[71 dB]

What do you mean 'let's assume?" Assume based on what research? Where did you get -20 dbFS? And who says that is "stupid waste of dynamic range???" As I have repeatedly shown, peak of 120 db SPL is supported in research of actual music halls. The threshold of hearing is below 0 dbFS, so the total dynamic range needed for transparency is 125 dB, not 94:



Reducing this to 94 dB means leaving 21 db on the floor. For no good reason whatsoever.

Is there any hope in our future that you guys argue based on proper references? Or are we doomed with assumption after assumption not verified in the slightest?

If an recording really needs (it doesn't, because 125 dB dynamic range in consumer audio is ridigulous idea) 125 dB, then shaped dither can do it, you see it in the picture. Are we done? Is there any hope in out future that you understand why 16 bit is enough in consumer audio?

 

[71 dB]

Noise is not music.

A lot of parents think the music their children listen to is noise. It doesn't matter how it's classified. What matters is the frequency content.

 

[71 dB]

Sitting here, you still have not read the research and yet have that opinion of it??? .

I recall reading those papers long ago. Familiar stuff. Unlike you, I didn't draw some crazy conclusions for 16 bit not beeing enough in consumer audio.

 

[RRod]

-35 dbFS is your minimum???? How did you determine that? Is it peak, RMS or average? What is the window size?

Here is the stats on my last clip I randomly posted:



You get numbers like that because there is fade to digital silence. If I select just the loud portions I get:



That is still well below your number. If I select a little less loud section I get this:




So we are down to -60 dbFS which is well below your -35 dbFS.

A full histogram shows the same:




And let's remember that these programs can NOT separate music from noise. We can hear sound/music well below noise floor so effective music dynamic range is higher than these programs show.

The above are also RMS values and not peak (really dips). And there is an averaging window.

This is why we don't want to split the peas this way. The method used by Fielder is to measure the lowest level noise of microphones and ADCs and use that as the lowest limit. After all, the music can get as soft as it wants.

Your two posts perfectly encapsulate the issues we are having with your arguments. In the ABX: What is your listening level? Anyone can ABX truncated 16 bit vs. 24 bit if you can arbitrarily up the pot, especially when you separate the reverb tail from the loud music right before it. For all you know this album has a full-scale 1kHz square wave thrown in for kicks somewhere. How are we supposed to interpret you result without any other loudness context?

Re RMS, -35dB RMS is for *the whole album*, and is indicative again of where I actually set my pot to listen to the material in my listening environment. The softest sections with musical material (non-fade/reverb) I've encountered are around -70dB for a 1s RMS. In essence you seem to think of listening-to-music in a way most of us don't. I think of setting the volume once (maybe with a bit of adjustment) in a library-quiet environment, as opposed to fiddling with the pot on fading tails with IEMS. Even then, what happens when your album is mastered at -20dB RMS instead of -35? That's either 2.5bits more quiet in the room or 2.5bits louder on the pot you need to hear the same tail issues. It never seems to end with you, and that's what doesn't make any sense.

 

[bigshot]

I wrote a very long post about technicalities of immersive audio so I guess then I may match your criteria of music alienation. I certainly don’t match any criterium of music erudition. And the immersive audio post received low feedback. Do you believe that immersive rendering of music is then somehow irrelevant and that such rendering will be restricted to the gaming and virtual reality niche?

Actually, I'm very interested in multichannel mixing techniques and I've posted about it here quite a bit. I think your post didn't get a lot of response because you overloaded the first post. Threads are discussions. You throw out a few points and see the reaction you get and go from there. If you put too much information up front, it becomes an article and you don't get much interaction. There should be a document of best practices for participation in internet forums. But I bet even if there was something like that, we'd still get the same sort of pointless nonsense we are getting in this thread!

 

[bigshot]

In the ABX: What is your listening level? Anyone can ABX truncated 16 bit vs. 24 bit if you can arbitrarily up the pot, especially when you separate the reverb tail from the loud music right before it.

You aren't going to get any kind of satisfactory answer to that question. I asked it a long time ago and got a song and dance.

 

[jgazal]

Actually, I'm very interested in multichannel mixing techniques and I've posted about it here quite a bit. I think your post didn't get a lot of response because you overloaded the first post. Threads are discussions. You throw out a few points and see the reaction you get and go from there. If you put too much information up front, it becomes an article and you don't get much interaction. There should be a document of best practices for participation in internet forums.

I see. You are absolutely right. I did it with good intentions. I thought it would be easier or more clear to readers to identify any mistake from my part if people knew exactly where I got each concept I was describing. And I wanted to get together in a single post all the insights you all gave in disperse threads. I will be more concise and patient to build one concept at each post. I am sorry.

 

[amirm]

I recall reading those papers long ago. Familiar stuff.

Then please link to some place you have posted them so that people have the full information.

 

[amirm]

You aren't going to get any kind of satisfactory answer to that question. I asked it a long time ago and got a song and dance.

Precisely what I get when I ask you to run a single blind test and report the results. You can start with the one in front of us.

 

[sonitus mirus]

You aren't going to get any kind of satisfactory answer to that question. I asked it a long time ago and got a song and dance.

It was already asked, and it was ignored.

https://www.head-fi.org/threads/24bit-vs-16bit-the-myth-exploded.415361/page-291#post-13854799

I'm a critical listener, hear me roar!

 

[bigshot]

Amirm, I've done lots of blind tests. I know my thresholds pretty well. 16/44.1 sounds exactly like 24/96. AAC 256 VBR is totally transparent, as is MP3 320 LAME. I can hear up to about 17kHz. Music with dynamics beyond 50dB is uncomfortable to listen to, and my normal listening levels are always under 80dB.

I think everyone who is serious about sound should know as much about the specs of human hearing as they do the specs of their equipment. If you don't know what you can and can't hear, you chase after numbers without a solid grasp of what they represent. It's also important to remove your ego from your listening tests. When you allow yourself to judge your self worth on your perceptual thresholds, you can be tempted to fudge the results by reaching for the volume knob or focusing in on tiny insignificant bits at the ends of tracks. That doesn't reflect any kind of practical application of knowledge. It isn't even knowledge. We're all human and we all hear music with human ears. The people who perceive things in music that other people can't are the ones who have studied music, not the ones that have studied audio specs.

We live in a great time for electronics. We can go to a Walmart and pick up a cheap player with perfect sound. Speakers and headphones are better than they've ever been. Quality standards continue to improve, even beyond our ability to hear the improvements. We're very lucky. We can go to Amazon and order just about any midrange piece of audio equipment they sell and have it delivered to our door... and it will perform to a very high standard. We don't need to worry about that sort of thing any more. We're free now to focus on our room acoustics, mixing and mastering and most of all musical creativity. Music is what matters.

 

[bigshot]

I see. You are absolutely right. I did it with good intentions. I thought it would be easier or more clear to readers to identify any mistake from my part if people knew exactly where I got each concept I was describing. And I wanted to get together in a single post all the insights you all gave in disperse threads. I will be more concise and patient to build one concept at each post. I am sorry.

No problem. It's an interesting subject and it really deserves its own active thread.

 

[amirm]

Re RMS, -35dB RMS is for *the whole album*, and is indicative again of where I actually set my pot to listen to the material in my listening environment. The softest sections with musical material (non-fade/reverb) I've encountered are around -70dB for a 1s RMS. In essence you seem to think of listening-to-music in a way most of us don't.

That is exactly it: I am not worried about "most of you" (putting aside that you have no data that represents most of you -- just assumptions). I am worried about all of you. As is the research.

When we talk about what the music industry should release to us, we need to be inclusive of the needs of everyone, at all listening levels, for all content, and all situations.

Now, if this was not achievable, OK, we fall back on what we can deliver. OP claims this should be 16/44.1 We absolutely can certainly capture and playback at > 20 bits and far faster sampling rates with no cost to any of us. So there can't be a motivation to back off on some arbitrary reason, or heaven forbid, job security for someone who thinks their job is to do that conversion for us.

Heck, if we want to go by what most people find adequate, lossy compression should be the limit. Why did OP say it has to be 16/44.1 lossless? Isn't what 99% of the world enjoys and so is more "real life" than lossless compression?

The argument gets slippery real fast. That's why we approach this topic methodically and determine what we can determine to be an audibly noise-free channel. We can prove and defend such a standard.

As for me, I have been in rooms including our own at work where my pant legs move with bass. I have also stood near my son playing drums and it doing the same thing. In neither case did I go deaf, or run right out of the room. Yes, the system is loud. This is not for everyone. But dynamics can be fun and it is not something we want to deprive people from based on averaging entire albums, etc. type of numbers. We are talking about instantaneous peaks and valleys. That is what our channel stores.

Answering your question regarding level of my ABX test, I don't know how to answer that as I am using headphones and I have no way of measuring levels there. I turned it up 'till I could hear what was there and did my ABX test.

What was the purpose of the test and can you run it and and post results of success or failure regardless of level? I am trying to get us calibrated subjectively.

Oh, I forgot to say that the DAC headphone combo I used retails for $60.

 

[RRod]

It was already asked, and it was ignored.

https://www.head-fi.org/threads/24bit-vs-16bit-the-myth-exploded.415361/page-291#post-13854799

I'm a critical listener, hear me roar!

Yeah, I figured I try again in the context of an actual example.

 

[amirm]

And to offset the people saying they don't listen loud, he is an online friend "basspig" with his system: