TOPIC:

After running a full calibration, my overall volume is low, despite turning the drc-88a volume all the way up to 0db.

I then noticed some very strange behavior of my Umik1.

If I reboot the PC, and run REW, the mic correctly measures the SPL in my quiet room of around 50db.

Then, run a dirac calibration, and immediately after, run REW again...this time the REW SPL measures the same quiet room around 30db higher.

This explains why my dirac calibration is low in volume (it thinks everything is 30db higher), but the question is why?

I'm using a CSL calibration file (90 degree 1/3 octave) with the mic sensitivity 1st line copied into it.

Any thoughts?

Hi, DLCT doesn't read the sensitivity figure, so whatever that is about must be some sort of REW issue. At any rate, if you want to be checking SPL readings it's better to provide a signal.

With regard to Dirac, just checking that you moved the slider on the Export tab (to full)? Also, on the output and levels screen, did you move any the output volume sliders?

Also, just verify that you have the latest firmware.

john.reekie wrote:

Hi, DLCT doesn't read the sensitivity figure, so whatever that is about must be some sort of REW issue. At any rate, if you want to be checking SPL readings it's better to provide a signal.

It seems that was a red herring. But, others have reported the same issue. REW reads SPL levels way higher than they are after running through a DLCT calibration. Rebooting the computer fixes it. Separate issue.

john.reekie wrote:

With regard to Dirac, just checking that you moved the slider on the Export tab (to full)? Also, on the output and levels screen, did you move any the output volume sliders?

Also, just verify that you have the latest firmware.

I've made sure to move the slider before exporting, and also realized that if I don't do that, I can simply turn the volume all the way up on the 88a afterwards, no difference.

I *did* have to move sliders on the individual channels. I've come to learn that this is the entire problem. If I do a simple 3 channel calibration, things are spot-on when I am done (I read 75db with an SPL meter running pink noise). If I try to include my surrounds in the calibration, I have to tweak the individual channel levels, and when I do that, the calibration ends up being down 8 or so db.

It seemed to me that since all channels appear to be maxed from the outset, the only way to deal with channels that need raising is to reduce all of the other channels instead. Is that the way it should be? Or, should I actually be able to raise the individual channel level on my two surrounds?

A real feature request here: those individual channel sliders should have a db label like the output level at the top of the window does now. Then we would know what they are actually set to rather than trying to figure it out from an otherwise non-obvious graphical representation. Furthermore, those sliders appear to be actually *broken* in that they do not represent the actual values that are being used if you finish a session of calibration, and then re-launch the app...But that's yet another separate issue that we can cover at another time. [It's actually quite obvious if you move a level to an extreme amount. The next time you launch the app, it will appear to be visually back where it started, but you can clearly see from the level when testing that channel that it is using the previous skewed amount.]

I'm on firmware 2.17.

Confirmed. If I launch the calibration tool, the sliders are all already at max, so I can't bump my surround channels, I can only reduce everything else.

Is it supposed to behave that way?

It sounds as though you are saying that your surround channels are substantially lower in sensitivity than the rest. If so, and your output gain jumpers are "down" then you can set them "up." There's a lot of info in the manual about gain structure, perhaps worth a look.

john.reekie wrote:

It sounds as though you are saying that your surround channels are substantially lower in sensitivity than the rest. If so, and your output gain jumpers are "down" then you can set them "up." There's a lot of info in the manual about gain structure, perhaps worth a look.

Yep, my surrounds are much quieter since they are going through a much less powerful amp.

My jumpers are already "up" (2v in, 2v out).

kgoroway wrote:

john.reekie wrote:

It sounds as though you are saying that your surround channels are substantially lower in sensitivity than the rest. If so, and your output gain jumpers are "down" then you can set them "up." There's a lot of info in the manual about gain structure, perhaps worth a look.

Yep, my surrounds are much quieter since they are going through a much less powerful amp.

My jumpers are already "up" (2v in, 2v out).

If you look on page 41 of the user guide, you will see that you can mix the output DIP switch settings. So if your surrounds are at a lower output level, perhaps you should try setting the LCR DIP switch settings to .9V, and the other channels to 2V. That will raise the output to the surround channels by 6dB relative to the LCR. Have you experimented with this approach?

ICBW, but I think this is what John is suggesting.

@Kgoroway,

Let's try to see if we can bring some ideas/feedback:

Level issue: That statement here is key: "Yep, my surrounds are much quieter since they are going through a much less powerful amp."
Dirac leveling algorithm is quite simple. It's sweeping each of the speakers and then looking at their respective level. If say your surround are a lot lower than the front, then Dirac algorithm rightly assumes that these speakers are just less powerful (when indeed it's the amplifier). In other words, for basic physical limitations, you shouldn't try to overdrive them by a lot. (else you'd run into distortion). They are therefore used as a basis for leveling and all other channels are aligned to that signal.

So the key issue is here more down to the mismatched amplifier (throwing off DIrac algorithm). If you don't have any sensitivity/volume control for this amplifier, then you'd want to use the DIP switch as correctly suggested by Jerry. That's the work around because the fader inside the software is a digital volume control (i. 100% = max volume). You indeed don't have 150% possibility, that would be asking for trouble to do volume boost in the digital domain.

REW question: We've never seen this but it sound like a sound card/software issue as to how REW/DIRAC is taking control of the USB audio class level control.
Did you try to kill Dirac Calibration tool when you run REW? You'd also want to make sure that you have ticked the box under preferences for control input level? Running latest REW we're guessing? It should be automatic.
Finally, please make a copy and paste of how you've manually pasted that sensitivity factor to your first line of the CSL cal file. Just want to make sure it's correct.

DevTeam

Kevin has been discussing this issue over on the AVS Forum as well. Thank you for the description regarding how the leveling process works. I think we are all looking forward to Kevin's report, should he decide to re-address the gain structure by altering the dip switch settings.

I think this is a good place to comment on what many of us think is a hardware design issue with the 88A. As you can imagine, connecting sixteen RCA cables to the back of the unit, especially when the 88A is mounted in an equipment rack, is not a trivial task. So, when the entire unit must be removed in order to open the case to access the dip switches, this makes experimenting with gain settings a difficult process. If the dip switch were mounted so that it could be accessed externally, i.e. without having to remove the 88A from the rack or disconnect cables, that would be a very welcome hardware change. Perhaps this suggestion could be passed on to the engineering team?

And finally, regarding the peculiar issue Kevin reported with the microphone gain in REW after completing a Dirac calibration, be advised that not everyone has seen this issue. I have been using the 88A for over six months, and every time I complete a Dirac calibration, I immediately launch REW to evaluate the calibration results. I have never experienced the issue Kevin has reported. This does not mean the issue doesn't exist, rather I mention this to add the perspective that it is not a pervasive issue. I will offer to assist Kevin to isolate root cause.

devteam wrote:

@Kgoroway,
Level issue: That statement here is key: "Yep, my surrounds are much quieter since they are going through a much less powerful amp."
Dirac leveling algorithm is quite simple. It's sweeping each of the speakers and then looking at their respective level. If say your surround are a lot lower than the front, then Dirac algorithm rightly assumes that these speakers are just less powerful (when indeed it's the amplifier). In other words, for basic physical limitations, you shouldn't try to overdrive them by a lot. (else you'd run into distortion). They are therefore used as a basis for leveling and all other channels are aligned to that signal.

So the key issue is here more down to the mismatched amplifier (throwing off DIrac algorithm). If you don't have any sensitivity/volume control for this amplifier, then you'd want to use the DIP switch as correctly suggested by Jerry. That's the work around because the fader inside the software is a digital volume control (i. 100% = max volume). You indeed don't have 150% possibility, that would be asking for trouble to do volume boost in the digital domain.

So, is it then expected behavior that the resulting filters will yield volume other than 75db at reference? While it seems clear that trying to boost the signal above 100% for the surrounds would be quite problematic, I don't see any problem with reducing the gain on the other channels. Is there some reason the processor doesn't do that?

devteam wrote:

REW question: We've never seen this but it sound like a sound card/software issue as to how REW/DIRAC is taking control of the USB audio class level control.
Did you try to kill Dirac Calibration tool when you run REW? You'd also want to make sure that you have ticked the box under preferences for control input level? Running latest REW we're guessing? It should be automatic.
Finally, please make a copy and paste of how you've manually pasted that sensitivity factor to your first line of the CSL cal file. Just want to make sure it's correct.

I haven't tried to "kill" the calibration tool, I only exited it normally. (REW can't see the mic while the calibration tool is still running) Are you suggesting that I try killing it from the task manager?

I'm not familiar with the "control input level" preference...I'll go looking for that.
Yes, running latest REW.

I'd absolutely speculate that this was a one-off problem, and nothing to worry about...except at least one other person on AVS echoed my report, claiming to witness the exact same issue.

First line of my mic calibration file: (SERNO changed)
"Sens Factor =-5.033dB, SERNO: 7002xxx"

Thanks for the help!

EDIT: Okay, some more information. I went looking for the "control input level" and found that I only have that option if I revert to the java drivers. Using the ASIO drivers, there is no "control input level" checkbox. That being said...I found that if I make REW do the "bad" thing, (run a DIRAC calibration), I can "fix" REW by switching back to the java drivers (SPL level is fine again), and then switch back once more to the ASIO drivers (SPL level is still fine). So, we can at least call that a work-around, and it likely points the problem at the ASIO drivers (ASIO4All in this case).

AustinJerry wrote:

Kevin has been discussing this issue over on the AVS Forum as well. Thank you for the description regarding how the leveling process works. I think we are all looking forward to Kevin's report, should he decide to re-address the gain structure by altering the dip switch settings.

I think this is a good place to comment on what many of us think is a hardware design issue with the 88A. As you can imagine, connecting sixteen RCA cables to the back of the unit, especially when the 88A is mounted in an equipment rack, is not a trivial task. So, when the entire unit must be removed in order to open the case to access the dip switches, this makes experimenting with gain settings a difficult process. If the dip switch were mounted so that it could be accessed externally, i.e. without having to remove the 88A from the rack or disconnect cables, that would be a very welcome hardware change. Perhaps this suggestion could be passed on to the engineering team?

And finally, regarding the peculiar issue Kevin reported with the microphone gain in REW after completing a Dirac calibration, be advised that not everyone has seen this issue. I have been using the 88A for over six months, and every time I complete a Dirac calibration, I immediately launch REW to evaluate the calibration results. I have never experienced the issue Kevin has reported. This does not mean the issue doesn't exist, rather I mention this to add the perspective that it is not a pervasive issue. I will offer to assist Kevin to isolate root cause.

Hi Jerry,

Thanks for the feedback and great hearing that the AVS community is once again very helpful! You guys are a great team and we can't be thankful enough of all the great help you're providing to the community! Thanks again.

- DIP switch: That's indeed a very valid request. I could certainly see how we could put this into place for a future hardware revision. An external DIP switch might require a bit of reorganization as we're quite tight on the rear panel, but worth looking into it... Thanks for the feedback.

- REW + ASIO + DLCT: Interesting issue that we'd have to see how we can reproduce.. Not sure of the root cause to be honest.. Sounds like the way the OS is taking over the UMIK-1 over Class 1 vs ASIO..

DevTeam

@ Kevin,

- Levelling: Not 100% sure to be honest if we're understanding your comment about leveling. I believe that the algorithm is actually already levelling (i.e. bringing the whole system lower). It's doing so because it's looking at the surround as the basis for the highest the system can achieve I guess. Are we missing maybe something? (hard to understand a system thousand of miles away.. )

- REW: Very interesting finding indeed that it seems like ASIO is locking the use of the device.. That little workaround sounds like a neat little trick for the time being as we don't know 100% what REW does.. Possible that you'd have to reboot REW to get UMIK-1 discovered again. i.e.
a) Close DLCT
b) Open REW (should discover the UMIK-1 at that point) with the prompt.

DevTeam

devteam wrote:

@ Kevin,
- Levelling: Not 100% sure to be honest if we're understanding your comment about leveling. I believe that the algorithm is actually already levelling (i.e. bringing the whole system lower). It's doing so because it's looking at the surround as the basis for the highest the system can achieve I guess. Are we missing maybe something? (hard to understand a system thousand of miles away.. )

I agree that the 88a is doing a good job at leveling the system. But here's the question as simply as I can put it.

If the surrounds are the quietest speakers, but I lower the other channels in the calibration screen such that they are all at -12 before running the calibration, is it expected that the resulting calibration will be lower than 75db at reference? And, if so, why? [There's certainly no problem with the surrounds reaching 75db at reference, in fact, if I use an SPL meter while running the calibration, I can easily achieve MUCH louder SPL than 75db from the surrounds by raising the output of the 88a]

Kevin,

To be honest, not 100% sure of what could be the reason here. Maybe something down to how the algorithm by Dirac figures out the leveling?
While we're able to figure out quite few things based on our experience, there is quite a fair amount of information that we don't have access to on how Dirac measures/figures out filters and levels. It's somehow part of their secret sauce.

Did you try the trick we mentioned earlier (i..e changing the DIP switch). Did that do the trick?

DevTeam

devteam wrote:

Kevin,

To be honest, not 100% sure of what could be the reason here. Maybe something down to how the algorithm by Dirac figures out the leveling?
While we're able to figure out quite few things based on our experience, there is quite a fair amount of information that we don't have access to on how Dirac measures/figures out filters and levels. It's somehow part of their secret sauce.

Did you try the trick we mentioned earlier (i..e changing the DIP switch). Did that do the trick?

DevTeam

Is it possible to reach out to Dirac to see if this is the expected behavior?

I didn't try to DIP switch trick for a few reasons.
#1) It's an absolute pain. (see above suggestion on making the DIP switches easier to get to)
#2) I'm absolutely certain that it would work, since I proved that the issue doesn't happen if I don't include the surrounds in the calibration
#3) It would, in effect, by crippling the best amp in my setup that I use for my most important speakers (front L/C/R)
#4) I can completely solve the problem by boosting the output signal by 7.5db for all of my channels in the AVR to get back to 75db at reference volume.

-Kevin