Almost missed this post, thanks for confirming.
So a DACT just might be an option for me. As the finer steps may still be digital, I just want to be able to attenuate to any level that might be fit for the occasion when I want that. Which means stepping out of the room to set that level, no biggy. If I need anywhere from 0.5 to 5 dB more or less, just grab the phone. But largely the digital room needed would be preserved under all conditions. By the way I do like that solution by Bruno. He's a smart cookie.
I can understand what he's doing, I just can't see me repeating that trick with the Universal Buffers or even more complex, build it myself to my liking.
I'm a mechanical engineer, with a lack of experience in the electronics world. Even though I'm learning as I go, can follow basic schematics etc. It's just not what I'm good at. The digital realm we use does not hold much secrets for me. Mechanically I can build whatever I can come up with. But there are simple reasons why I bought my amps from you.
Perhaps i could learn, but I'd rather just do what I do best. That is listening to the music and analysing where I can advance what I have for myself.
I'll probably first build what I said I would. 3x Universal buffers in a case. But it might be smart to leave room for a volume control like the DACT. If I feel that would gain me another step on this slippery slope we like to call progress
.
Thumb up had always myself drooled looking those nice hardware DACT units, but implementation is not in sync saing i'm not an electronics wizard, but guess you are humble there and really think you a master with your hands and able to familiarize yourself with the complicated technical solutions of the world
You know what makes it hard? Finding out what's true and what's superstition in this hobby. I'm on a side of the forum where most members are running, or chasing really, after the next great driver almost all the time, expecting to find audio nirvana that way. Possibly a rewarding hobby for lots of them, but my path is a little different.
I stick to what i've got, change one thing and evaluate. Most of the time things get tested at least more than once to know if I really arrived at the right conclusion. Slowly but surely it has taught me a lot, and made this thread a long long story and got me great sound.
Can anyone give us a short description of a world beating recipe for great sound? Ask 10 different people on this forum and you'll get 10 different answers.
Maybe my steps seem weird at times, but that's because I want to continue in a controlled way. Basically I'm still trying to better the setup i had with the Pioneer amplifiers. The Atom amp has shown the way, as the Goldmund clone sure is the better choice. But I've lost some time chasing my tail here looking to fix something with DSP progress while I should have gone with my first hunch, which was adding a pré amplifier. Looking back over our conversations off line that was all I talked about right after trying to get instant good results with the Goldmund without your HP-1 helping it out.
Aren't many digital processors using 40bit or better signal processing? For instance, the little Behringer mixer I mentioned does.
I don't think being down 15 to 20 dB is a problem then, even after allowing 17dB headroom. I'm down that much all the time given the high efficiency drivers being used. (with amps at +26dB).
I guess i was going yikes about the multi-pot, because I've had more trouble with pots going crusty over time, than any other problem with high-end home gear. Conrad-Johnson, even Manley. (I don't live in a corrosive environment)
But sorry if i yiked out of place
It really is all about what we want and enjoy, huh?
There is often a difference between hardware and software when it comes to DSP calculations and headroom. Many hardware systems are processing limited and will clip interstage. My Najda would and I read recently that MiniDSP automatically reduces the input volume when boost is added to avoid clipping. These boast bit depths that would seem to make that irrelevant but in hardware it is not always.
In Jriver the software is written to allow the full headroom as it is running on a computer for which 64 bit floating point calculations are not processor intensive, so you get 384dB, practically impossible to clip without trying very very hard.
How much of that can come out of a DAC depends on the bit depth the DAC works at. A 32 bit DAC like the Sabre has 12 bits between input and output linearity, so 72dB of attenuation can be applied before the DAC is losing resolution, the downside is that you also lose signal to noise ratio as nc535 pointed out as digital volume cannot attenuate the noise in the way an analogue volume control does.
It is the SNR reduction that changes the sound more than the change in resolution for these DAC's with 40dB being a more practical limit.
Wesayso's 24 bit DAC is not so lucky and all this comes in 8 bits (48dB), earlier, 4bits (24dB) being the change point.
Those stepped attenuators are nothing like a pot, they are detented switches, no wipers to get scratchy
Basically what I do need is the best way of getting a crude 6 channel volume control.
All I'd need is -5 dB, -10 db etc to up to -50 dB or something in 5 dB steps. At the highest quality level I could get, for those 6 channels.
Making it a pain to setup a good loudness curve in the process though. Which would be the next hurdle to take.
All I'd need is -5 dB, -10 db etc to up to -50 dB or something in 5 dB steps. At the highest quality level I could get, for those 6 channels.
Making it a pain to setup a good loudness curve in the process though. Which would be the next hurdle to take.
How many dB's of attenuation do you want to have analogue control over? It could be possible to switch the gain boost out of the buffers to remove whatever they are adding and if it was a fixed amount switched resistor attenuators could be built in, then you could use a few very high tolerance resistors, cheaper and more accurate than a stepped attenuator if it met your needs.I'll probably first build what I said I would. 3x Universal buffers in a case. But it might be smart to leave room for a volume control like the DACT. If I feel that would gain me another step on this slippery slope we like to call progress
Edit looks like we crossed over, 50 is still possible but would need more switch contacts.
There is often a difference between hardware and software when it comes to DSP calculations and headroom. Many hardware systems are processing limited and will clip interstage. My Najda would and I read recently that MiniDSP automatically reduces the input volume when boost is added to avoid clipping. These boast bit depths that would seem to make that irrelevant but in hardware it is not always.
In Jriver the software is written to allow the full headroom as it is running on a computer for which 64 bit floating point calculations are not processor intensive, so you get 384dB, practically impossible to clip without trying very very hard.
How much of that can come out of a DAC depends on the bit depth the DAC works at. A 32 bit DAC like the Sabre has 12 bits between input and output linearity, so 72dB of attenuation can be applied before the DAC is losing resolution, the downside is that you also lose signal to noise ratio as nc535 pointed out as digital volume cannot attenuate the noise in the way an analogue volume control does.
It is the SNR reduction that changes the sound more than the change in resolution for these DAC's with 40dB being a more practical limit.
Wesayso's 24 bit DAC is not so lucky and all this comes in 8 bits (48dB), earlier, 4bits (24dB) being the change point.
Those stepped attenuators are nothing like a pot, they are detented switches, no wipers to get scratchy
Exactly!
Now one way to move forward could be 32 bit DAC output. But what DAC to get?
Between the Xonar and the M1 (on mains), the M1 wins. Not once, not twice but 3 times tested including my family (without them knowing they are being tested).
So that too could be a next step. One I've certainly though about. A device like that Octo DAC seems interesting enough. But it's a huge step with risks.
I got the M1 DAC for next to nothing. Got it for doing a job for someone because I never figured I'd fork out the money needed for a DAC like that. Here it was a reward for a job well done, so it felt like being a free way to test if i liked it. And I did. I know there are DAC's for less and DAC's for way more than what it did cost at the time. But what I can't predict is what the next purchase is going to sound like, not without extensive testing, reworking the DSP a couple of times and living with it. Only after quite a bit of time I'd be able to conclude if it's better or worse. With lots of measurements to hopefully provide clues and answers. This is a hard hobby to have, by the time all of it fits our needs we might have grown deaf from old age!
There is often a difference between hardware and software when it comes to DSP calculations and headroom. Many hardware systems are processing limited and will clip interstage. My Najda would and I read recently that MiniDSP automatically reduces the input volume when boost is added to avoid clipping. These boast bit depths that would seem to make that irrelevant but in hardware it is not always.
In Jriver the software is written to allow the full headroom as it is running on a computer for which 64 bit floating point calculations are not processor intensive, so you get 384dB, practically impossible to clip without trying very very hard.
How much of that can come out of a DAC depends on the bit depth the DAC works at. A 32 bit DAC like the Sabre has 12 bits between input and output linearity, so 72dB of attenuation can be applied before the DAC is losing resolution, the downside is that you also lose signal to noise ratio as nc535 pointed out as digital volume cannot attenuate the noise in the way an analogue volume control does.
It is the SNR reduction that changes the sound more than the change in resolution for these DAC's with 40dB being a more practical limit.
Wesayso's 24 bit DAC is not so lucky and all this comes in 8 bits (48dB), earlier, 4bits (24dB) being the change point.
Those stepped attenuators are nothing like a pot, they are detented switches, no wipers to get scratchy
Thx, good DAC explanations.
Yep, the last high quality analog pre-amp I got has a stepped attenuator volume control. Never given any noise/trouble at all, other than I wish the steps were finer at the lower levels.
I've also thought of a relay switch attenuator... covering 6 channels means it is still hard to find 'the' solution. The right answer could be a 32 bit DAC
. But as said, it is a gamble. What if I don't like it after paying a XXXX.XX number for it. I'm not the person who's going to like it because of it's price tag. I either like it, grow to like it or it has been a disaster. I think a Topping D50 is a safe and cheap bet for dipping your toe into the Sabre waters, as I said before it has my slight preference over my other two sabre DAC's.
It is cheap solid small black and sounds good. The cost of entry is not that high and if you don't like it you might be able to send it back or sell it and not lose too much. It would also be a great desktop headphone DAC if you didn't like it as much as the M1. Perhaps worth testing first before going down the analogue rabbit hole
If you get one switch the filter away from the default to linear before you listen, they have followed fashion rather than sense and set the default to an MQA style filter.
Anybody put credence in the claims that best achievable system S/N ratio is about 124dB.
And that a 24bit DAC has the capability of doing up to about 145dB SNR.....?
If all that is true, why would we ever need to jump to 32bit?
Here's a well made test from a 24 bit RME babyface pro DAC..
It shows half volume, -32dB....so i thought it's worth posting...
And that a 24bit DAC has the capability of doing up to about 145dB SNR.....?
If all that is true, why would we ever need to jump to 32bit?
Here's a well made test from a 24 bit RME babyface pro DAC..
It shows half volume, -32dB....so i thought it's worth posting...
Attachments
I know and I agree with you, that in your case the safest way to go is with a DACT sort of analog vol. control and use the digital for the finer steps. I just felt a need to tell what my experience was and what I ended up with in my system. The Bruno way to go is for electronic diy nerds as myself and not for everyone.Almost missed this post, thanks for confirming.
So a DACT just might be an option for me. As the finer steps may still be digital, I just want to be able to attenuate to any level that might be fit for the occasion when I want that. Which means stepping out of the room to set that level, no biggy. If I need anywhere from 0.5 to 5 dB more or less, just grab the phone. But largely the digital room needed would be preserved under all conditions. By the way I do like that solution by Bruno. He's a smart cookie.
I can understand what he's doing, I just can't see me repeating that trick with the Universal Buffers or even more complex, build it myself to my liking.
I'm a mechanical engineer, with a lack of experience in the electronics world. Even though I'm learning as I go, can follow basic schematics etc. It's just not what I'm good at. The digital realm we use does not hold much secrets for me. Mechanically I can build whatever I can come up with. But there are simple reasons why I bought my amps from you.
Perhaps i could learn, but I'd rather just do what I do best. That is listening to the music and analysing where I can advance what I have for myself.
I'll probably first build what I said I would. 3x Universal buffers in a case. But it might be smart to leave room for a volume control like the DACT. If I feel that would gain me another step on this slippery slope we like to call progress
I'm too lazy download and read their manual for that DAC but in its named Pro will say there's a chance they implemented a mix of digital volume control combined relay switched analog gain, think they are specialist to do that trick because they care but also its not low cost and we can see over ASR site that no other brand can pull them off 1st place in the head phone low 50milivolts SNR test
I've also thought of a relay switch attenuator... covering 6 channels means it is still hard to find 'the' solution. The right answer could be a 32 bit DAC
About 32 bit DAC possibility, If you havent ordered those buffers yet you can call Okto and see if its possible join the touring unit they have around EU, that is free except it cost shipping and users time for a week or two it will take.
Last edited:
Now if i really had two weeks time to play with it... sadly I do have to work etc.. and barely get anything done these days .
First step for now: get 3 buffers to play with (waiting for Tom's additional power supply to start playing with it).
Leave room for a DACT solution there, but might never need it down the road.
.First step for now: get 3 buffers to play with (waiting for Tom's additional power supply to start playing with it).
Leave room for a DACT solution there, but might never need it down the road.
the limited time sadly is true, but I do get to enjoy what's there already. In the build up of the coming F1 season I've been watching the series: Drive to Survive. What beauty is that on the bigger screen we have now with surround sound! I wish races were that exiting!
It's lot of fun to be able to enjoy it too. Even if I'd really want more play time to advance faster. I should never forget what i have right now. It is a joy to experience!
It's lot of fun to be able to enjoy it too. Even if I'd really want more play time to advance faster. I should never forget what i have right now. It is a joy to experience!
Not if you just want to replace the M1 and keep the rest the sameBut I do need 6 channels...
I run the D50 fed from the coax digital output of the Scarlett sometimes in preparation for the multichannel needs of ambience and subs. The ambience and subs could still run as they are now, perhaps worth a try before spending a large amount of money on a different route?
The Johnson/Thermal/self noise of resistors is a limiting factor in the analogue stage of audio devices. Most good ones set the signal to noise limit at around 125dB. This is one of those limits that is very hard to improve upon some designers can get a few more out. It is also possible to measure below the noise floor with precision oscillators.
That image seems misleading to me as it looks like it is measuring the noise, that is the part that does not change with attenuation but what does is the signal to noise ratio so instead of the difference being 0 to -140 it is -32 to -140.
I have attached the ESS Digital vs Analogue Presentation in zipped pdf, that is the clearest explanation with diagrams I have seen that describes what happens. You can see the noise as the baseline with the signal being the thing that changes.
32bit dacs have benefits when digital volume control is being used, without it in a normal stereo system where there is little to no processing 24bit is enough.
Not if you just want to replace the M1 and keep the rest the same
The ambience and subs could still run as they are now, perhaps worth a try before spending a large amount of money on a different route?
I'd really like to find a solution covering all of it, all 6 channels.
Even on the subwoofers I do use processing, as a sort of Linkwitz transform.
How else am I going to get what I'd want from 12" subwoofers in a pretty small enclosure. Always having to deal with those compromises, you know...
Trying to keep things small has drawbacks...