How is doing attenuation in the DAC any different than doing attenuation in the DSP? I agree you should level match your drivers / speakers but why does it matter where you do it when both methods are attenuating digitally?
Michael
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Topping DM7 8-Channel DAC Review
- Thread starter amirm
- Start date
How is doing attenuation in the DAC any different than doing attenuation in the DSP? I agree you should level match your drivers / speakers but why does it matter where you do it when both methods are attenuating digitally?
Michael
Think about it...
I am slow, why don't you explain?
Do you think that your MOTU is not attenuating digitally when you apply an output channel trim?
Michael
I have spent way to much time this past week playing with this stuff though using my Okto not a topping. Same thing.
I've been doing it without the luxury of Acurate or Audiolense to get gain matching going between W/M/T +2 subs and have been experimenting with different XO structures with various slopes. If the change is dramatic I need to re-address gains. I do it all in software in the measurements because if the levels were static in the DAC it would be far more difficult, confusing and time consuming with no benefit. All of my amps have similar output gain so it's mostly taming the sensitivity of the drivers. I tried and measured 10 variations on XO structure to see how the room would respond. Very informative but took gobs of time.
I realize my use case is fairly unique and most will just "set-and-forget" with their convolution filters but it demonstrates why we do things the way we do.
I've been trialing Mitchco's multi-ch Hang Loose Convolver and it's got great promise. Otherwise I'm doing everything in REW, rePhase and Audacity to create my 8ch WAV files. I started a Dirac trial last week too, it's been fun watching Dirac try to deal with totally broken XO files. I've had good results but there's just not enough user control in Dirac for a tinkerer like me.
Just in case anyone is getting itchy to try out a Topping or other multi-ch DAC my signature links to my post detailing how I set up convolution files and XO. Systemwide signal correction, works on Mac & PC.
I've been doing it without the luxury of Acurate or Audiolense to get gain matching going between W/M/T +2 subs and have been experimenting with different XO structures with various slopes. If the change is dramatic I need to re-address gains. I do it all in software in the measurements because if the levels were static in the DAC it would be far more difficult, confusing and time consuming with no benefit. All of my amps have similar output gain so it's mostly taming the sensitivity of the drivers. I tried and measured 10 variations on XO structure to see how the room would respond. Very informative but took gobs of time.
I realize my use case is fairly unique and most will just "set-and-forget" with their convolution filters but it demonstrates why we do things the way we do.
I've been trialing Mitchco's multi-ch Hang Loose Convolver and it's got great promise. Otherwise I'm doing everything in REW, rePhase and Audacity to create my 8ch WAV files. I started a Dirac trial last week too, it's been fun watching Dirac try to deal with totally broken XO files. I've had good results but there's just not enough user control in Dirac for a tinkerer like me.
Just in case anyone is getting itchy to try out a Topping or other multi-ch DAC my signature links to my post detailing how I set up convolution files and XO. Systemwide signal correction, works on Mac & PC.
- Joined
- Jan 7, 2020
- Messages
- 959
- Likes
- 1,279
Thanks @mitcho - that makes sense. In the okto you can adjust levels through the dac itself via the menu control panel and dont need to do it in the driver software interface. That is one sdvantage over the Topping.
The problem with different amps with different gain strucures is they are often not linear through the volume range. While setting the levels in the dac at one master volume the relative levels at different master volume settings wont be necessarily be the same. For a multichannel active speaker setup amp and driver matching is critical.
Given the topping is using ess dac chips should be easy to solve this issue hopefully.
What? No way this is a true statement. Do you have an example of such an amp?
It would sound absolutely ridiculous if you experienced non-linear output levels depending on input level.
Michael
- Joined
- Jan 7, 2020
- Messages
- 959
- Likes
- 1,279
Yeah, my issue was with so called "custom" input buffers for hypex modules. I bought into the hype several years ago. That mistake has been fixed after much pain!
Edit: by the way not suggesting nonlinear relationship (although that was a problem I encounteted), but is the linear slope of the input level to output level the same for different amps with diffetent gain structures?
That is how an amp works, it takes an input and multiplies it which is a linear operation.
Here is a slightly simplified (ignores amplifier clipping) actual example. I was using amps with 26 dB gain for all my drivers. I switched the sub amps from using a single stereo amp to dual amps in BTL configuration increasing the gain to 32 dB. In order to maintain the same balance between my subs and everything else I needed to increase the output level of all my other amps by 6 dB.
Prior to changing the subs my tweeters were at -8 dB compared to my subs. Maximum output level with a 2 V input to the tweeters was 2 x 10^(-8/20) x 10^(26/20) = 15.9 V, to the subs it was 2 x 10^(26/20) = 39.9 V. As a sanity check you can see the relative level is 20 x log (39.9 / 15.9) = 8 dB. If I decrease the input level to a 0.25 V input that relationship still holds. 0.25 x 10^(-8/20) x 10^(26/20) = 1.99 V to the tweeter, 0.25 x 10^(26/20) = 4.99 V to the sub, relative is still 20 x log (4.99 / 1.99) = 8 dB.
After switching the sub amps I changed the tweeter attenuation to -2 dB to maintain the relative level difference of 8 dB, I have now increased the maximum output level of the system but the relative values stay the same. Again looking at a 2 V input, 2 x 10^(-2/20) x 10^(26/20) = 31.7 V to the tweeter and 2 x 10^(32/20) = 79.6 V to the sub, relative is still the same at 20 x log (79.6 / 31.7) = 8 dB.
So if you know the amplifier gain difference you can exactly compensate in DSP and the relative output levels stay consistent regardless of input levels.
Michael
Yes, I fully agree with you on this really important (but rather infrequently discussed) point!
Please note, however, that my perspectives on this point are not related to your linear/non-linear volume control, but based on power matching and (subjective) sound quality matching.
I actually had rather long and intensive amplifier exploration journey in my multichannel project (you may find Hyperlink Index here and here); the summary of my amplifier exploration can be found here. My latest system setup as of May 30 2022 can be found here.
(Almost) all of the home-use Hi-Fi amplifiers, I mean integrated amps and power amps, are designed for full range operation, i.e. to cover ca. 20 Hz - 30 kHz. This means that we should be very much careful in evaluating and selecting each amplifier to directly and dedicatedly drive each of the SP drivers, in my case woofers (WO), Be-squawkers (Be-SQ), Be-tweeters (Be-TW) and horn super tweeters (ST). These BE-SQ, Be-TW and ST are highly efficient in response to amp's power input. (My large and heavy L&R sub-woofers, Yamaha YST-SW1000 to go down to 16 Hz, has its dedicated amplifier in it.)
As I shared here, even Greg Timbers uses "reasonable and low budget" Pioneer Elite A-20 2-Ch integrated amplifier for UHF compression drivers (super tweeters) in his extraordinary expensive multichannel stereo system with JBL Everest DD67000 which he himself designed and developed.
Last edited:
Hi mitch! Thanks for explaining this
But if the DAC levels are already at max (example) in mixer, does it matter in practise whether you let AL compensate for this 6dB difference between drivers, or you manually adjust it in DACs ASIO/ALSA mixer panel? If it is already maxed in mixer, the only thing you can manually do is reduce levels in mixer?
But what has been achieved in the end in practical terms?
Assuming the mixer levels are not the same as the main volume knob (which is ESS DAC chip volume control)
dartinbout
Active Member
- Joined
- Jan 15, 2020
- Messages
- 275
- Likes
- 278
Good afternoon,
I heard back from Topping in regards to volume control of the individual channels.
The say,
"Dear Dartin
Thank you for contacting TOPPING support. We are happy to assist.
The volume of each channel can not be adjusted on our driver, you can adjust the settings on your computer, I have attached the settings tutorial to you. Hope this will help you.
Best Regards
--------------------
TOPPING SERVICE"
I brought a cheese, fruit and crackers for all to enjoy, during the coming "whine" tasting.
I heard back from Topping in regards to volume control of the individual channels.
The say,
"Dear Dartin
Thank you for contacting TOPPING support. We are happy to assist.
The volume of each channel can not be adjusted on our driver, you can adjust the settings on your computer, I have attached the settings tutorial to you. Hope this will help you.
Best Regards
--------------------
TOPPING SERVICE"
I brought a cheese, fruit and crackers for all to enjoy, during the coming "whine" tasting.
Can you please post the Setting Tutorial??Good afternoon,
I heard back from Topping in regards to volume control of the individual channels.
The say,
"Dear Dartin
Thank you for contacting TOPPING support. We are happy to assist.
The volume of each channel can not be adjusted on our driver, you can adjust the settings on your computer, I have attached the settings tutorial to you. Hope this will help you.
Best Regards
--------------------
TOPPING SERVICE"
I brought a cheese, fruit and crackers for all to enjoy, during the coming "whine" tasting.
dartinbout
Active Member
- Joined
- Jan 15, 2020
- Messages
- 275
- Likes
- 278
Sure.
Sokel
Master Contributor
- Joined
- Sep 8, 2021
- Messages
- 6,228
- Likes
- 6,355
They can't be serious...
dartinbout
Active Member
- Joined
- Jan 15, 2020
- Messages
- 275
- Likes
- 278
Our first guest! Will you be having a glass of red or white whine?They can't be serious...
Sokel
Master Contributor
- Joined
- Sep 8, 2021
- Messages
- 6,228
- Likes
- 6,355
Glass?Pass the bottle
What it achieves is being able to turn it up to 0dB on your volume and having it sound as clear as -20dB. However it'll be 90dB on a meter instead of 100dB because of the headroom management.
Hearing music at 90+db without distortion of any kind is a decent tradeoff for amplification power.
I feel like I am taking crazy pills following this thread, maybe I am missing something?
Let's compare two cases, 1) channel attenuation in DSP and 2) channel attenuation using a downstream DAC as advocated by @jtwrace. For this example let's say we want a 10 dB difference between drivers.
My first argument is that both cases are identical as long as you are not boosting in DSP. Let's say I send a 0 dBFS signal to the DSP, one channel will be at -10 dBFS and one will be at 0 dBFS, it doesn't matter whether I do the 10 dB relative attenuation in DSP or in the downstream DSP. Someone please show me how there is a difference between these two cases.
However, if I have boosts in my DSP I now potentially have issues in case 1. Let's say I have a 6 dB boost applied to the driver that I want at 10 dB lower level. Now I can clip at the DSP output because I have the possibility of exceeding 0 dBFS. It doesn't matter that I am attenuating by 10 dB after the DSP if I send 0 dBFS to the DSP and boost by 6 dB I will clip at the DSP output. I understand that not accounting for such a boost (especially if you are you are using downstream volume control in the DAC or preamp) is bad practice but you are more vulnerable in case 1. In case 2 if I boost by 6 dB in DSP but also cut by 10 dB net DSP output is -4 dBFS and I do not have a clipping issue.
Someone please give me a concrete example of why you do NOT want to do relative channel attenuation in DSP but rather downstream in the DAC. I just cannot come up with one.
Michael
Last edited:
I thought I was agreeing with you. I do.
In my previous post I mentioned I was trialing Dirac which takes a hefty cut off the top of your gain with it's filters (because it'll arbitrarily boost outside of curtain on both channels) but because it is not digitally clipping I can make up the gain afterwards in the chain. But only up to the limits of amplification power. Any EQ with proper cutting of peaks only is going to take a certain amount of amplification power to make up the difference.
Without any DSP my system will easily exceed 105dB at 3 meters out but sound like shit.
Edit: I do some boosting but a pretty shallow Q to the midrange.
In my previous post I mentioned I was trialing Dirac which takes a hefty cut off the top of your gain with it's filters (because it'll arbitrarily boost outside of curtain on both channels) but because it is not digitally clipping I can make up the gain afterwards in the chain. But only up to the limits of amplification power. Any EQ with proper cutting of peaks only is going to take a certain amount of amplification power to make up the difference.
Without any DSP my system will easily exceed 105dB at 3 meters out but sound like shit.
Edit: I do some boosting but a pretty shallow Q to the midrange.
Last edited:
Similar threads
- Poll
