|
|
Joined: Nov 2000
Posts: 2,766 Likes: 6
Robert.W Great Green Arkleseizure
|
|
Robert.W Great Green Arkleseizure
Joined: Nov 2000
Posts: 2,766 Likes: 6 |
George,
after reading your post, I understand that this -3dB is needed for the headroom and center level in the different presets. Why M have a positive dB adjustment for the center I dont know, I mean the center speaker is allways louder as the mains and have to be calibrated lower . Subtracting 3bit from the original signal reduces the theoretical dynamics. Having 96dB on the CD as a maximum and 0dB as a minimum,you get "only" 93dB after the prozessor.
But you cant compare this 1:1 to the practical dynamic at your home. The dynamic at home are limited in the maximum SPL for your ears and your equipment, and at the minimum level for the ambient noise. So have maximum 100dB(thats really loud), your rooms have at least 40dB ambient noise,maybe with luck only 35dB. So your practical dynamics are 100 minus 40(35) means 60(65)dB. So the (bit) dynamics from your CD get "scaled" to a SPL dynamic of 60dB. When you loose 3db from 96dB you have a new scale from 93 to 60 dB instead of 96 to 60dB. You can imagine that this is not a big difference on the paper.
And the most important thing is, this dynamics you will only get when you turn the volume nearly full up, so the 0dB (full scale) from your CD get to 100dB SPL. If you dont listen to this nearly full up volume, you will loose dynamics again.
But how its sound ? I testet only in stereo with two 8k and stereo preset.
I get a test record from Nippon Columbia (Denon) C39-7147 with a lot of usefull test signals on it. First I want to test when the DSP get digitally saturated. There are different frequencies with full scale means 0dB on it.And you have that for L and R channel, so you can test it independly.
So I modified the configfile with the text editor as George described. The only difference is that I let one channel at 0 dB, so I can compare easily with just jumping to the next track. For example the L speaker is calibrated to 0dB and the right one to +3dB. As the CD have tracks for each channel you can skip between the tracks and compare it with your ear.
As you put +dB in the file the output of the prozessor get louder and regarding to impressions it correspond with the volume display-means 3dB more positive you have to subtract 3 units from the volume display to get the same SPL.(Sorry have only my ears no SPL meter).
So I dit make some files with left 0dB and right starting with +15dB down to 3,5 dB. Then I listen with skip the tracks between the different dB levels. I can clearly hear when the DSP is digitally saturated.
Its easy to hear it with 1kHz and with 10kHz. With 1kHz the frequency stays at the same as expected, but the overtones? are different. Its like when you hear a guitar or mandoline play the same note(same frequency) but it sounds different.
With 10kHz have to pop the volume a little to hear the 10kHz and maybe turn the head a little until you will hear it.But the good thing on that is that this more sensitive as the 1 kHz tone.With 10kHz you hear a second tone with a different(lower) frequency when the signal is saturated. So I found out that + 3.5 dB saturated the sound and 3.0 dB not. Its a good test to modify with higher positive dB maybe 9 or 12 to easily get the impression how a saturated signal sounds.
So I finally use two file with 0dB(L&R) and +3dB(L&R) to make a sound test with a very well known CD. I dit change that 3 times to get shured about my impressions. The sound with +3dB (play with 3 volume increments lower) seems to smoother with voices maybe even smoother when the voices are at low levels.The piano and accustic instruments sound a littler more "natural" with a little more body. I will test more on that with different recordings so I can describe that better and check that Iam not "head pregnant"..... I also will test this +3dB thing with DVDA to find out how that will be sound.
George thanks for playing and testing to found that out, I allways thinking its just 3dB lower volume what can be corrected with the volume knob.But its definetly more.
The bad thing is, that the center cannot adjusted to that +3dB, as then the speaker is to dominate. I have before all that testing the L/R at 0dB and the center at -1.5dB.
When I now calibrate the center also to +3dB, the speaker is to loud.Of coarse you can adjust the level of the center in that presets where a seperate center level adjustment is,but what are you doing with the other presets where no extra center level adjustment is? That means, that in that preset the center is to loud.
So only two possibilities are left: a)Calibrate the center to +1,5dB(in my case with 1.5dB difference between the mains and center) and the mains to +3dB. b)Adjust L/R/C to +3dB and listen only to that presets where you can adjust the center level in the preset(to 1.5dB in my case)
George it would be interesting now to compare the sound of the DSP with direct plug in a CD transport in the speaker and going through the prozessor as there should be no SPL difference.
But this issue is something what M should remove immediately. Its a minor thing for them and the system sounds better.
thanks George
Robert
800v3/218/271/8k.2/5k5s  /5kC/DSP3200/Reviver/ TT-ELP/Graham Slee Revelation/Trinnov Altitude 32/OPPO 205 __Ayon/HB/Shunyata/Miltzow -- power cables, __Ayon/Jungson/Miltzow --analog, Miltzow-- digital, __Vortex/WMA ,
|
|
|
|
|
Joined: May 2000
Posts: 8,364 Likes: 14
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: May 2000
Posts: 8,364 Likes: 14 |
Raw test files are at http://softronix.com/pictures/levelperfect/ if you want to look at them.
- Fiddler (IMHO)
If you think listening to music is fun, try learning an instrument. HT: G61RSL, 4x5K, 5.5KHC, 2xSC6000, Transporter, HD621, Philips BDP7501 4K, Tivo, 65" LG OLED 4K C6 Computer: Foobar2000, 5K, Mojo, HiFiMan HE1000 V2
|
|
|
|
|
Joined: Nov 2001
Posts: 691 Likes: 4
Formerly Meridian506 Paranoid android
|
|
Formerly Meridian506 Paranoid android
Joined: Nov 2001
Posts: 691 Likes: 4 |
Thanks for posting the waveforms George. Looks like a big question to add to the HH2/3 list 
Living room: DSP5000.3 WFH Office: M33 Bedroom: P200ZQ
|
|
|
|
|
Joined: May 2000
Posts: 8,364 Likes: 14
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: May 2000
Posts: 8,364 Likes: 14 |
George,
George it would be interesting now to compare the sound of the DSP with direct plug in a CD transport in the speaker and going through the prozessor as there should be no SPL difference.
Robert I did do that. It's identical. I even measured it and posted wave forms to show it. I said that a several times in fact  Good work Robert. I think if you can just get your mains to +3dB the increase in detail is great. If you can get center too that's icing on the cake. I agree 1.5dB boost on center is probably too much. By the way Meridian could EASILY come up with firmware to allow us to manually set an offset in the volume in the DSP speaker. This would allow all DSP speakers to be set to +3dB with no calibration level compromise.
- Fiddler (IMHO)
If you think listening to music is fun, try learning an instrument. HT: G61RSL, 4x5K, 5.5KHC, 2xSC6000, Transporter, HD621, Philips BDP7501 4K, Tivo, 65" LG OLED 4K C6 Computer: Foobar2000, 5K, Mojo, HiFiMan HE1000 V2
|
|
|
|
|
Joined: Dec 2000
Posts: 866 Likes: 2
Pan-dimensional being
|
|
Pan-dimensional being
Joined: Dec 2000
Posts: 866 Likes: 2 |
Subtracting 3bit from the original signal reduces the theoretical dynamics. Having 96dB on the Cd as a maximum and 0dB as a minimum,you get "only" 93dB after the prozessor. Actually this is not the case. It would be if the resulting signal would be 16 bits, but in 24 bits the original 96dB S/N should stay the same with the volume 3dB reduced. I also applied the 'mod' to my setup and the 1st impression is pretty positive. I need to do more testing, but at least I did not hear any nasty side effects. There are a few things I do not understand: 1) I have no idea whether the 500 series processors are level perfect, but they sound better when you give them a headroom of 3 to 4 dB. Obviously the G68 has already 3dB headroom, but the G68 sounds better without it. This is just the opposite! 2) George's reports indicate it only work at exactly +3dB, so when the headroom is removed completely. This is pretty strange, because for starters there should be no difference at all changing the volume 3dB in a 24 bits range. This could mean a bug in the G68 firmware. This would be a nice fix along all other things for the multiple times promised but long time overdue firmware update for the G68.
|
|
|
|
|
Joined: Nov 2000
Posts: 2,766 Likes: 6
Robert.W Great Green Arkleseizure
|
|
Robert.W Great Green Arkleseizure
Joined: Nov 2000
Posts: 2,766 Likes: 6 |
I did do that. It's identical. I even measured it and posted wave forms to show it. I said that a several times in fact  Where is your post,its short or long time ago? Dit you make that with the modified config file? If you make it that way and there is no difference at all- then the prozessor do nothing (no dithering and.. whatever ) to the signal ? I thought it will make things similar as the 518 can do it? Robert
800v3/218/271/8k.2/5k5s  /5kC/DSP3200/Reviver/ TT-ELP/Graham Slee Revelation/Trinnov Altitude 32/OPPO 205 __Ayon/HB/Shunyata/Miltzow -- power cables, __Ayon/Jungson/Miltzow --analog, Miltzow-- digital, __Vortex/WMA ,
|
|
|
|
|
Joined: May 2000
Posts: 8,364 Likes: 14
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: May 2000
Posts: 8,364 Likes: 14 |
Subtracting 3bit from the original signal reduces the theoretical dynamics. Having 96dB on the Cd as a maximum and 0dB as a minimum,you get "only" 93dB after the prozessor. Actually this is not the case. It would be if the resulting signal would be 16 bits, but in 24 bits the original 96dB S/N should stay the same with the volume 3dB reduced. I also applied the 'mod' to my setup and the 1st impression is pretty positive. I need to do more testing, but at least I did not hear any nasty side effects. There are a few things I do not understand: 1) I have no idea whether the 500 series processors are level perfect, but they sound better when you give them a headroom of 3 to 4 dB. Obviously the G68 has already 3dB headroom, but the G68 sounds better without it. This is just the opposite! 2) George's reports indicate it only work at exactly +3dB, so when the headroom is removed completely. This is pretty strange, because for starters there should be no difference at all changing the volume 3dB in a 24 bits range. This could mean a bug in the G68 firmware. This would be a nice fix along all other things for the multiple times promised but long time overdue firmware update for the G68. I totally agree with your comments. I'm suspecting it's a bug too. If the math is done right .5 bits of 24 should not make much difference. Especially internally it's probably 48bits !!! I'd be curious to test this out on a 565 and a 568.1 and .2 (with Beta Firmware) all on DSP Speakers. Maybe the "bug" got introduced at the same time as the Trifield Width bug did. But I don't want folks to point fingers at Meridian. Be kind, be curious but don't be rude to them. Hopefully they are readying this (they often do) and will look into it. We have no proof it's a bug and it may be a well understood compromise they chose to make.
- Fiddler (IMHO)
If you think listening to music is fun, try learning an instrument. HT: G61RSL, 4x5K, 5.5KHC, 2xSC6000, Transporter, HD621, Philips BDP7501 4K, Tivo, 65" LG OLED 4K C6 Computer: Foobar2000, 5K, Mojo, HiFiMan HE1000 V2
|
|
|
|
|
Joined: May 2000
Posts: 8,364 Likes: 14
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: May 2000
Posts: 8,364 Likes: 14 |
I did do that. It's identical. I even measured it and posted wave forms to show it. I said that a several times in fact  Where is your post,its short or long time ago? Dit you make that with the modified config file? If you make it that way and there is no difference at all- then the prozessor do nothing (no dithering and.. whatever ) to the signal ? I thought it will make things similar as the 518 can do it? Robert I'll give you a break because English is obviously not your first language  •It gives us 3 things. You can indeed make DSP speakers louder than 0 and some folks have struggled with this. With out a doubt, now when I compare Direct Wired Transporter to DSP6000 it sounds exactly the same as going through G68 with Direct DSP preset.I believe not only are the levels matched (so I can compare them instantly level matched) but a veil that the G68 was presenting is now gone. There is no reason it would not continue to do Resolution enhancement (which includes dithering). It comes out 24 bit (See Wiki Wave forms). It's still doing upsampling too. It's avoiding a costly processing step or possibly a buggy processing step.
- Fiddler (IMHO)
If you think listening to music is fun, try learning an instrument. HT: G61RSL, 4x5K, 5.5KHC, 2xSC6000, Transporter, HD621, Philips BDP7501 4K, Tivo, 65" LG OLED 4K C6 Computer: Foobar2000, 5K, Mojo, HiFiMan HE1000 V2
|
|
|
|
|
Joined: Feb 2002
Posts: 8,816 Likes: 87
President of the Imperial Galactic Government
|
|
President of the Imperial Galactic Government
Joined: Feb 2002
Posts: 8,816 Likes: 87 |
George,
you put the consequences very well above w.r.t. what this actually means. Is it a bug or simply a design decision? If it's a design decision then it would be helpful to understand it a little more. If it's a bug then it needs looking at.
Is it linked to the Trifield bug...unlikely IMHO but that bug does set a precedent w.r.t. M's software controls and testing procedures. Anything is possible.
I doubt that M will comment here but one would hope they are following this thread closely as it is by far the most important but of customer tech feedback for quite some time.
Rgds, Nick
|
|
|
|
|
Joined: Nov 2000
Posts: 2,766 Likes: 6
Robert.W Great Green Arkleseizure
|
|
Robert.W Great Green Arkleseizure
Joined: Nov 2000
Posts: 2,766 Likes: 6 |
Thanks for the brake, but I think my english is slightly better than your german;-)) Dit i was unpolite or what?
What I mean,if direct plug in the DSP sounds the same as with the +3dB via the prozessor as you posted- then there should be not difference at all. And if there is no sound differences, then what is doing the prozessor ? I know it can upsample and can make 24 bit resolution, but when you hear no difference-what is left? Can you explain that,please?
800v3/218/271/8k.2/5k5s  /5kC/DSP3200/Reviver/ TT-ELP/Graham Slee Revelation/Trinnov Altitude 32/OPPO 205 __Ayon/HB/Shunyata/Miltzow -- power cables, __Ayon/Jungson/Miltzow --analog, Miltzow-- digital, __Vortex/WMA ,
|
|
|
|
|
Joined: Nov 2000
Posts: 2,766 Likes: 6
Robert.W Great Green Arkleseizure
|
|
Robert.W Great Green Arkleseizure
Joined: Nov 2000
Posts: 2,766 Likes: 6 |
Thanks for the explaination from 16bit to 24 bit, makes sense. I dit tested the +3dB boost with a 861 not an G68.
800v3/218/271/8k.2/5k5s  /5kC/DSP3200/Reviver/ TT-ELP/Graham Slee Revelation/Trinnov Altitude 32/OPPO 205 __Ayon/HB/Shunyata/Miltzow -- power cables, __Ayon/Jungson/Miltzow --analog, Miltzow-- digital, __Vortex/WMA ,
|
|
|
|
|
Joined: Dec 2000
Posts: 866 Likes: 2
Pan-dimensional being
|
|
Pan-dimensional being
Joined: Dec 2000
Posts: 866 Likes: 2 |
We have no proof it's a bug and it may be a well understood compromise they chose to make. I disagree. The proof is in the fact that the effect (absense of veil) is only noticeable at one specific level setting, while it should be the same at any level setting.
|
|
|
|
1 members (RobertW),
363
guests, and
3
robots. |
|
Key:
Admin,
Global Mod,
Mod
|
|
|
Forums18
Topics31,249
Posts297,436
Members5,503
| |
Most Online4,788 May 22nd, 2026
|
|
|
|