends to exacerbate my tinnitus.

It's the only way to enjoy a relatively flat response speakers for me: I have to crank the volume to get that club level sound environment.

 

[MindsMirror]

If your music format is redbook (16/44.1) - and I have to assume that the OP's is from his comments - and especially when you read the title of the thread.


We know that the difference in bit depth has to do with dynamic range - and that 16 bit is all you'll ever need for playback (so he's covered already). What possible use would there be to resample to the highest possible setting in Windows, and what possible advantage is there - when he is already covered at 16 bit?

You seem to be arguing semantics - when it was you who said


So 'll ask instead - what ADVANTAGE is there in the "sage" advice you gave? Why do you need to set it to the highest? Please be detailed.

You gain nothing out of setting a higher sample rate or bit depth (it doesn't magically restore data missing from your source files in the first place). If anything, it's taxing on the hardware and a potential problem source. By default, the windows mixer resamples all audio to the specified sample rate. Most audio is distributed as 44100khz/16 bit or 48000khz/16 bit, therefore it would be more logical to use that.

But Amos - what is the point? All you are doing is padding it with zeros. So there is no gain. So why would anyone suggest this "Set the bit depth to the highest option. The reason is that converting 16 bit audio up to 24 or 32 bit has no negative impact on sound quality, so there's no reason not to set it to the highest."

And by setting the output in Windows mixer to 24/192 - Windows IS resampling and dithering to that bit-depth and sample rate. So again - what is the point?

You are confused. Changing the bit depth does not require resampling. I never said to set the sample rate to the highest option. I specifically said to set the sample rate to match what you listen to most often, to avoid resampling. I also never said that increasing the bit depth restores missing data. I said it does no harm. Unless your computer is from the 1980's and runs on a 16 bit architecture, there is no situation where setting your output to 24 bit is worse than setting it to 16 bit. Here are a few situations where setting it to 16 bit is strictly worse than 24 bit.

• Playing anything that is not encoded in a lossless PCM format.
  
• Playing anything encoded in a lossless PCM format at higher than 16 bits.
• Playing 16 bit lossless PCM audio at anything other than 100% volume.
• Playing anything that does not match the chosen sample rate
• Using any DSP or EQ.
• Mixing multiple audio sources.

You undoubtedly will do many of these on a regular basis. In all of these situations, by setting it to 16 bit instead of 24 bit you are essentially saying "Windows, I want you to add -96dB of quantization noise to my audio instead of -144dB."

There is only one very specific and restrictive situation where choosing 16 bit 44.1KHz is not worse, and by not worse I mean the result is exactly the same as choosing 24 bit 44.1KHz, that is:

• Playing lossless 16 bit 44.1KHz PCM audio exclusively from one source with no DSP, mixing, or volume control allowed, i.e. "bit perfect" output.

 

[Brooko]

I don't know why you're making such a big deal out of this - is it about just "scoring points" - I'm getting a bit sick of it to be honest. But lets play the game.

I am not confused. You don't know me, nor the extent of my knowledge. Note that I haven't labeled you - so please have the courtesy to avoid doing so to me. Thanks. I also didn't say it was better or worse - I simply said it was not required (in the instance the OP described) to set the bit depth to the highest possible setting.

Now lets go through your bullet points:

Here are a few situations where setting it to 16 bit is strictly worse than 24 bit.
Playing anything that is not encoded in a lossless PCM format.

Why would it be worse - which is what you stated? Its a lossy file - so 99% of the time it'll be 16 bit depth at 44.1. What possible audible gain are you going to get by setting it higher? Answer - there is none.

Here are a few situations where setting it to 16 bit is strictly worse than 24 bit.
Playing anything encoded in a lossless PCM format at higher than 16 bits.

Well unless he specifically wants to run in bit perfect (WASAPI/ASIO), again how can it be worse when there is enough dynamic range in 16bit that any difference is going to be inaudible.

Here are a few situations where setting it to 16 bit is strictly worse than 24 bit.
Playing 16 bit lossless PCM audio at anything other than 100% volume.

Again - go and do some real world tests and see if you can hear quantization noise - for normal playback you won't.

Here are a few situations where setting it to 16 bit is strictly worse than 24 bit.
Playing anything that does not match the chosen sample rate

Who's getting confused - pretty sure we were discussing bit depth.

Here are a few situations where setting it to 16 bit is strictly worse than 24 bit.
Using any DSP or EQ.
Mixing multiple audio sources.

I'm quoting these two together. The OP mentioned listening - nothing about recording or mixing samples. Even with EQ or DSP (and if he's asking the sorts of general questions he is, any EQ would be relatively simple don't you think), as long as he's not applying cumulative EQ and mixing multi-layered tracks, he's not going to need 144 dB of dynamic range. I agree the higher bit depth if he's mixing and recording, and there it is definitely worth going to 24 bit - but you're using a specific case which the OP didn't mention.

Now - can I make a suggestion - none of this is helping the OP at all. I stand by my initial statement that there is no need to set the bit depth for a listening environment to above 16 bits (I will add here "unless specifically required for recording / complex processing). For the majority of his use, 16 bit will fully capture the dynamic range for everyday playback, and any difference is going to be inaudible. If you wish to carry on this debate, I suggest we take it to Sound Science, because all its doing is creating a point scoring debate here.

 

[MindsMirror]

I don't know why you're making such a big deal out of this - is it about just "scoring points" - I'm getting a bit sick of it to be honest.

No one is forcing you to read or contribute to this thread. If you are sick of this discussion then by all means don't let me keep you. I'll just post my replies for anyone else who's interested.

In all of the points I listed, the audio is processed in some way. Before processing it is converted to a higher bit depth, 32 or 64 bit floating point, to minimize errors. After processing it is converted back down to your chosen bit depth, where dither or quantization noise is added.

Why would it be worse - which is what you stated? Its a lossy file - so 99% of the time it'll be 16 bit depth at 44.1

Lossy formats are not stored as PCM audio. They only become PCM after decoding, which will be done at 32 or 64 bit floating point and then converted to 16 or 24. There's no reason to choose 16 bits over 24 bits at this step.

Who's getting confused - pretty sure we were discussing bit depth.

Still you I'm afraid. The resampling process is done at a higher bit depth. Afterwards it is converted back down to your chosen bit depth. Again there is no downside to going to 24 instead of 16.

What possible audible gain are you going to get by setting it higher?

I didn't say it would be audible. However there are real situations where it could be. Say you have a sound card with 2VRMS output, like a Soundblaster Z which is quite common. You're using ER4SR, a reasonably common IEM. It has a sensitivity of 98dB @0.1VRMS (124dB at 2VRMS). Set your output to 16 bits, and the noise floor comes in at 96dB below the max output of 2V. With the ER4SR that puts it at a SPL of 28dB which would certainly be audible given the high isolation of the IEM. The Soundblaster Z lists its SNR as 116dB, so if you set the output to 24 bits it will reduce that noise to a SPL of 8dB, a noticeable real world improvement.

Now - can I make a suggestion - none of this is helping the OP at all. I stand by my initial statement that there is no need to set the bit depth for a listening environment to above 16 bits (I will add here "unless specifically required for recording / complex processing). For the majority of his use, 16 bit will fully capture the dynamic range for everyday playback, and any difference is going to be inaudible. If you wish to carry on this debate, I suggest we take it to Sound Science, because all its doing is creating a point scoring debate here.

This discussion is relevant to the topic of the thread, the OP is not the only one who will read it. If this thread belongs in sound science I presume the moderators will move it there.

 

[Brooko]

I didn't say it would be audible.

And there you go ........ my point exactly. Thanks