|
Post by Marvin on Mar 17, 2011 22:41:21 GMT -5
Good suggestion, Jason. I have been quite busy lately, so tweaking has not been a priority. I do hope to try some of these things.........someday.
Marvin
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 18, 2011 4:41:54 GMT -5
My money's on digital clipping. Yes, I get the same effect if I boost a frequency too much with the miniDSP. I turn down the input gain of the miniDSP and it goes away. It sounds a lot like a speaker distorting from damage. Its easy for me to work around.
|
|
|
Post by autocrat on Mar 18, 2011 19:46:11 GMT -5
Back on the first page (nearly 2 months ago!) I suggested it might be digital clipping. That might be right or not, in any event it seems to have become the default speculative cause of the observed distortions.
I'd really appreciate someone from Emotiva stepping in and commenting on what people are observing - if it is digital clipping, then it can do some real damage to speakers, so I think we should all know what were dealing with.
Thanks.
|
|
|
Post by Marvin on Mar 18, 2011 23:49:02 GMT -5
Back on the first page (nearly 2 months ago!) I suggested it might be digital clipping. That might be right or not, in any event it seems to have become the default speculative cause of the observed distortions. I'd really appreciate someone from Emotiva stepping in and commenting on what people are observing - if it is digital clipping, then it can do some real damage to speakers, so I think we should all know what were dealing with. Thanks. +1 If that is a limitation of digital audio and is not unique to the UMC (post by theresa re. mini DSP distortion), then suggestions from EMO on why it happens and how to avoid it would be most welcome.
|
|
|
Post by marrypoppins on Mar 19, 2011 7:10:01 GMT -5
I'd really appreciate someone from Emotiva stepping in and commenting on what people are observing - if it is digital clipping, then it can do some real damage to speakers, so I think we should all know what were dealing with. Thanks. I am not from Emotiva, but I've made my own investigations using FW 7.04: First I've made Setup->Load Default then Input Setup -> Cable -> EQ Mode : Manual EQ 1 then Speaker Setup -> Speaker Size and Crossover-> L/R Front: 250Hz Speaker Setup -> Speaker Size and Crossover-> Front Xover Slope: 24dB then Emoq/EQ -> Manual EQ 1-> Front Right: all freq. gains to 10 except for 63Hz: 9 Then I took my notebook and connect it using HDMI to HDMI2. Set the windows sound level to maximum. Then I start REW (Room EQ Wizard) and generate there a 63Hz tone at level -10db. So far so good: I hear 63Hz tone from my SW and front speakers keep silent. Then I change the 63Hz eq gain on UMC1 from 9 to 10 and... my front right speaker starts to produces tone >~ 250Hz even louder than my SW. What that mean? I think that it is definitely digital distortion(but not a digital clipping), because additionally to origianl 63Hzh frequency I hear probably 4*63=252Hz tone (it comes from loudspeaker with Crossover 250Hz.) You can do the same test without SW and Crossovers using only one speaker and you will definitely hear how the source signal change its frequency at this EQ freq gain 10. If I set the db level in REW to -11db then no signal come from front speaker(no distortion) even when EQ 63Hz gain is on 10. If I set the db level to -3 in REW, then I have no distortion on EQ 63hz gain:2, but distortion starts in this case from EQ 63hz at gain 3. So for me it means a bug in EmoQ. I even can explain what the kind of bug do they have in their DSP code: it looks like EmoQ has overflow problem. Instead of digital clipping EmoQ does distortion if signal too near the maximum. So (as I already written) I do not recommend to use boosting in EmoQ only gains <= 0. Otherwise explosions in DD and DTS could produce distorted sound in your main speakers (Actually I've heard that distortion from explosions while watching "Star Wars")! Could somebody else prove my experiment?
|
|
|
Post by Nemesis.ie on Mar 20, 2011 8:06:33 GMT -5
MP: (Welcome!) I set my C-200 volume control to -6 or -5 which matches PCM sound (for 2 channel) to the overall level of other formats which I bitstream. That way I just leave it like that and use the UMC-1 volume control all the time, it seems to work fine.
The QSC RMX amps look interesting, are the fans quiet?
jason1976: 7.04 does seem to have a cleaner sound/lower noise floor even with EQ/Emo-Q used.
|
|
|
Post by marrypoppins on Mar 20, 2011 9:41:53 GMT -5
If you set C-200 to -5 -6 than I suppose you are on safe side for (Multi)PCM if you do not have EmoQ boosting gains more than +5 +6. And you are on safe side for dialog levels forum.blu-ray.com/audio-theory-discussion/121087-understanding-dialog-normalization.html of DD and DTS. But you are not safe for explosion levels of DD and DTS. Original fan of QSC is very very loud (40-50db), but I replaced it with 12db fan + resistor + thermo-switch so now it is < 6db when amplifier temperature <60C and 12db if amplifier goes hot(but that was never the case in my HT usage: -36..-15 DBFS). MP.
|
|
|
Post by Nemesis.ie on Mar 20, 2011 9:56:48 GMT -5
If you set C-200 to -5 -6 than I suppose you are on safe side for (Multi)PCM if you do not have EmoQ boosting gains more than +5 +6. And you are on safe side for dialog levels forum.blu-ray.com/audio-theory-discussion/121087-understanding-dialog-normalization.html of DD and DTS. But you are not safe for explosion levels of DD and DTS. Original fan of QSC is very very loud (40-50db), but I replaced it with 12db fan + resistor + thermo-switch so now it is < 6db when amplifier temperature <60C and 12db if amplifier goes hot(but that was never the case in my HT usage: -36..-15 DBFS). MP. Hi MP - I was really just mentioning those settings for overall level balance whether or not any EQ (on the Emo-side) is being used. As an aside and more pertinent to the discussion - I wonder if with Emo-EQ only being used for frequencies above the XO does this problem appear - i.e. is the distortion purely coming from some kind of "digital harmonic" of the signals crossed over to the sub? In my setup, I do not use any UMC-1 EQ for frequencies below the XO, i.e. I set the lower bands for the speakers to 0 and do all my sub EQ via the MiniDSP and have not really noticed any distortion. BTW, where are you based? I'm guessing somewhere in Europe? Great to see another person posting detailed observations. I recently changed the layout of my room (and have been sick/on holiday, busy with work too) and it's not finished yet so I'm not geared up to run any REW tests for the moment for verification. I do want to though as my MiniDSP sub EQ definitely needs to be redone.
|
|
|
Post by marrypoppins on Mar 20, 2011 11:50:15 GMT -5
Hi Nemesis:
The problem I've described applies to all frequencies. I used XO only to split original 63hz and distorted harmonics ~300Hz and show that they are exists by outputting them on different devices SW and loudspeaker in my case. You can turn off XO (set loudspeaker to Full), then you will hear how the tone has been changed in the same loudspeaker, but you can't tell in this case that it was UMC1 problem: it could be amplifier or loudspeaker.
Input signal is distorted in other frequencies as well. For example if you use 1kHz source signal tone then you probably got 5kHz distorted tone, and it will be mixed with source tone 1Khz and it will be played on the same speaker and you can detect it only with you ears. I've took coefficient 5 for distortion from the pictures I've attached to my original post. There we can see that from 1 signal wave after EmoQ (if my suggestion is right) we get 5 waves in the same period of time.
If you take 4kHz signal you get ~20kHz distorted signal, probably you can hear it as well. But if you take 16kHz input signal then you get 80kHz!!! distorted signal. Nobody (I mean people) can hear 80kHz, but that probably could damage your speakers if your amplifier is capable to amplify that frequency and loudspeaker is not capable to filter or reproduce it (I suppose that UMC DAC is capable to produce that frequency because it supports 24bit/192Khz). So be careful and take care of your loudspeakers.
And you are right the latest few years I am living in Europe.
MP.
|
|
|
Post by Nemesis.ie on Mar 20, 2011 12:22:22 GMT -5
Thanks MP - I think this warrants some serious investigation and also comment/testing from Emotiva. P.S. I had a feeling you were in Germany too for some reason.
|
|
|
Post by jason1976 on Mar 23, 2011 16:10:18 GMT -5
Hi Nemesis,
Actually after playing with this some more, the distortion is back, albeit in a lesser fashion. After I did the reset, it was much less noticeable. I'm sad to say that I can still hear it with EMOq engaged and minimal boosting.
I did notice the noise floor get better with the 04 update.
I used to have my HTPC connected via Optical and never had an issue. When upped my video card and switched to HDMI was when I experienced the distortion. I'm wondering if this isn't some sort of level mismatch with the HDMI inputs.
There are so many variables with this issue which makes it hard to pinpoint. I really wish EMO would at least acknowledge the issue. Lonnie are you out there??
|
|
|
Post by Nemesis.ie on Mar 24, 2011 5:08:53 GMT -5
If you are seeing (hearing) it from your graphics card, do you have any other HDMI device to compare it with? Also, what drivers are you using and what player?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 24, 2011 7:34:42 GMT -5
Since it's a problem exclusive to you, as far as is known, its hard to pinpoint such a problem. I had the problem that if the miniDSP for my center channel was touched it would cause the tweeter amp channel to shut down. I finally switched channels on the miniDSP and replaced the cable going to the amp and the problem is gone. I had no idea that it was either the miniDSP or the cable and still don't know which it is and since I am the only one with this problem, was in somewhat the same place as you. I would send my UMC-1 back to Emotiva and have it serviced or replaced. It could be your power source
|
|
|
Post by jason1976 on Mar 24, 2011 8:33:16 GMT -5
Well I don't think it is just a problem with me, there were several other people that had a similar problem. The Original poster had the problem on Analog inputs and it was separate from HDMI. It only happens in the front Mains that I have found, just like Marry Poppins was testing.
I also have a Panasonic Bluray player that I can test as well. I'll just burn a CD of the offending music and see if I can replicate it.
My distortion starts when I'm boosting by only 3db, so it isn't like I'm trying to over correct. Most of my other tweaks are cuts to the highs. All EQs have some form of a master gain control and it seems to me that this would be an easy problem to solve with the settings that are provided in the service menus.
|
|
|
Post by Marvin on Mar 24, 2011 9:13:24 GMT -5
OP here. Jason is correct, I have distortion on analog and digital inputs--really any type of input--if you boost too much. Even with modest boost of a few bands, some songs distort in certain chords or passages.
Still haven't updated to 04, but intend to someday, but from others reports, don't expect that to cure the distortion problem.
Marv
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 24, 2011 9:19:40 GMT -5
You could take a look at what the distortion looks like with REW and a measurement mic. Even 3db of boost can overload the miniDSPs, in my experience, so I lower the input level to the miniDSP by the amount of the biggest boost.
|
|
|
Post by Nemesis.ie on Mar 24, 2011 12:34:56 GMT -5
Interesting, using REW's generated biquads, I've not seen/heard distortion using boosts of even 12db+. Are you using biquads or moving sliders/plot points? It could be very room dependent too of course. Maybe we need an REW/MiniDSP thread (and lots of time and energy for doing tests too!).
|
|
|
Post by jason1976 on Mar 24, 2011 22:27:26 GMT -5
I wouldn't mind helping with the tests, but I have a feeling that this isn't going to get addressed. I know it doesn't effect everyone, but this really shouldn't require new code to get written. Something is off in the gain and input level structure.
|
|
|
Post by jason1976 on Mar 24, 2011 23:42:42 GMT -5
Hi Nemesis: The problem I've described applies to all frequencies. I used XO only to split original 63hz and distorted harmonics ~300Hz and show that they are exists by outputting them on different devices SW and loudspeaker in my case. You can turn off XO (set loudspeaker to Full), then you will hear how the tone has been changed in the same loudspeaker, but you can't tell in this case that it was UMC1 problem: it could be amplifier or loudspeaker. Input signal is distorted in other frequencies as well. For example if you use 1kHz source signal tone then you probably got 5kHz distorted tone, and it will be mixed with source tone 1Khz and it will be played on the same speaker and you can detect it only with you ears. I've took coefficient 5 for distortion from the pictures I've attached to my original post. There we can see that from 1 signal wave after EmoQ (if my suggestion is right) we get 5 waves in the same period of time. If you take 4kHz signal you get ~20kHz distorted signal, probably you can hear it as well. But if you take 16kHz input signal then you get 80kHz!!! distorted signal. Nobody (I mean people) can hear 80kHz, but that probably could damage your speakers if your amplifier is capable to amplify that frequency and loudspeaker is not capable to filter or reproduce it (I suppose that UMC DAC is capable to produce that frequency because it supports 24bit/192Khz). So be careful and take care of your loudspeakers. And you are right the latest few years I am living in Europe. MP. Got an idea for you. Do you have the option of switching your video/audio card to Multi Channel PCM? On the ATI Realtek adaptor I have I can force Multichannel PCM to the UMC. When I did this I noticed the input levels dropped considerably vs PCM and it was much harder to overload EMOq. The PCM decoded signal was 5-10 dbs hotter than Multi PCM. I'm just wondering if the something in the PCM level decoding is off. It also might explain the OPs issue because analog was coming in and converted to PCM.
|
|
|
Post by Nemesis.ie on Mar 25, 2011 7:08:08 GMT -5
Likewise I set my Popcorn hour's volume to -6 (from 0) to equalise PCM level with other formats so maybe that helps with this issue.
|
|