Why is Roon limited to 24 bit instead of 32 bit

Best regards.

I am new to Roon, as is Roon. However I am a longtime J River user, and as such I can compare many features between the two.
Why is Roon limited to 24 bit files and not to 32 bit if the modern CPU’s ( i5 and over as recomended by Roon) can process them as can the modern d/a?

Also, is room correction and / or electronic crossover capabilities being considered for Roon?

Great starting product…

Are you going to be able to hear a difference !!!

Easily - it’s half as many bits again!


Sorry to ask. Does that mean you don’t have the answer?

Then why reply.

Maybe I can hear it. Its the same argument people state againt hi def formats in general. Afterall, cd is perfect sound forever. Or that LPs do not sound any better than CDs.

If we start that road, why go to the moon if we will not live there… or Why install Roon if one can not listen to more than one record at a time and a cd player is a perfectly capable reproduction machine, as are all in one table top radios.

1 Like

Sorry back to you but didn’t realise I was not allowed to ask a question.

Well, the ASIO driver for my computer sound card lets me set the bit depth at the driver level. I do not let JRiver do any processing in itself, so it is the ASIO driver which is padding all content less than 32 bit to 32 bit. Why? No particular reason. It doesn’t make a difference to the sound as any padding done is nothing but zeroes to fill up the stream.

Does JRiver have the ability to pad bit depth? I don’t know since I’ve never bothered to look. But, JRiver or Roon padding bit depth would not make a difference as there is really no change being done to the sound stream; just a little extra processing work. There are some that would argue that, making the DAC work less improves the sound; but I’m not one of them.

That being said, Roon has never been able to detect my ASIO driver so I’ve never been able to test if Roon down grades the stream from 32 to 24 bit, which is what I am assuming @yoyoc is discussing. I would suggest that he should make a post in the Support area and treat Roon’s down grading bit depth as a bug to let the dev’s know. My sound card’s drivers can do 24 or 32 bit no problem, are always down graded and dithered by Roon to 16 bit. Still hoping to work that out sometime as 24 bit music sounds much better in 24 bit than 16, imho.

My assumption is that 32bit is still a very small niche market. There are hardly any titles available as far as I know. Let alone even music mastered at 32bit.

Seeing at how roon has picked up for instance DSD support, I think they will seriously look at it as soon as 32bit gains some traction. Maybe MQA will be in the long term roadmap.

But best that roon answers your question.

Nick B… of course one can ask a question. If I misunderstood my mistake.
It did not sound like a question, but a rethoric question, which beats the purpose of my posting.

It is a small market, as is any HD format. Still most peope state that MP3 is a fine format standard.

In my system, I have some 24 bit music does sound better in 32 bits. even if its mofe zeroes, it means lower noise floor and larger dynamic range. Which in turn means more realistic sound.

A response was either withdrawn or edited out. I will reply anyways to the uncalled for post. I was going to flag it, but, somehow its gone.
As far as I know any PC with a modern CPU has the computational capacity to resolve 32 bits data, some even 64 bits, some d/a as the ones using the ESS chips have it too, as does the one I use ( Geek out uses ESS 32 bit d/a chips inside, also the ability to process dsd and 2x dsd).
Thus I don’t see why there would be a real question about my system having the capacity to resolve the difference between 24 bit stream of data and 32 bits stream, or whether it is state of the art or not. Probably the question was posed improperly by the poster. Maybe its lack of understanding.
Now, if the real question posed was whether my system is state of the art and has the ability to reproduce the sonic difference between the two. I already answered that I believe it does have it.
Finally, such a post denotes that my first assertion to the response was correct. I ask that no post going against the rules of the forum be posted in response to a valid question by me.
If another person’s system does not have the ability to reproduce what mine can, that is not my problem. It is my opinion that if a person has an insufficient system setup, or an organic incapacity to hear subtleties, that person should not take it on others, or spend money on a lifetime subscription to Roon. It would be a waste.

1 Like

From what I understand there isn’t a single DAC chip out there that can truly do more than 20 or 21 bits. I have no idea what the theory behind it is but I have read enough articles to know that there is no such thing as true 24bit, let alone 32bit, it’s all marketing hype. This is not intended to flame the OP, just what I have read over the years playing around with computer audio.

1 Like

I have not read such, but now that you mention it I will search for it. Can you point me to any of such articles?
Of course you might be completely right, but with the cheap computational power now available, seems odd to me.

It’s nothing to do with CPU power that, it is to do with how accurate the electronics in the DAC can perform the conversion to the analogue domain. Currently the best DACs in the world can only resolve to circa 22 bit accuracy. Thus anything over this is lost. There’s also been a lot of research recently that suggest that human hearing at it best can’t resolve anything higher than this granularity. I’ll have a dig around to see if I can find some of the white papers that are available online.


i will be interesting reading.

There is a very interesting gentleman who is a record producer and works only in high definition, his take on the current “hi-def” music is very illuminating. The site is called Real HD Audio - www.realhd-audio.com. His take is that anything above 24/96 is wasted as if the whole of your audio reproduction chain is not capable of playing at HD definition then you cannot hear the difference as it never gets to your ears. Sign up for his daily e-mails and look and listen to some of the samples on his site.



Thanks NickB. I will.

Correct. robbbby

First rhetorical question: Why?
Answer: Any attempt to quantize below 20bits is futile, because the component noise floor of even the best circuits is at the 20bits level.

Second rhetorical question: So then why use 24bits?
Answer: Computers process bits and bytes. In 1993 the ISO/IEC 2382-1 de facto standard became eight bits in one byte. We need 8 bits to encode a value. Since 16 bits is 2 bytes, next 8 bit increment is 24 bits (16 + 8 = 24), despite the last 4 bits being relatively useless.


1 Like

The extra bit are not “relatively useless” in the realm of mathematics. Mathematics is used a lot with DSP functions so they are quite nice to have around. In the end, when all the math is done, 21 or 22 bits is enough to pass on to any DAC but having 24 to pass on is not a problem.

You’re not wrong, but your point is under the assumption that all 24 bits are recorded accurately. As previous comment said,

I believe he same rule applies to microphone and other recoding devices. If the last several bits recorded are just noise, bringing them to DSP will not make any difference.

Yep. The 20 bits level is at the noise floor of the circuit. Mathematically beneficial, but not content beneficial. The remaining bits are dithered.

I don’t think you get what I am saying. You want the extra bits when doing any manipulation of the music data. That would include any upsampling, downsampling, volume adjustment, other DSP adjustments, etc. Once that manipulation is done, then the bits above 21 or 22 are not “content” beneficial. In other words, you would never want 20 bit files if you had 24 bit files available to you.

1 Like