SME Posted August 29, 2016 Share Posted August 29, 2016 for completeness, a high Q ULF LS has some ripple in jriver. It equates to a minimum corner frequency of ~8Hz for completely predictable behaviour. I'm not sure if this is to be expected or not, probably irrelevant anyway but I thought I'd mention it (can't attach a pic due to space issues, not sure if http://yabb.jriver.com/interact/index.php?action=dlattach;topic=106708.0;attach=22139;imageis visible either) Is that supposed to display an image? It doesn't work for me. Quote Link to comment Share on other sites More sharing options...
3ll3d00d Posted August 29, 2016 Share Posted August 29, 2016 Is that supposed to display an image? It doesn't work for me.Try https://i.imgsafe.org/3ce614b15e.png Quote Link to comment Share on other sites More sharing options...
SME Posted September 1, 2016 Share Posted September 1, 2016 The ripple in those looks pretty minor, but it's still a sign of things not working right. IIRC, JRiver makes a point of using double precision arithmetic throughout, so I worried that my own DSP might have trouble with these as it handles audio as single precision float. (The biquad coefficients and intermediate data are double precision at least.) However, my filters look perfect from 1-100 Hz with Q=1.5, 2, 4, etc. I have no idea what is going wrong with JRiver. Have you seen problems with shelf filters on other platforms too? I know my MiniDSP 2x4 had a lot of issues with low frequency filters in general because it did computation in fixed point without enough precision. Quote Link to comment Share on other sites More sharing options...
3ll3d00d Posted September 1, 2016 Share Posted September 1, 2016 I haven't looked at other platforms tbh, not many have the ability to set something that low down. Quote Link to comment Share on other sites More sharing options...
Patzig Posted October 20, 2016 Share Posted October 20, 2016 This is a great thread, thank you for posting. A couple quick questions... 1. Applying these filters in JRiver: the LCR and back/surround filters are pretty straight forward, but should the "Sub" channel be used to apply the filters labeled for "LFE" in this thread? I ask because I'm not sure if the LFE filters mentioned here are meant for the LFE channel only, or actually meant for the sub channel (where I send all <80Hz frequencies from all speakers too) 2. Why are there sometimes multiple gain filters? Star Wars for example...2 x +6db and a +3db filter on the LFE instead of 1 x +15db filter. Do DSP's/JRiver not like filters applied over 6db? 3. Formal request for some BEQ on Star Trek Beyond would be GREAT, if possible! Thank you Quote Link to comment Share on other sites More sharing options...
3ll3d00d Posted October 20, 2016 Share Posted October 20, 2016 1. Applying these filters in JRiver: the LCR and back/surround filters are pretty straight forward, but should the "Sub" channel be used to apply the filters labeled for "LFE" in this thread? I ask because I'm not sure if the LFE filters mentioned here are meant for the LFE channel only, or actually meant for the sub channel (where I send all <80Hz frequencies from all speakers too) the filters are applied to input channels (e.g. LFE) not output channels (e.g. subwoofer). The JRiver sub channel simply refers to channel 4 which means the LFE input channel if the PEQ block is before bass management and the subwoofer output channel if the PEQ block is after bass management. This means you need to place a PEQ block before bass management and apply these filters there. Quote Link to comment Share on other sites More sharing options...
Kvalsvoll Posted October 29, 2016 Author Share Posted October 29, 2016 This is a great thread, thank you for posting. A couple quick questions... 1. Applying these filters in JRiver: the LCR and back/surround filters are pretty straight forward, but should the "Sub" channel be used to apply the filters labeled for "LFE" in this thread? I ask because I'm not sure if the LFE filters mentioned here are meant for the LFE channel only, or actually meant for the sub channel (where I send all <80Hz frequencies from all speakers too) 2. Why are there sometimes multiple gain filters? Star Wars for example...2 x +6db and a +3db filter on the LFE instead of 1 x +15db filter. Do DSP's/JRiver not like filters applied over 6db? 3. Formal request for some BEQ on Star Trek Beyond would be GREAT, if possible! Thank you The filters are for each channel individually, LFE means LFE channel before adding any rerouted bass from other channels. So, for JRiver you should implement filters for each of the LCR+surround+LFE channels according to the description given for the film. When LFE and bass from LCR+surround is routed to the sub output, you can apply filters to the sub output if you don't have access to the individual signals. It is not the same, but should give good results in many cases. Use filters specified for LFE channel. A good bass system has a dsp able to do this filtering. The multiple filters with smaller gain vs one filter with huge gain is to avoid too much overshoot due to the q, the response is not the same. 1 Quote Link to comment Share on other sites More sharing options...
Patzig Posted November 3, 2016 Share Posted November 3, 2016 the filters are applied to input channels (e.g. LFE) not output channels (e.g. subwoofer). The JRiver sub channel simply refers to channel 4 which means the LFE input channel if the PEQ block is before bass management and the subwoofer output channel if the PEQ block is after bass management. This means you need to place a PEQ block before bass management and apply these filters there. The filters are for each channel individually, LFE means LFE channel before adding any rerouted bass from other channels. So, for JRiver you should implement filters for each of the LCR+surround+LFE channels according to the description given for the film. When LFE and bass from LCR+surround is routed to the sub output, you can apply filters to the sub output if you don't have access to the individual signals. It is not the same, but should give good results in many cases. Use filters specified for LFE channel. A good bass system has a dsp able to do this filtering. The multiple filters with smaller gain vs one filter with huge gain is to avoid too much overshoot due to the q, the response is not the same. Thank you for the replies, sounds good. I watched a few movies with these settings...good stuff. Quote Link to comment Share on other sites More sharing options...
maxmercy Posted November 8, 2016 Share Posted November 8, 2016 Star Trek: Beyond BEQ This film improves significantly. After applying BEQ, to get equivalent dialogue level, you have to playback at +7dB your normal listening level. Pre/Post PvA: BEQ Solution: LFE Channel: 1. Gain -7dB 2. Low Shelf 27Hz, Q of 0.707 (Slope of 1), +6dB x5 filters (total of +30dB Shelf) LCRS: 1. Gain-7dB 2. Low Shelf 30Hz, Q of 0.707 (Slope of 1), +6dB x4 filters (+24dB total Shelf) 3. Low Shelf 30Hz, Q of 0.707 (Slope of 1), +3dB 4. HighPass Filter, 1.5Hz, 6dB/octave New Ratings after BEQ played back at Reference (+7dBRef): Level: 5 Stars (112.6dB Composite) Extension: 4 Stars (12Hz) Dynamics: 5 Stars (30.5dB) Loudest moment of the film when played back at +7dBRef is 126dB, 2dB under a 7.1 Worst Case Scenario, and only for an instant. Hottest prolonged effect (1/8 of a second) is 116dB. If your system can handle Reference with TDKR, you can run this solution at Equivalent Ref (+7dBRef). Try out the solution and report your opinions. JSS 2 Quote Link to comment Share on other sites More sharing options...
Kvalsvoll Posted November 8, 2016 Author Share Posted November 8, 2016 Nice, @maxmercy. How is your impression of the improved lf, does it sound good? Some movies improve a lot in the fr graph, but not so much in terms of sound quality. With very high boost at low freqs you often end up amplifying just noise. Quote Link to comment Share on other sites More sharing options...
maxmercy Posted November 8, 2016 Share Posted November 8, 2016 It is much better, good impact and appropriate heft for the oncreen action... The sound designers look like they knew what they were doing, and then the shelf was put in later. Zero Infrasonic noise noise in LFE, had to use the ~DC highpass on the LCRS, though. It's no 9 or Star Trek 2009 or War of the Worlds, but it is much improved. JSS 2 Quote Link to comment Share on other sites More sharing options...
bluescale Posted December 11, 2016 Share Posted December 11, 2016 LCR: 1. Gain -7dB 2. Low Shelf 25Hz, Q 1.13, +6dB 3. Low Shelf 26Hz, Q 1.13, +6dB 4. Low Shelf 27Hz, Q 1.13, +6dB 5. Low Shelf 52Hz, Q 0.5, +5dB LFE: 1. Gain -7dB 2. Low Shelf 18Hz, Q 1.13, +6.5dB (3 filters for 19.5dB correction) 3. Low Shelf 36Hz, Q 0.5, +5dB Surrounds: 1. Gain -7dB 2. Low Shelf 22Hz Q 1.13, +7dB 3. Low Shelf 23Hz Q 1.13, +7dB 4. Low Shelf 24Hz Q 1.13, +7dB 5. Low Shelf 25Hz Q 1.13, +7dB 6. Low Shelf 22Hz Q 0.5, +8dB Pacific Rim is now near 5-Star with this correction, and from the opening scene with 'Axehead', you know it right away. I highly recommend people try this one. MiniDSP nano-AVR .xml will be coming when I get time, or program an .xml yourself with the data above and miniDSP's spreadsheet: Biquad Spreadsheet JSS I'm finally going back and redoing my BEQ filters for JRiver by converting Q to S. I'd love it if someone could take a look at my calculations and make sure I've done the conversions correctly: LCR: 1. Gain -7dB 2. Low Shelf 25Hz, S 2.35, +6dB 3. Low Shelf 26Hz, S 2.35, +6dB 4. Low Shelf 27Hz, S 2.35, +6dB 5. Low Shelf 52Hz, S 0.51, +5dB LFE: 1. Gain -7dB 2. Low Shelf 18Hz, S 2.32, +6.5dB (3 filters for 19.5dB correction) 3. Low Shelf 36Hz, S 0.51, +5dB Surrounds: 1. Gain -7dB 2. Low Shelf 22Hz S 2.28, +7dB 3. Low Shelf 23Hz S 2.28, +7dB 4. Low Shelf 24Hz S 2.28, +7dB 5. Low Shelf 25Hz S 2.28, +7dB 6. Low Shelf 22Hz S 0.53, +8dB Quote Link to comment Share on other sites More sharing options...
3ll3d00d Posted December 11, 2016 Share Posted December 11, 2016 You could run a loopback measurement through those filters to see what the shape looks like. Quote Link to comment Share on other sites More sharing options...
maxmercy Posted December 11, 2016 Share Posted December 11, 2016 What formula are you using to convert Q to Slope? JSS Quote Link to comment Share on other sites More sharing options...
3ll3d00d Posted December 11, 2016 Share Posted December 11, 2016 The formula to convert Q to S is listed in http://www.musicdsp.org/files/Audio-EQ-Cookbook.txtand is 1/Q = sqrt((A + 1/A)*(1/S - 1) + 2) where A = 10^(dBgain/40) If you solve for S you get 1/((((1/Q)^2-2)/(A+1/A))+1) Quote Link to comment Share on other sites More sharing options...
bluescale Posted December 11, 2016 Share Posted December 11, 2016 You could run a loopback measurement through those filters to see what the shape looks like. Sorry if this is seems like asking the obvious, but I want to make sure I understand correctly. You mean run a loopback measurement of the entire movie to see if the measurement matches what was posted in the graphs? I could do that. I suppose I was hoping for something that didn't take ~2 hours to verify. That, of course, assumes that this is easy for someone to verify my work. After I posted my question, I realized there is a quick way to do a sanity check. I took a look at a coupe of the more recent posts where maxmercy posted both S and Q. In both cases, if I plug his Q value into my Excel spreadsheet, my S and his match. Quote Link to comment Share on other sites More sharing options...
3ll3d00d Posted December 11, 2016 Share Posted December 11, 2016 I mean measure the response of your combined set of filters by running a sweep through jriver, compare against the expected correction. Quote Link to comment Share on other sites More sharing options...
bluescale Posted December 11, 2016 Share Posted December 11, 2016 What formula are you using to convert Q to Slope? JSS If you solve for S you get 1/((((1/Q)^2-2)/(A+1/A))+1) If you'll recall, 3II3d00d, I had to come up with a slightly different formula for Excel to work correctly. I'm not sure why our experiences different, but here's what I finally settled on for S: S = (A^2*Q^2+Q^2)/(A^2*Q^2-2*A*Q^2+A+Q^2) For A I used what comes from the Audio EQ Cookbook: A = 10^(dBgain/40) More specifically, I set up a spreadsheet where you input the desired dB gain in cell B1 and the target Q value in B2. Then I have a formula for A in E1 and S in E2. Here's what those formulas look like: A: 10^(B1/40) S: (E1^2*B2^2+B2^2)/(E1^2*B2^2-2*E1*B2^2+E1+B2^2) Quote Link to comment Share on other sites More sharing options...
3ll3d00d Posted December 11, 2016 Share Posted December 11, 2016 An easy online alternative is to paste it into https://www.mathpapa.com/algebra-calculator.htm i.e. Paste into the box and click evaluate, you can then enter values for gain and q 1/((((1/Q)^2-2)/((10^(d/40))+1/(10^(d/40))))+1) Quote Link to comment Share on other sites More sharing options...
maxmercy Posted April 10, 2017 Share Posted April 10, 2017 Rogue One: A Star Wars Story BEQ Solution: LFE: Gain: -7dB Low Shelf: 20Hz, Slope 1.2, Q 0.776, +8dB High Shelf: 50Hz, Slope 1.5, Q 0.873, +8.5dB LCR: Gain: -7dB Low Shelf: 20Hz, Slope 1.75, Q 0.943, +8dB (2 filters for total gain of 16dB) Low Shelf: 40Hz, Slope 0.5, Q 0.5, +2dB Surrounds: Gain: -7dB Low Shelf: 20Hz, Slope 1.5, Q of 0.872, +8dB (3 filters for gain of 24dB) Low Shelf: 40Hz, Slope 0.5, Q of 0.5, +3.5dB Playback for equivalent dialog level will be +7dB. No scene is louder than 126dB if played back at equivalent reference. Let me know what you guys think. The high shelf is OK on LFE as the very steep DTS filter is in place on the original LFE track. JSS 2 Quote Link to comment Share on other sites More sharing options...
Kvalsvoll Posted April 14, 2017 Author Share Posted April 14, 2017 Lots of movies has this 30hz boom-boost and falls off below - question is, are they similar enough to allow for one simple predefined filter solution. BassEQ with several filters for each channel is fine, and will give the best results. However, if I just want to watch a movie - NOW - this is not a practical solution. It takes to much time and effort to implement either you are doing it in dsp for playback or remastering the soundtrack. Since many movies share similar frequency balance defects, one or a small selection of predefined filters could improve things a lot - actually improving the overall sound quality by magnitudes - and it would be easy to select if predefined in the bass system dsp. A bass-system dsp with option for presets could be configured with a few filters, covering most of the movies: 1: Default - flat response (flat ref to chosen target). 2: ULF boost - +10dB < 25hz. 3: ULF boost + boomfix - +10dB < 25hz, -3dB at 30hz. Example #3: When I want to watch SW-RO, I go to data-bass, observe the frequency spectrum chart, and choose preset #3. Done, start the movie. This is something that could work for a lot of people. For new readers: Now, of course, the soundtrack should be played "as intended", but that will have to wait until they actually present a soundtrack with decent sound quality having reasonable full frequency range balance, and this will not happen until they install proper sound systems in their studios. When "as intended" is boomy bass with lack of realism and dynamics and refinement there is huge potential for improvement, and if you really believe there is no significant difference to balanced, full frequency range sound, you are lucky because you can look forward to having a new experience for what sound reproduction actually can be like. 1 Quote Link to comment Share on other sites More sharing options...
Kvalsvoll Posted April 16, 2017 Author Share Posted April 16, 2017 Rogue One: A Star Wars Story BEQ Solution: LFE: Gain: -7dB Low Shelf: 20Hz, Slope 1.2, Q 0.776, +8dB High Shelf: 50Hz, Slope 1.5, Q 0.873, +8.5dB LCR: Gain: -7dB Low Shelf: 20Hz, Slope 1.75, Q 0.943, +8dB (2 filters for total gain of 16dB) Low Shelf: 40Hz, Slope 0.5, Q 0.5, +2dB Surrounds: Gain: -7dB Low Shelf: 20Hz, Slope 1.5, Q of 0.872, +8dB (3 filters for gain of 24dB) Low Shelf: 40Hz, Slope 0.5, Q of 0.5, +3.5dB Playback for equivalent dialog level will be +7dB. No scene is louder than 126dB if played back at equivalent reference. Let me know what you guys think. The high shelf is OK on LFE as the very steep DTS filter is in place on the original LFE track. JSS Fixed SW-RO yesterday, briefly saw some scenes earlier today - huge improvement, and I found the sound to be quite good, despite all the negative comments. This movie has great dynamic range, the loudness varies between scenes, some are softer and some are loud. Removed the 30hz-boom and may have ended up with quite high level below 20hz, added only a moderate midbass-lift. Sounded quite balanced on the scenes I saw - the beginning and some later. I will post a spectrum and the filters later. Quote Link to comment Share on other sites More sharing options...
Kvalsvoll Posted April 16, 2017 Author Share Posted April 16, 2017 SW-RO BassEQ filters: lcr:sfm 16hz 2.0 +10dBlfe:gain -10dBsfm 16hz 2.6 +12dBsfm hs 32hz 1.6 +1.6dBhpf 12hz 24dB/octhard limiter -10dBgain +10dB Spectrum for fixed LFE: In this scene (from the opening) you can now feel the spacecraft passing by: Quote Link to comment Share on other sites More sharing options...
SME Posted April 16, 2017 Share Posted April 16, 2017 SW-RO BassEQ filters: lcr: sfm 16hz 2.0 +10dB lfe: gain -10dB sfm 16hz 2.6 +12dB sfm hs 32hz 1.6 +1.6dB hpf 12hz 24dB/oct hard limiter -10dB gain +10dB ... Is this your own BEQ solution for Rogue One? What does sfm mean? How much content do you lose with "hard limiter -10dB" in your chain? FWIW, my BEQ implementation has no headroom limitation aside from what the equipment can do. Quote Link to comment Share on other sites More sharing options...
Kvalsvoll Posted April 16, 2017 Author Share Posted April 16, 2017 Is this your own BEQ solution for Rogue One? What does sfm mean? How much content do you lose with "hard limiter -10dB" in your chain? FWIW, my BEQ implementation has no headroom limitation aside from what the equipment can do. This is the quick-fix-BassEQ, the one JSS presented is the state-of-the-art. Perhaps a bit too quick with the copy-paste: sfm = shelf filter min = Low Shelf sfm hs = High Shelf The effect of the hard limiter remains to be verified, I did not get to those scenes in my brief testing. The reason for this is that I ended up getting some heavy ulf peaks in some very few scenes, the 12hz high-pass fixed most, and the hard limiter does the rest. This is on the lfe channel, it should not introduce audible distortion and no content is lost. Doing the same on any other channel would destroy the sound, as the limiter would distort and remove all other higher frequency content when the signal is limited. Doing the hard limiter in the signal processing makes no difference compared to letting the bass system dsp limit the signal during playback - the result is the same. The ulf boost is a bit heavy - JSS uses only 8dB where i put in a +12dB boost with a much higher q on the filter, making for much more level in the 5hz - 20hz range. I wanted to test a simpler filtering scheme, more suitable for bass system dsp presets. If that works, it becomes so easy to play back a movie with BassEQ that anyone with a similar dsp can do it simply by looking up the movie on data-bass and select the most suitable preset. Remember that the alternative is NO EQ, with too low ulf level and too much 30hz boom. Even if it isn't perfect, the improvement will be very significant, to the extent that the experience is lifted to a different level. 1 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.