|
Post by ttocs on Jun 28, 2020 11:33:30 GMT -5
Regarding the lower volume, it now means absolutely nothing to me. When I look back at when I noticed a lower volume with Dirac 1 and the XMC-1 and didn't like it, but nothing I could do about it, I wanted to understand why this was. Not too long ago it was pointed out to me that the reason for this is due to the target curve being a REDUCTION of the peaks. So now I get it. Can't have same volume out when the processed volume has been reduced. The key word here is "processed".
This doesn't specifically answer why Neural:X processing is lower in volume, but I've got to believe that there's something about the way it operates that makes it do this. So does the Marantz - while playing Neural:X at the same volume as DD - play at the same volume as Neural:X or DD on the RMC? Only those who take voltage measurements will know for sure.
edit: I just got done measuring ONLY the Right Front Height:: Sm Front Height and found a nice sharp sound in The Dark Knight when the knife is in the guys mouth in the pool table scene. There is no difference in volume between Neural:X and DSU when player is set to PCM. This suggests to me that Neural:X is doing what I've read it does, emphasize the immersive speakers at the expense of the beds. So instead of RAISING the volume of the immersive speakers, it keeps those the same and LOWERS the volume of the other channels. Further, there was no discernable difference in dynamics.
Prior to the above test I did a Factory Reset so all speaker Levels and Distances are set to zero.
|
|
|
Post by Mikomill on Jun 28, 2020 12:36:34 GMT -5
Of course I didn’t. If following this thread it was written in context. You had to have read my previous posts. But maybe I should write each comment as a new comment to not be misinterpreted.. The sound is better (and higher in volume) with heights chosen in setup rather than tops. This with surround mode used and output only to 5.1 speakers. Hope this is clear enough? Why tops drag down surround mode in a straight output is really strange though. If I hadn’t switched between these speaker settings I would never have noticed this. I think a lot more people need to do this sort of testing to realize they too have issues. ...better yet, you need to accept the fact that other users DON'T have the issue that you think you have regardless of how hard you're trying to convince them that they do!
|
|
|
Post by hsamwel on Jun 28, 2020 13:15:29 GMT -5
Got it...so it isn't that you know the bit rate has changed, it just sounds like it...so, kind of like if in stereo the difference between a very low res recording and a top notch one. And also, per one of the other posters, some are expecting sound out of certain speakers with Neural X and it's not there at all (as another issue). Right? So, if I get it right the distinct issues people are having are actually: 1) ~6 dB level drop 2) Audio quality isn't what they expect and it either: a) sounds less dynamic b) sounds like a lower quality audio recording (muddy, etc.) 3) Isn't playing all the speakers people think should be playing. Correct? I'm looking into how to do a survey that would isolate which issues people are having and how they might interact. Frankly, it looks like Survey Monkey would be the best tool to gather that data. The polling system here is pretty limited, but I can create a survey in Survey Monkey and include a link here. And, to be clear - my interest is to ensure Emotiva has the best possible problem definition. As I've said many times in my career, a problem well-defined is the first step to a solution. Mark Its not just that speakers are being left out, its sound that is completely gone which I believe is a redirect of sound to speakers that don't exist in the setup. As I mentioned earlier it is my belief that the 6dB or greater drop is lost sound energy that has been misdirected. This would explain why scene to scene its different. If the missing sound is some thunder from a storm it will sound like a loss of dynamic range. If its an explosion it may drop the midrange leaving the bass and making things sound muddy. In other scenes it may just be incidental atmospheric sound in which case it will hardlybe noticedand just sound quieter I have 6dB drop from Surround mode as well with speakers set to tops, not with heights, from a DTS source. Did test again today from a different DTS-HD MA 7.1 track. Still the same. Well, I have a test I’d like everyone with atmos speakers could do. Play a scene from a DTS 5.1/7.1 source (legacy or MA doesn’t matter). Bitstream! Have your setting set to 7.x.4 with tops. Play the same scene with first Surround then again with Neural:X. Volume difference? Then change top speakers to heights. Play the same scene with Surround and then with Neural:X. Volume difference? I did read a source that stated Neural:X only upmixes up to 48khz. That would explain why DTS-HD MA 5.1 96khz 24bit tracks don’t get upmixed to heights and only to surround back I guess?! Play a PCM 5.1 source. Do you get any sound from top spealers with Neural:X?
|
|
|
Post by hsamwel on Jun 28, 2020 13:53:14 GMT -5
I think a lot more people need to do this sort of testing to realize they too have issues. Done it already and guess what....... realise I don’t have the issues. I bowed out of this discussion before and now I feel I should again because I just don’t get it. Good luck with your quest. Ok. So you have the same volume with Surround mode with either tops or heights selected in speaker settings? Note I’m not talking about Neural:X which always have the 6-7dB drop. I mean just by playing a DTS track. This would indicate that something is indeed wrong with either my HW or FW..
|
|
|
Post by ttocs on Jun 28, 2020 14:10:28 GMT -5
Note I’m not talking about Neural:X which always have the 6-7dB drop. I mean just by playing a DTS track. Playing the native dtsHD MA 5.1 Bitstream, the volume is the same with Direct and Surround, no upmix.
|
|
|
Post by hsamwel on Jun 28, 2020 15:02:13 GMT -5
Note I’m not talking about Neural:X which always have the 6-7dB drop. I mean just by playing a DTS track. Playing the native dtsHD MA 5.1 Bitstream, the volume is the same with Direct and Surround, no upmix. Yeah it is for me as well. But I meant between tops and heights settings with only Surround mode. Actually just just found an much easier way to test. Thanks to the live update system. Start a DTS 5.1/7.1 track and play it with Surround mode. Then go to speaker settings Rear height and just cycle through all speakers. If everything works fine you should have NO difference as you are playing Surround mode. In my RMC-1 I get a clear negative volume difference if I choose ”Sm Top Rear”, ”Lg Top Rear” or ”DD Enabled Rear”. All other settings seems to keep the higher volume. Does not matter what settings I have in Front height. So it seems Surround mode gets the same volume as Neural:X if I have Top Rear active. This does not change Neural:X though. Atleast not for me anyway.
|
|
|
Post by PaulBe on Jun 28, 2020 15:12:01 GMT -5
I think a lot more people need to do this sort of testing to realize they too have issues. ...better yet, you need to accept the fact that other users DON'T have the issue that you think you have regardless of how hard you're trying to convince them that they do! Yep. And, there is a hole in this bucket that will never be fixed.
|
|
|
Post by markc on Jun 28, 2020 15:39:35 GMT -5
Regarding the lower volume, it now means absolutely nothing to me. When I look back at when I noticed a lower volume with Dirac 1 and the XMC-1 and didn't like it, but nothing I could do about it, I wanted to understand why this was. Not too long ago it was pointed out to me that the reason for this is due to the target curve being a REDUCTION of the peaks. So now I get it. Can't have same volume out when the processed volume has been reduced. The key word here is "processed". This doesn't specifically answer why Neural:X processing is lower in volume, but I've got to believe that there's something about the way it operates that makes it do this. So does the Marantz - while playing Neural:X at the same volume as DD - play at the same volume as Neural:X or DD on the RMC? Only those who take voltage measurements will know for sure. edit: I just got done measuring ONLY the Right Front Height:: Sm Front Height and found a nice sharp sound in The Dark Knight when the knife is in the guys mouth in the pool table scene. There is no difference in volume between Neural:X and DSU when player is set to PCM. This suggests to me that Neural:X is doing what I've read it does, emphasize the immersive speakers at the expense of the beds. So instead of RAISING the volume of the immersive speakers, it keeps those the same and LOWERS the volume of the other channels. Further, there was no discernable difference in dynamics. Prior to the above test I did a Factory Reset so all speaker Levels and Distances are set to zero. This is part of the problem. I didn't have any immersive speakers, just a simple 7.1 setup yet with Neural:X I was getting a 6dB drop in volume and change is sound. I always suspected that sound was being extracted to speakers I don't have set as present in my speaker preset configuration. &.1 with all the rest set to "none"
|
|
|
Post by ttocs on Jun 28, 2020 16:53:33 GMT -5
I didn't have any immersive speakers, just a simple 7.1 setup yet with Neural:X I was getting a 6dB drop in volume and change is sound. I always suspected that sound was being extracted to speakers I don't have set as present in my speaker preset configuration. &.1 with all the rest set to "none" Sound is not being extracted to speakers other than what you have setup. I just got done confirming this. But the Rear Surrounds play sound effects louder than the Surrounds.
|
|
|
Post by ttocs on Jun 28, 2020 16:53:56 GMT -5
Initially to prove that there was no sound going to channels other than what the system is setup for, I setup my system for 7.1, disconnected every output on the XMC-2, then using only the Center channel XLR cable I went one by one and replayed the same 5 seconds of audio that is in every channel that the system "could" be setup for. In other words, the suspenseful music and sound effect are resident in every native channel plus every channel that's possible to upmix to that is available on the XMC-2. This sound only occurred in the 7 channels plus sub while in Neural:X mode, not any other channels.
Then there's the RAISED VOLUME in the Rear Surrounds with Neural:X. Yes, it's about 3dB higher volume for this sound effect in the Rears than in the Surrounds or other speakers. So my theory is strengthened by this info that the upmixed channels are at the same volume predetermined by setup, and all other non-subwoofer channels are lowered in volume to make the Neural:X "effect" more prominent.
I did yet another test using Direct, which would only use 5.1 of the 7.1 channels available, vs Surround which only uses the 5.1 channels the audio track contains. Comparing the Right Surround channel in each audio mode resulted in the sound being louder by 3dB with Direct. This is not to "prove" anything, just to provide another comparison.
Neural:X messes with stuff far more than I would've imagined previously. It is affected by speaker phase, distance/level, and is preprogrammed to "enhance" the moving sounds by increasing their volume relative to the volume of the other non-upmixed channels. Some movies, it can be stated, are "improved" using this technology, and others suffer. My opinion is that they all suffer.
Each successive test I do forces me to conclude that the loosely termed "info" from DTS is vague on purpose.
I will state once again, if there is ANY information from DTS about how their Neural:X is supposed to work on every processor made, then please find it. I'm done looking after finding nothing. Saying that it will utilize any configuration out there is very misleading unless it is stated that x component will be routed to y speaker under z conditions.
I will state without reservation that the main channels are forced by Neural:X to be lower in volume so that the effects are louder by comparison and no other reason.
What this all says to me is that I'm very happy I personally don't want to watch any movie in anything other than the original soundtrack.
|
|
|
Post by markc on Jun 28, 2020 17:35:56 GMT -5
I will state without reservation that the main channels are forced by Neural:X to be lower in volume so that the effects are louder by comparison and no other reason Good that you tested it, so thanks. although, you misunderstand. .We are not blaming Neural:X, but the way Neural:X is implanted on these Emotivas Neural:X works fine! (on other manufacturers gear). It’s effects may not be to everyone’s taste, just like using Dolby Surround Upmixer, but the algorithms work. The Marantz 8805 does not suffer this volume loss. I am also 100% certain that 5.1 to 7.1 upmixing is not programmed within Neural:X to decrease the volume. It is the Emotiva doing that bit wrong. Again: My XMC-2 can’t do simple rear surround extraction of bitstreamed 5.1 DTS to 7.1 audio by de-matrixing the side surrounds. that is a fault in implementation, not the DTS coding itself.
|
|
|
Post by megash0n on Jun 28, 2020 17:50:03 GMT -5
I think a lot more people need to do this sort of testing to realize they too have issues. ...better yet, you need to accept the fact that other users DON'T have the issue that you think you have regardless of how hard you're trying to convince them that they do! Foolish nonsense.
|
|
|
Post by Mikomill on Jun 28, 2020 18:18:33 GMT -5
...better yet, you need to accept the fact that other users DON'T have the issue that you think you have regardless of how hard you're trying to convince them that they do! Foolish nonsense. Alright, let's move on. This subject doesn't need numerous posts deterring from the topic. We are all welcome to our own thoughts of about this Foolish Neural:X nonsense.
|
|
|
Post by Mikomill on Jun 28, 2020 18:19:43 GMT -5
I will state without reservation that the main channels are forced by Neural:X to be lower in volume so that the effects are louder by comparison and no other reason The Marantz 8805 does not do this volume loss. Yes it does, you just aren't "testing it" correctly.
|
|
|
Post by ttocs on Jun 28, 2020 18:28:14 GMT -5
The Marantz 8805 does not do this volume loss. This one is easy to test. Take a Marantz box and an Emotiva box, use PCM on the player to compare DSU to Neural:X (can't use bitstream - DSU won't upmix dts audio via bitstream), measure the voltage-in and voltage-out, compare results. I'd like to know. i am also 100% certain that 5.1 to 7.1 upmixing is not programmed within Neural:X to decrease the volume. It is the Emotiva doing that bit wrong. Please note that it's only the channels native to the audio track that are lower in volume. All the upmixed channels are louder. The upmixed channels are as loud as all the DSU channels are. This is what I measured. I didn't expect this result. It's just what the result happens to be. This is something anyone with an SPL meter can test, but do yourself a favor and only test one channel at a time, and pick a sound that is already in all of the native channels otherwise the test is useless. Again: My XMC-2 can’t do simple rear surround extraction of bitstreamed 5.1 DTS to 7.1 audio by de-matrixing the side surrounds. that is a fault in implementation, not the DTS coding itself. But my test today showed that dtsHD MA 5.1 does indeed upmix to 7.1. Neural:X does it using bitstream, and DSU (Dolby Surround) does it from a dts audio track when using PCM.
|
|
|
Post by megash0n on Jun 28, 2020 20:28:00 GMT -5
I will state without reservation that the main channels are forced by Neural:X to be lower in volume so that the effects are louder by comparison and no other reason Good that you tested it, so thanks. although, you misunderstand. .We are not blaming Neural:X but the way Neural:X is implanted on these Emotivas Neurak:X works fine! (on other manufacturers gear). It’s effects may not be to everyone’s taste, just like using Dolby Surround Upmixer, but the algorithms work. The Marantz 8805 does not do this volume loss. i am also 100% certain that 5.1 to 7.1 upmixing is not programmed within Neural:X to decrease the volume. It is the Emotiva doing that bit wrong. Again: My XMC-2 can’t do simple rear surround extraction of bitstreamed 5.1 DTS to 7.1 audio by de-matrixing the side surrounds. that is a fault in implementation, not the DTS coding itself. Somehow this continues to be missed.
|
|
|
Post by markc on Jun 29, 2020 1:29:33 GMT -5
The Marantz 8805 does not do this volume loss. This one is easy to test. Take a Marantz box and an Emotiva box, use PCM on the player to compare DSU to Neural:X (can't use bitstream - DSU won't upmix dts audio via bitstream), measure the voltage-in and voltage-out, compare results. I'd like to know. i am also 100% certain that 5.1 to 7.1 upmixing is not programmed within Neural:X to decrease the volume. It is the Emotiva doing that bit wrong. Please note that it's only the channels native to the audio track that are lower in volume. All the upmixed channels are louder. The upmixed channels are as loud as all the DSU channels are. This is what I measured. I didn't expect this result. It's just what the result happens to be. This is something anyone with an SPL meter can test, but do yourself a favor and only test one channel at a time, and pick a sound that is already in all of the native channels otherwise the test is useless. Again: My XMC-2 can’t do simple rear surround extraction of bitstreamed 5.1 DTS to 7.1 audio by de-matrixing the side surrounds. that is a fault in implementation, not the DTS coding itself. But my test today showed that dtsHD MA 5.1 does indeed upmix to 7.1. Neural:X does it using bitstream, and DSU (Dolby Surround) does it from a dts audio track when using PCM. 1) An ear will do it in a flash! We are not talking subtle changes but a slam dunk change on engaging Neural:X. YOu could quantify it to 0.01V is you had a meter, but it isn't required. 2 )I don't want my native front channels changed in volume at all when the two side surrounds of 5.1 are de-matrixed to 7.1. They shouldn't be. It is not part of dts coding and extra headroom is hardly required in the rear surround channels! 3) My Neural:X does indeed upmix to 7.1! That is not what I said. I said it can't do simple extraction of audio from side surrounds and leave all other channels untouched! DTS 5.1 to 7.1 upmixing is just being done incorrectly on the Emotiva's! There should not be a volume drop. By should not be, I don't mean "ought not to be, because I don't like it". I mean "should not be because the dts up-mixer doesn't do this". It never has in any iteration since 2000. That was 20 years ago.
|
|
|
Post by cwt on Jun 29, 2020 6:23:17 GMT -5
|
|
|
Post by markc on Jun 29, 2020 11:21:36 GMT -5
Yes, I think that again emphasises that the Emotiva is not doing DTS up-mixing right using Neural:X. If anything he is saying there is a barely detectable INCREASE in overall room volume with Neural:X ("Important to note: in casual listening, when all the ear-level speakers are running too, it's MUCH harder to hear the volume difference between Dolby Surround and DTS:Neural:X, but it might be important to you, so I thought it worth mentioning.") We are suffering a marked room volume decrease, that's with all speakers. Our issue is not because the heights speakers are increasing in volume and the lower bed speakers getting a compensatory decrease. If is more a case of sucking the sound out the room. Again: A 6db decrease in volume is 50% of the signal voltage arriving at the speakers. Hardly trivial. Again: We are not debating Dolby's decision on how they implement the Dolby Surround Upmixer or the decision made by DTS as to how they will extract and redistribute audio using Neural:X, or how much helicopter noise vs ambient crowd noise is up on the ceiling with each. It is the change to the rest of the audio when using Neural:X on the Emotiva's to that audio left behind in the bed speakers, and at a significantly lower volume. Maybe we should focus on the rather basic elementary function of 5.1 to 7.1 dematrixing. That isn't dependant on whether Dolby or DTS choose to extract ambient or effects or delay dependant or phased or non-phased. It is a simple procedure that Neural:X in the Emotiva's cannot do without flattening all the front channels when it should not be. (Only the side surround of the 5.1 should be changed, and decorrelated and correlated audio spread from the side surrounds to rear surrounds)
|
|
|
Post by cwt on Jun 29, 2020 13:29:16 GMT -5
We are suffering a marked room volume decrease, that's with all speakers. Our issue is not because the heights speakers are increasing in volume and the lower bed speakers getting a compensatory decrease. If is more a case of sucking the sound out the room. Again: A 6db decrease in volume is 50% of the signal voltage arriving at the speakers. Hardly trivial. Its notable with the options in the poll that people could interpret things separately ; saying yes to the 6db reduction but not perceiving any muddiness . The 8 votes saying yes is at odds with the number [including yourself];which are experiencing issues with the bed channel only configuration . This is the bane of any complicated poll .I get the impression also that bass management crossover settings are another variable that should be factored for. I hope this is settled with the next firmware one way or the other ;thinking you know if we had THX adaptive correlation this would affect the time and phase of a stereo pair of surrounds ; the sound field wouldnt collapse if you were off axis and you would have a more spacious sound Its good to note that Lonnie will 1 day reverse the dolby cross mixing rules ;with only 2 upmixers there is plenty of scope for say PL2X again or even the shelved emo dsp mode.Its niggles like this that makes you wait for dirac before buying
|
|