Amplifier Gain, Insertion Loss... What's Best

[whoareyou]

I would like to simplify my system. It uses a single 2 channel together with 3 channel amplifer.
The main speakers are on the 2 channel amp, and that amp has a much lower gain than the 3 channel amplifier.
To correct for the difference in gain, I insert a preamp in between my DAC and the 2 channel power amp

Without preamp in the chain, you can see how the impact of the difference between amps(attachment) . I didn't attach graph with preamp in chain, but
that frequency response shows all levels to be basically the same.

Audiolense can provide the digital correction but is this the way to go, or is this too much of a difference and should I keep the preamp in the chain?
PS - It sounds good to my ears, so maybe that's all that matters? I also have plenty of power for my setup so that's not a concern.

Thanks

 

[juicehifi]

As long as you have enough gain to play loud enough this gain difference shouldn’t matter. But some times other factors enters, such as noise and impedance matching etc. so it doesn’t hurt to try the preamp.

 

[ddude003]

If I have read your DAC and Amplifier specs correctly, I would suggest a line-stage preamplifier between your DAC and Amp...
DAC RCA unbalanced line out is 2.1 volts and your Amp is balanced in is 2.0 volts... If your DAC does not explicitly state that it has a preamp in its analogue out section, which I could not verify from the docs, then I would definitely add a line stage pre amp...

I have a Chord DAC which has 1, 2, and 3 volts out, which I set at 2 volts and run into a PrimaLuna line stage Preamp and on to a McIntosh MC152 Amplifier... This is the best combination I have found for my system... However, your milage may vary...

 

[whoareyou]

If I have read your DAC and Amplifier specs correctly, I would suggest a line-stage preamplifier between your DAC and Amp...
DAC RCA unbalanced line out is 2.1 volts and your Amp is balanced in is 2.0 volts... If your DAC does not explicitly state that it has a preamp in its analogue out section, which I could not verify from the docs, then I would definitely add a line stage pre amp...

I have a Chord DAC which has 1, 2, and 3 volts out, which I set at 2 volts and run into a PrimaLuna line stage Preamp and on to a McIntosh MC152 Amplifier... This is the best combination I have found for my system... However, your milage may vary...

For the FR and FL channels I use a Topping Pre90 to provide additional gain to the Benchmark amp. In the above measurement, that's what's been removed, resulting in the difference between the channels
shown in the graph. Although, the pre90 is transparent I would still like to simplify system, so if I could "replace" the Pre90 with digital correction I'd be very happy with that solution.

Until your post, I completely forgot I had entered equipment info on the site. It's a bit out of date, but basically what you are saying for the exasound is correct.
That DAC, although very good, can't provide enough power to drive the amps together with gain loss introduced by AL. Without digital correction, I'm sure the Exasound would have enough power to drive the amps directly. It needs a preamp on every channel and for that I use a Parasound Halo 7.1 preamp in combination with the Exasound.

 

[ddude003]

It seems like you have this well thought out and under control... Let your ears be your guide...

 

[Kal Rubinson]

Without digital correction, I'm sure the Exasound would have enough power to drive the amps directly.

You do not state which exaSound DAC you have but at least some of them permit per channel level adjustments.

 

[whoareyou]

You do not state which exaSound DAC you have but at least some of them permit per channel level adjustments.

It's the e38, which does allow for this, but I have issues taking measurements with this DAC. Because of that, I measure with a Motu MK5 and I create/use the same correction filters based on the Motu's measurements. If I could, I'd use Jriver as central location to match channel levels, but I can't figure a way to measure with a Jriver/AL audio chain.

Anyway, so far so good with the AL solution, but I need more listening time on it.

 

[ddude003]

If you are using REW you can use the Generator function to save a sweep tone with a Timing ref as a .wav file that you can play with Jriver thru your signal chain... REW will recognize the Timing ref and take the measurement...

This is how I do it with SweepTone.wav > Audirvana+HangLooseConvolver with FIRs > WiFi > Router > Switch > U1 mini > Chord DAC > Pre > Amp > Speaker(s)... Nothing but a usb mic connected to my laptop for the sweep measurement(s)...

 

[whoareyou]

If you are using REW you can use the Generator function to save a sweep tone with a Timing ref as a .wav file that you can play with Jriver thru your signal chain... REW will recognize the Timing ref and take the measurement...

This is how I do it with SweepTone.wav > Audirvana+HangLooseConvolver with FIRs > WiFi > Router > Switch > U1 mini > Chord DAC > Pre > Amp > Speaker(s)... Nothing but a usb mic connected to my laptop for the sweep measurement(s)...

I do not use REW, only Audiolense for my measurements / filters. Also, I moved away from my UMIK-1 because of inconsistent timing measurements i.e. leave microphone in same position and measure several times and results were not consistent. This is with AL and I have no results based off REW measurements.

 

[ddude003]

I guess I have ran out of ideas then... It would be interesting if REW and Audiolense could exchange files... Then you could play sweep.wav files in your player and thru your signal chain and measure in REW... Then load the captured data into Audiolense for further processing...

If only...

 

[jrobbins50]

I use no preamp in my system and adjusted for the insertion loss by using digital gain within HQPlayer. Works really well for me. JCR

 

[whoareyou]

I use no preamp in my system and adjusted for the insertion loss by using digital gain within HQPlayer. Works really well for me. JCR

Just curious. Do you meausre with both Audiolense and HQPlayer in your audio chain or doing something else?

 

[ddude003]

I am kind of surprised that the Audiolense Convolver wouldn't have some kind of gain management... The HangLoose Convolver does...

 

[sledwards]

Just curious. Do you meausre with both Audiolense and HQPlayer in your audio chain or doing something else?

I am using the s88 DAC in my system with HQPlayer. I remove the s88 and use the Motu Ultralite Mk5 for AL measurements. I could never get clean sweeps with s88 for playback and Motu for mic input. I can adjust the gain levels per channel on the Motu prior to measuring. After I create convolution filters in AL, I put the s88 back into the system and load the filters into the Matrix Pipeline in HQPlayer where I can adjust the gain to the levels I had dialed in on the Motu measurement. Hope this helps. BTW, HQPlayer allows the user to quickly change between filter configurations, and while it does not level match, as long as there are not significant insertion losses between the different filters, you can still make quick A/B assessments.

 

[whoareyou]

I am using the s88 DAC in my system with HQPlayer. I remove the s88 and use the Motu Ultralite Mk5 for AL measurements. I could never get clean sweeps with s88 for playback and Motu for mic input. I can adjust the gain levels per channel on the Motu prior to measuring. After I create convolution filters in AL, I put the s88 back into the system and load the filters into the Matrix Pipeline in HQPlayer where I can adjust the gain to the levels I had dialed in on the Motu measurement. Hope this helps. BTW, HQPlayer allows the user to quickly change between filter configurations, and while it does not level match, as long as there are not significant insertion losses between the different filters, you can still make quick A/B assessments.

I also use Mk5 for measurements, and for same reason. Once filters are created, I switch between a couple of DACs, but always use 0DB across devices because of uncertainty of how each DAC's db scales match up with one another. I wasn't sure if, for example, 6db attenuation on one was equal to 6db attenuation on other (I probably just should have tried it, and it makes sense this should work). I'll probably give this approach a try and see how I like it.

PS: AL seems to be doing a very nice job of digitally correcting the levels. So far, I see no issue with it for my setup.

Thanks

 

[jrobbins50]

And I will join the chorus. I have stopped using my Okto dac8pro for AL measurements and now use the Motu mk5. My experience is that the dial in of the attenuation adjustments in the Motu matched up with HQP. The difference is that the Motu attenuates and HQP adds gain. So, I just have everything in the same dB relationship to each other and it works very well. JCR

 

[whoareyou]

Ok, you guys have convinced me to give this a try some time down the road, but for now I'm going to stick with the AL digital correction as it sounds very good to me.

The difference is that the Motu attenuates and HQP adds gain. So, I just have everything in the same dB relationship to each other and it works very well. JCR

Adding gain doesn't risk clipping?

 

[Dali_20]

The important thing is that the final response of the drivers is aligned in magnitude, then how you get there makes no difference.
Audiolense should do this automatically, therefore the single gain of the various channels is irrelevant.
Clearly if you have big differences in gain there will be the channels with the greatest gain that will be more penalized as SNR by the digital attenuation of the filters.
But digital music is already digitally modulated in magnitude, so it doesn't change much...

Maybe for safety use digital volume attenuation of the DAC with higher output as a limiter, so you avoid accidents...

But it is not clear to me how the volume is currently adjusted. Digitally I assume right?

 

[jrobbins50]

To WhoAreYou’s question re clipping. I have not noticed any.

Mitch Barnett could weigh in here. He was the one who said the right way is to equalize the volume levels among drivers, rather than allowing the attenuation solely within AL. @Mitchco , perhaps you could comment? JCR

 

[whoareyou]

To WhoAreYou’s question re clipping. I have not noticed any.

Mitch Barnett could weigh in here. He was the one who said the right way is to equalize the volume levels among drivers, rather than allowing the attenuation solely within AL. @Mitchco , perhaps you could comment? JCR

Perhaps this is is more important when using AL to build your own crossovers.

 

[Mitchco]

@whoareyou it is hard to tell from your screen shot in the first post what the dB vertical division is? Also, probably more constructive to see the correction filters response... complete with vertical (1dB per div) and horizontal scales.

Yes, AL will make up the gain difference. But if there is too much of a gap between levels, you may need to add more "max corr" to close the gap. Adding more overall correction "can" take away from the music dynamics, aside from introducing more filter insertion loss. Since you have lots of power/headroom, the latter does not matter, but the former might...

It boils down to good engineering practices in gain staging. In my stereo triamp system, I match the levels of the three drivers as closely as possible to the intended target response before correction. Since I use compression drivers on the tweeters, they are way more sensitive then the mids and woofers. So rather than "pad down" the tweeter or turn down any analog or digital gain, I chose an amp with considerably less voltage gain.

With the driver levels matched, I can get away with minimal overall correction, which still provides the smooth response, but retains the dynamics I like. There is a sweet spot to how much max corr one should apply, your ears will tell you...

 

[whoareyou]

@Mitchco - Edit: The vertical divisions are 2DB (originally had typo of 6DB) It's pretty much spot on with the difference in published gain between my two amps. The total difference in gain is around 6db and that's pretty much the difference in gain between my amps.

I've been very happy with the results having AL do the correction, but since I can't instantly compare corrections, I have no way of knowing if one sounds better than the other. What I do know is the gain corrected filters sound very good, and I don't feel like I'm missing anything (except a few preamp related wires that were somehow always getting tangled up .

As far as the max gain, I was concerned with how this may be impacted, but the max stayed within same range, and is completely driven by my nasty room null. All simulations also look very similar (but as you suggest, I really should go back and look at some of the graphs with finer resolution).

I was just about to purchase a new lower gain amp, but decided to give this a try. I'm happy enough with result such that I've rethought the purchase.

 

[whoareyou]

Regarding the dynamics, does any analysis capture how "much" has been lost? Can you infer anything about the gain from the quality and amplitude of the step response, i.e. cleaner and higher amplitudes translate to better dynamics? Or have I gone off the deep end into pseudo-science ?

 

[juicehifi]

The correction limiter operates on a per driver basis, and the limit only operates on lokal dips, not global gain. So I expect the excact same result in most situations with or without digital gain adjustment outside of Audiolense. The logics used for separate the local from global gain is a bit fuzzy, so there may be instances where a difference occurs, but normally not.

Analog gain-matching can have some merits, though, to optimise S/N ratio at a reasonable loud level.

 

[whoareyou]

So, I remeasured after adjusting digitally via DAC. The correction filters may look sightly better, but can I hear a difference? Not sure / probably not.
For my room, AL vs DAC approach seems to make little difference.