Tilt control

[Geoffrey Armstrong]

A friend/customer of mine recently sent me this article re-published from Stereophile's archives and originally from 1991:

https://urldefense.proofpoint.com/v2/url?u=https-3A__www.stereophile.com_content_humidity-2Dconcert-2Dhall-2Dsound-2Dspectral-2Dtilt&d=DwQFaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=fIjp4PqByLla5KkLpLt3xfztydNaNIS7GhxiYSiM7Ko&m=EZW7l50-mQQEri3zfapZAKsGByqyWijoM3q87jiPF-Y&s=58i5fGf6dNJK1YYjK3umaAFfkawTT_LZUkl8WLc0488&e=

 

It basically advocates the use of a "Tilt control".

 

Do you have any recordings, that sound too bright, cool or thin? Or recordings that are too dark, muddy, warm?

 

A 'Tilt control" can correct these by allowing you to pivot, the frequency balance around a central frequency axis (e.g. 1000hz). Other than warming up or cooling down a recording to "correct" the spectral frequency balance, the "Tilt control" has no effect on the sound.

 

As explained in the article, a 'Tilt control" is not the same as tone controls, such as Bass and Treble controls that used to be common on audio gear. The "Tilt Control" is both subtle and powerful at the same time. Subtle in terms of how it effects the frequency balance; but powerful in the sense that just a small change can transform recordings that have problems in these areas.

 

The bad news is that 'Tilt controls" were introduced by Quad and can only be found on their vintage 34 and 44 pre-amps and their newest pre-amp. Audiophiles and High-End manufacturers avoided them like the plague, probably confusing them with "tone" controls, which were banished from any respectful high-end gear from about the same time that Quad introduced this innovation.

 

I went searching to find out if such a thing exists in software form today that could benefit any modern computer based audio system and I found this…

 

https://urldefense.proofpoint.com/v2/url?u=https-3A__www.tb-2Dsoftware.com_TBProAudio_stilt.html&d=DwQFaQ&c=euGZstcaTDllvimEN8b7jXrwqOf-v5A_CdpgnVfiiMM&r=fIjp4PqByLla5KkLpLt3xfztydNaNIS7GhxiYSiM7Ko&m=EZW7l50-mQQEri3zfapZAKsGByqyWijoM3q87jiPF-Y&s=ZizfNZ2k09NmoeagVK5BO_vEGykJ0BRR3uMhv4E60Ds&e=

 

In this page the company acknowledge Quad's innovation. They also include a health warning that it does increase latency. A dirty word in digital audio on a par with jitter in its undesirability. The latency they say will be mitigated by the DAW (digital audio workstation). That's because, of course, it's aimed at audio professionals working in the studio, to produce the recordings we listen to. Not really intended to be used by audiophiles as a post-processing DSP fix, after the fact.

 

The good news though, is that it works, and it's free!

 

You will need a software player that allows hosting of VST or Audio Unit plug-ins.

 

1. Audirvana
2. J River

… are two that I've tried.

 

sTilt consists of one main dial control. Pictured as a clock-face 12:00 O'clock would be applying no effect, five past twelve would be cooling the recording down a little, creating greater transparency in a recording that started off too warm and where the lower frequencies might have been obscuring details further up the frequency range. 11:55 does the opposite; warms up an overly cool/bright recording. In both cases I've found the recording in question becomes more "transparent" in the true sense. A recording that needed to be cooled down, will now have cleaner and seemingly more powerful bass, that might even seem to go deeper. A recording that needed warming up will now have instruments and voices that now sound more natural, more like themselves.

 

What about that latency issue?

 

sTilt comes with several quality settings. The degree of latency will increase as you go further up the quality scale. The Max quality button appears in red, supposedly to warn you of this. Max, is the quality setting I used though, after trying "Medium" and being disappointed with the quality of the result.

 

Latency will also increase depending on the amount of tilt you apply. I found five past the hour or five before the hour sufficient to correct most recordings I've tried it on. Some call for a more extreme correction at approaching 11:50 or 12:10, which will lead to greater latency.

 

With J River on Windows I couldn't detect latency at the 11:55 and 12:05 settings; but I definitely heard it at 11:50 or 12:10. Once heard I believe I could also detect at the 5 minute settings, although it bothered me less.

 

Audirvana on Windows fared worse, with the latency becoming unbearable for me.

 

Audirvana on Mac on the other hand fared best of all. I detect no obvious audible latency, even at the more extreme settings.

 

Could this be because Mac has inherently less latency in its Core Audio stack and drivers, and is favored by audio pros for this reason? I don't know. A couple of Google searches threw up articles that suggest this.

 

Microsoft have recognized this problem and have worked to reduce it, apparently, on the latest versions of Win 10. Still I detected it easily and that was when talking to an ASIO driver.

 

Incidentally, if you use Audirvana as a UPnP server, the benefits of any AU processing, such as sTilt you may apply, will also be sent to the audio renderer, which could of course, be running on a different machine with a different OS. I've no idea how Linux fares on the latency issue.

 

Thoughts anyone?

 

[jamesg11]

Really interesting.

Recordings that weren’t responsive to ‘tilt’?

Connection to room correction requirements?

[copy_of_a]

11 hours ago, Geoffrey Armstrong said:

A 'Tilt control" can correct these by allowing you to pivot, the frequency balance around a central frequency axis (e.g. 1000hz). Other than warming up or cooling down a recording to "correct" the spectral frequency balance, the "Tilt control" has no effect on the sound.

I love „Tilt“ EQs since they are mostly invisible. I wouldn‘t say they have no effect on the sound as this depends on the filters and the center frequency. The center of the audible frequency range (20Hz - 20kHz) in a logariythmic scale is 632Hz and flat „Tilt“-EQs with a center frequency close to 632Hz maximally preserve the relationship of the instruments ... only making the overall presentation brighter or darker... just as you said.

I‘ve once tried sTilt but to me it sounded somehow artificial. Possibly due to it being linear phase. And minimum phase sounded weak to me.

The cleanest, most transparent „Tilt“ I know of is from TDR Audio in the Slick EQ Mastering plugin and/or the Slick EQ GE plugin (technically it‘s the same in both the plugins).

Another great choice is Sonimus „Sweetone“. It‘s a vintage-style processor but if you disable the emulated preamp you get a clean and extremely nice softly shaped tilt centered at 650Hz. You can also switch it to a center frequency of 2k ... which is a great option when you think about target house curves ...

As far as latency goes this the price to pay with linear phase EQs; although the host (eg. Audirvana) actually should compensate ... which is also really a non-issue when playing back audio files.

Can‘t comment wrt Windows...

[Geoffrey Armstrong]

Thanks so much. I didn't realize there were other choices out there and I will definitely check out the two you mentioned.

 

I agree the host should really compensate and this seems to be the case with Audirvana on Mac, though not on Windows, judging by my listening.

 

Obviously in a Pro environment, there could potentially be a lot of DSP effects plug-ins in use. I would expect recording and mastering engineers would try to use those with the lowest latency and again the latency should be compensated by the DAW.

 

What I heard of as obvious latency was akin to added artificial reverb. At least I attributed the effect to delays caused by latency. It could be that some of us have become tolerant of this when using various forms of DSP.

 

The big question for me is should we avoid DSP to keep latency to an absolute minimum, or does this detract from the main benefits of computer based audio?

 

[Tintinabulum]

Just for the record, Meridian used "tilt" on their digital active speakers D600s in the late 80's. 

[Geoffrey Armstrong]

18 minutes ago, jamesg11 said:

Really interesting.

Recordings that weren’t responsive to ‘tilt’?

Connection to room correction requirements?

Well I'll start with recordings that were responsive to tilt. Joni Mitchell's Blue, which sounds too bright on every system I've heard it on. Listen when she sings loud on "All I want" (the first track). Sure she was young and her voice was bright; but on digital this gets into the "glassiness" category. Warm her up and we're back to something more akin to a good analogue system of the period. As though this was the kind of system the album was mastered for. This goes for the high-rez version on Qobuz and the MQA on Tidal (especially the MQA).

 

A favorite more recent recording of mine is Alejandra Ribera's La Boca. Nothing too much wrong with this recording. When I take the dial 5 minutes in the positive direction though, more detail is revealed that, I believe, was a little obscured by lower frequencies.

 

Also try warming up some of the brighter BSO recordings mentioned in the Stereophile article. I found these needed a more aggressive setting to sound their best.

 

My feeling is; the recordings that need it, will need the same approximate correction, independently of systems, rooms and tastes. Frequency spectrum inbalances are not corrected to suite individual tastes and circumstances, in my view. They are needed to bring the frequency balance back to where it should have been in the first place. It's not a fix for your room, it's a fix for the recording itself. You've selected and tuned your system and room to be where you believe it should be, probably using some known references. Most recordings shouldn't need tilting. Those that do though, may be indispensable recordings in your collection.

 

So this is an area where I feel we could all find some common ground, regardless of systems, rooms and tastes. Perhaps it would even be possible to test this by listing recordings that needed "tilt" correction in one direction or another and see where we might all be in agreement.

 

I have a few very different systems at my disposal. The essential characteristics of these recordings though, holds true for me, regardless of the system or room.  I see no reason tilt "correction" shouldn't be used in conjunction with room correction.

[copy_of_a]

24 minutes ago, Geoffrey Armstrong said:

What I heard of as obvious latency was akin to added artificial reverb. At least I attributed the effect to delays caused by latency. It could be that some of us have become tolerant of this when using various forms of DSP.

 

The big question for me is should we avoid DSP to keep latency to an absolute minimum, or does this detract from the main benefits of computer based audio?

Hups?!

No... there should not be any audible effect.

Propper implemented latency compensation delays the proccessed audio as a whole (by XY samples) ... without effects like „reverb“ or so. It should simply start to playback a little later...

Now, if the latency compensation is off this could be caused by the plugin, by the host or by the way they interact. Unfortunately this is difficult to find out. I have some plugins that do not work (wrt latency compensation) in Audirvana (on Mac) but the very same plugins work just fine in other hosts. Tricky...

I‘d simply use one that works - and for Tilt I‘d go minimum phase or mixed phase but not linear phase...

[Geoffrey Armstrong]

7 minutes ago, Tintinabulum said:

Just for the record, Meridian used "tilt" on their digital active speakers D600s in the late 80's. 

So Quad weren't the only ones. Of course Boothroyd Stewart and Quad were neighbours, so I'm sure they shared ideas over a few beers occasionally :)

 

[Geoffrey Armstrong]

2 minutes ago, copy_of_a said:

Hups?!

No... there should not be any audible effect.

Propper implemented latency compensation delays the proccessed audio as a whole (by XY samples) ... without effects like „reverb“ or so. It should simply start to playback a little later...

Now, if the latency compensation is off this could be caused by the plugin, by the host or by the way they interact. Unfortunately this is difficult to find out. I have some plugins that do not work (wrt latency compensation) in Audirvana (on Mac) but the very same plugins work just fine in other hosts. Tricky...

I‘d simply use one that works - and for Tilt I‘d go minimum phase or mixed phase but not linear phase...

Thanks. That's the approach I've taken. Audirvana on Mac with sTilt works well for me at the moment. As I said, I'll try the others you mentioned though, to see if I get even better results.

[copy_of_a]

Just now, Geoffrey Armstrong said:

Thanks. That's the approach I've taken. Audirvana on Mac with sTilt works well for me at the moment. As I said, I'll try the others you mentioned though, to see if I get even better results.

TDR Slick ( M or GE) is as good as it gets. Really top notch super clean tools!

Sweetone has a some „colour“ (even without the preamp) but it‘s so damn musical that it‘s impossible to make things sound bad. I dig it! 

[GregWormald]

Thanks Geoffrey, I used the tilt on my Quad 33/303 and am looking forward to seeing what it does on a few recordings.

[Geoffrey Armstrong]

5 hours ago, GregWormald said:

Thanks Geoffrey, I used the tilt on my Quad 33/303 and am looking forward to seeing what it does on a few recordings.

Thanks Greg. I look forward to your findings.

[bogi]

Thanks Geoffrey, nice to try something like this.

I didn't find an easy and free usable case with VST3 (64bit) plugin, so I tried the VST2.4 32 bit one.

Question to you more skilled: Would bring the VST3 any sound improvement against VST2.4 or it is only about the effectivity of digital signal processing?

 

I tried it in Foobar2000 and it worked well, although I don't like much the Foobar2000 sound compared to HQPlayer. Thus the idea to find a way to convert files by applying tilt processing from the plugin.

 

Foobar2000 Converter is able to add VST2.4 plugins into DSP chain when converting files. So I was able to apply the plugin (conversion speed was 1.4x against listening speed), but the resulting track sounded bad with strong echo in the recording, which was not present when the plugin was applied during playback.

 

Then I tried JRiver 22 (I didn't bought newer since I'm using it only occasionally). I was able to add the VST 2.4 plugin and save the conversion result, but the unwanted echo was again here.

 

Then I searched for  command line tools which could be able to apply VST plugin to source file and to produce destination file. What I found is MrsWatson http://teragonaudio.com/MrsWatson.html

It really enables to apply one or more VST2.4 plugins on a source file and generate destination file and it didn't produce the unwanted echo with sTilt. Plugin parameters should be provided on command line or using special configuration file. The trouble was to find value ranges of the sTilt plugin parameters because it is not sufficiently documented in this regard. Some help appears here https://www.tb-software.com/TBProAudio/download/stilt_manual.pdf and here https://github.com/teragonaudio/MrsWatson (scroll down) but it is not quite sufficient when you need to enter parameter values numerically. But I figured it out (on Windows, but in other OS it should work similarly):

1. Place mrswatson.exe and sTilt.dll in the same directory and change to that directory in Command Prompt

2. Get list of plugin parameters:
 

c:\Users\bogi>mrswatson --display-info --plugin sTilt
- 00000000 000000 MrsWatson version 0.9.8 initialized, build 20150122
- 00000000 000016 Plugin 'sTilt' is of type VST2.x
- 00000000 000016 Opening VST2.x plugin 'sTilt'
- 00000000 000172 Information for VST2.x plugin 'sTilt'
- 00000000 000172 Vendor: TBProAudio
- 00000000 000172 Version: 1410
- 00000000 000172 Unique ID: TBP6
- 00000000 000172 Plugin type: effect, category 1
- 00000000 000172 Version: 1410
- 00000000 000172 I/O: 2/2
- 00000000 000188 InitialDelay: 0 frames
- 00000000 000188 Parameters (13 total):
- 00000000 000188   0: 'AB' (1.000000)
- 00000000 000188   1: 'Quality' (0.000000)
- 00000000 000188   2: 'Slope' (0.500000)
- 00000000 000188   3: 'CFrequency' (0.045020)
- 00000000 000188   4: 'Clip Protection' (1.000000)
- 00000000 000188   5: 'Unit gain' (0.000000)
- 00000000 000188   6: 'Limit Spectrum' (0.000000)
- 00000000 000188   7: 'Mode' (0.000000)
- 00000000 000203   8: 'Monitor' (0.000000)
- 00000000 000203   9: 'Mix' (1.000000)
- 00000000 000203   10: 'Gain' (0.500000)
- 00000000 000203   11: 'Toggle Tooltips' (1.000000)
- 00000000 000203   12: 'Filter Mode' (0.000000)
- 00000000 000203 Programs (2 total):
- 00000000 000203   0: 'Default'
- 00000000 000203   1: 'Full Reset'
- 00000000 000203 Current program: 'Default'
- 00000000 000203 Common canDo's:
- 00000000 000203   sendVstEvents: Don't know
- 00000000 000203   sendVstMidiEvent: Don't know
- 00000000 000203   receiveVstEvents: Don't know
- 00000000 000219   receiveVstMidiEvent: Don't know
- 00000000 000219   receiveVstTimeInfo: Yes
- 00000000 000219   offline: Don't know
- 00000000 000219   midiProgramNames: Don't know
- 00000000 000219   bypass: Yes
E 00000000 000219 Output source could not be opened, exiting

 

3. An example call - applying the plugin on 192k source file:
The parameter 1 (Quality) was set to 1 (maximum value) and param. 2 (Slope - that's the plugin's main knob) is set to 0.4 (40) of range 0 - 1 (0 - 100):

c:\Users\bogi>mrswatson --parameter 1,1 --parameter 2,0.4 --plugin sTilt --input "R:\RAMDISK\Vrchol sezony.wav" --output out.wav
- 00000000 000000 MrsWatson version 0.9.8 initialized, build 20150122
- 00000000 000000 Setting 2 channels
- 00000000 000000 Setting sample rate to 192000Hz
- 00000000 000000 Plugin 'sTilt' is of type VST2.x
- 00000000 000015 Opening VST2.x plugin 'sTilt'
- 00000000 000203 Set parameter 1 on plugin 'sTilt' to 1.000000 (Max)
- 00000000 000203 Set parameter 2 on plugin 'sTilt' to 0.400000 (40)
- 00000000 000234 Starting processing input source
- 37329408 031953 Total processing time 31sec, approximate breakdown:
- 37329408 031953   MrsWatson Initialization: 230ms (0.7%)
- 37329408 031953   MrsWatson Input Source: 921ms (2.9%)
- 37329408 031953   MrsWatson Output Source: 1sec (4.7%)
- 37329408 031953   sTilt Audio Processing: 29sec (91.6%)
- 37329408 031969   sTilt MIDI Processing: 0ms (0.0%)
- 37329408 031969 Read 18664449 frames from R:\RAMDISK\--02 - Vrchol sezony.wav
- 37329408 031969 Wrote 37329408 frames to out.wav
- 37329408 031969 Shutting down
- 37329408 031969 Closing plugin 'sTilt'
- 37329408 031984 Goodbye!

 

This way also other than sTilt plugin could be applied in batch to set of files, for example a music album.

Do you know other free solutions to batch convert a music album files with any VST plugin?

[Geoffrey Armstrong]

That’s very interesting bogi. I had no idea it was even possible to apply a vst plug-in as a command line tool. In that case I should be able to apply it to a file before sending on to HQPlayer. I’ve already experimented along these lines and can confirm the result of sending a pre-tilted file onto HQPlayer is far superior to my ears than using any other audio engine.

 

I started to try out the Tokyo Dawn plug-in recommended by copy_of_A this morning. I’m still evaluating it; but I think it’s a little better than sTilt. I wonder if that plug-in could also be used in a command line fashion via the method you’ve detailed here?

 

I also experienced the “echo” effect you described on Windows; but as previously mentioned, not on Mac.

 

That’s the sound of latency, I believe. The Tokyo dawn plug-in may not suffer from this.

[Geoffrey Armstrong]

I’ll just qualify what I said there a bit further, because whilst HQPlayer is definitely my preferred player, Audirvana is also very satisfactory to listen to and J River as well as Foobar can give more than acceptable results.

 

Let me put it this way; I am driven first and foremost by the music. So if a recording that I want to listen to needs a little Tilt correction and I could only do that with, say Audirvana, I would rather listen to the corrected file through Audirvana than the non-corrected version through HQPlayer.

 

Audirvana also has support for Qobuz and Tidal, of course, so incoming streams from those services can also be Tilt corrected.

 

The holy grail for me though is to send that corrected version to HQPlayer for further DSP up-sampling and conversion to DSD, as long as there are no serious latency issues leading up to the final output.

 

If you use HQPEmbedded, you should be able to achieve this by sending the Tilt corrected result from Audirvana, acting as the VST/AU host and UPnP server to HQPEmbedded.

 

Audirvana running on a separate machine should also avoid any latency issues as the pre-prosessed file url is being sent onto the machine running the UPnP renderer. So this is a two box solution.

 

The file served from Audirvana via UPnP should incorporate the effects of any VST/AU plug-ins. The same should hold true when using J River or Foobar as UPnP servers, I believe.

 

HQPEmbedded is a Linux only solution at the moment though.

 

I’m looking to feed a Tilt corrected output file to HQPlayer desktop on either Windows or Mac.

 

 

[bogi]

The TDR SlickEQ Mastering plugin works in Foobar2000 well, no issues. It does not produce unwanted echo in Foobar2000 converter, which can be used also in command line mode. It seems to me that the tilt control can be easiest achieved by enabling the bigger middle knob and setting here the Brightness level.

I don't know how to set parameters in command line of MrsWatson to correspond to the middle Brightness knob setting:

 

c:\Users\bogi>mrswatson --display-info --plugin "TDR SlickEQ M"
- 00000000 000000 MrsWatson version 0.9.8 initialized, build 20150122
- 00000000 000000 Plugin 'TDR SlickEQ M' is of type VST2.x
- 00000000 000016 Opening VST2.x plugin 'TDR SlickEQ M'
- 00000000 000188 Information for VST2.x plugin 'TDR SlickEQ M'
- 00000000 000188 Vendor: Tokyo Dawn Labs
- 00000000 000188 Version: 202
- 00000000 000188 Unique ID: Td6a
- 00000000 000188 Plugin type: effect, category 4
- 00000000 000188 Version: 202
- 00000000 000188 I/O: 2/2
- 00000000 000188 InitialDelay: 0 frames
- 00000000 000188 Parameters (81 total):
- 00000000 000203   0: 'Band 1 Active' (1.000000)
- 00000000 000203   1: 'Band 1 Shape' (1.000000)
- 00000000 000203   2: 'Band 1 Freq' (0.334605)
- 00000000 000203   3: 'Band 1 Gain' (0.500000)
- 00000000 000203   4: 'Band 1 Q' (0.500000)
- 00000000 000203   5: 'Band 1 Stereo W' (0.500000)
- 00000000 000203   6: 'Band 1 Stereo B' (0.500000)
- 00000000 000203   7: 'Band 2 Active' (1.000000)
- 00000000 000203   8: 'Band 2 Shape' (0.000000)
- 00000000 000219   9: 'Band 2 Freq' (0.399849)
- 00000000 000219   10: 'Band 2 Gain' (0.500000)
- 00000000 000219   11: 'Band 2 Q' (0.500000)
- 00000000 000219   12: 'Band 2 Stereo W' (0.500000)
- 00000000 000219   13: 'Band 2 Stereo B' (0.500000)
- 00000000 000219   14: 'Band 3 Active' (1.000000)
- 00000000 000219   15: 'Band 3 Shape' (0.000000)
- 00000000 000219   16: 'Band 3 Freq' (0.658826)
- 00000000 000219   17: 'Band 3 Gain' (0.500000)
- 00000000 000219   18: 'Band 3 Q' (0.500000)
- 00000000 000235   19: 'Band 3 Stereo W' (0.500000)
- 00000000 000235   20: 'Band 3 Stereo B' (0.500000)
- 00000000 000235   21: 'Band 4 Active' (1.000000)
- 00000000 000235   22: 'Band 4 Shape' (0.000000)
- 00000000 000235   23: 'Band 4 Freq' (0.338175)
- 00000000 000235   24: 'Band 4 Gain' (0.500000)
- 00000000 000235   25: 'Band 4 Q' (0.500000)
- 00000000 000235   26: 'Band 4 Stereo W' (0.500000)
- 00000000 000235   27: 'Band 4 Stereo B' (0.500000)
- 00000000 000235   28: 'Band 5 Active' (1.000000)
- 00000000 000235   29: 'Band 5 Shape' (0.000000)
- 00000000 000250   30: 'Band 5 Freq' (0.565412)
- 00000000 000250   31: 'Band 5 Gain' (0.500000)
- 00000000 000250   32: 'Band 5 Q' (0.500000)
- 00000000 000250   33: 'Band 5 Stereo W' (0.500000)
- 00000000 000250   34: 'Band 5 Stereo B' (0.500000)
- 00000000 000250   35: 'Band 6 Active' (1.000000)
- 00000000 000250   36: 'Band 6 Shape' (1.000000)
- 00000000 000250   37: 'Band 6 Freq' (0.641939)
- 00000000 000250   38: 'Band 6 Gain' (0.500000)
- 00000000 000250   39: 'Band 6 Q' (0.500000)
- 00000000 000266   40: 'Band 6 Stereo W' (0.500000)
- 00000000 000266   41: 'Band 6 Stereo B' (0.500000)
- 00000000 000266   42: 'HP Active' (0.000000)
- 00000000 000266   43: 'HP Steepness' (0.666667)
- 00000000 000266   44: 'HP Freq' (0.000000)
- 00000000 000266   45: 'Elliptical Filt' (0.000000)
- 00000000 000266   46: 'Elliptical Filt' (0.647640)
- 00000000 000266   47: 'LP Active' (0.000000)
- 00000000 000266   48: 'LP Steepness' (0.666667)
- 00000000 000266   49: 'LP Freq' (0.872417)
- 00000000 000266   50: 'Center Filter A' (0.000000)
- 00000000 000266   51: 'Center Filter T' (0.000000)
- 00000000 000281   52: 'Center Filter A' (0.500000)
- 00000000 000281   53: 'LF Sat Active' (0.000000)
- 00000000 000281   54: 'LF Sat Amount' (0.500000)
- 00000000 000281   55: 'HF Sat Active' (0.000000)
- 00000000 000281   56: 'HF Sat Amount' (0.500000)
- 00000000 000281   57: 'Output Gain' (0.500000)
- 00000000 000281   58: 'Output Width' (0.500000)
- 00000000 000281   59: 'Output Balance' (0.500000)
- 00000000 000281   60: 'EQ Range' (0.750000)
- 00000000 000281   61: 'EQ Range Mode' (1.000000)
- 00000000 000297   62: 'Auto Gain' (1.000000)
- 00000000 000297   63: 'Bypass' (0.000000)
- 00000000 000297   64: 'Mode' (0.166667)
- 00000000 000297   65: 'Quality' (0.000000)
- 00000000 000297   66: 'Display Scale' (0.000000)
- 00000000 000297   67: 'Analyzer Scale' (0.000000)
- 00000000 000297   68: 'Show Freq Magni' (0.000000)
- 00000000 000297   69: 'Analyzer Source' (0.000000)
- 00000000 000297   70: 'Use New Auto Ga' (1.000000)
- 00000000 000297   71: 'Nonlinearity Mo' (1.000000)
- 00000000 000313   72: 'LF Exciter Acti' (0.000000)
- 00000000 000313   73: 'LF Exciter Amou' (0.000000)
- 00000000 000313   74: 'LF Exciter Freq' (0.246373)
- 00000000 000313   75: 'HF Exciter Acti' (0.000000)
- 00000000 000313   76: 'HF Exciter Amou' (0.000000)
- 00000000 000313   77: 'HF Exciter Freq' (0.557886)
- 00000000 000313   78: 'Analyzer Stereo' (0.000000)
- 00000000 000313   79: 'Mixed Phase' (1.000000)
- 00000000 000313   80: 'Fixed Latency M' (0.000000)
- 00000000 000313 Programs (31 total):
- 00000000 000313   0: 'Default'
- 00000000 000328   1: 'Mastering Starting Poin'
- 00000000 000328   2: 'Tight Control'
- 00000000 000328   3: 'Soft Bells'
- 00000000 000328   4: 'Boiler Plate'
- 00000000 000328   5: '4 Band Party'
- 00000000 000328   6: 'FX: Low Frequency Monoi'
- 00000000 000328   7: 'FX: Low Frequency Excit'
- 00000000 000328   8: 'FX: High Frequency Exci'
- 00000000 000328   9: 'FX: Stereo Smear'
- 00000000 000328   10: 'FX: Sweet Telephone'
- 00000000 000328   11: 'USER001'
- 00000000 000344   12: 'USER002'
- 00000000 000344   13: 'USER003'
- 00000000 000344   14: 'USER004'
- 00000000 000344   15: 'USER005'
- 00000000 000344   16: 'USER006'
- 00000000 000344   17: 'USER007'
- 00000000 000344   18: 'USER008'
- 00000000 000344   19: 'USER009'
- 00000000 000344   20: 'USER010'
- 00000000 000344   21: 'USER011'
- 00000000 000360   22: 'USER012'
- 00000000 000360   23: 'USER013'
- 00000000 000360   24: 'USER014'
- 00000000 000360   25: 'USER015'
- 00000000 000360   26: 'USER016'
- 00000000 000360   27: 'USER017'
- 00000000 000360   28: 'USER018'
- 00000000 000360   29: 'USER019'
- 00000000 000360   30: 'USER020'
- 00000000 000360 Current program: 'Default'
- 00000000 000360 Common canDo's:
- 00000000 000375   sendVstEvents: No
- 00000000 000375   sendVstMidiEvent: No
- 00000000 000375   receiveVstEvents: No
- 00000000 000375   receiveVstMidiEvent: No
- 00000000 000375   receiveVstTimeInfo: Yes
- 00000000 000375   offline: Don't know
- 00000000 000391   midiProgramNames: Don't know
- 00000000 000391   bypass: Yes
E 00000000 000391 Output source could not be opened, exiting

 

[Geoffrey Armstrong]

I used the TDR Slick Eq GE Edition which has a tilt control in the bottom “drawer”. I settled on a center frequency of 825hz and turned the dial to the right to cool down, to the left to warm up. I’m still in the early stages of evaluating this and hope I’m getting it right. Also hope these settings can be accessed via the command line. Slick Eq is obviously more complicated as there are far more Eq controls and filters. sTilt has the advantage of being purely a tilt control.

[bogi]

The GE Edition parameters seem to be easier understandable

 

- 00000000 000219   18: 'Tilt Freq' (0.335792)
- 00000000 000219   19: 'Tilt Gain' (0.500000)
- 00000000 000219   20: 'Tilt Bypass' (0.000000)
- 00000000 000219   21: 'Tilt Type' (0.000000)

 

The default Tilt Freq value probably means those 650 Hz on 1/3 of the knob range.

The Tilt Gain range is probably from 0 to 1.

 

We may want to bypass unused parts by setting values to 1 (default is 0):

- 00000000 000219   27: 'LOW Bypass' (0.000000)
- 00000000 000234   28: 'MID Bypass' (0.000000)
- 00000000 000234   29: 'HIGH Bypass' (0.000000)

 

LP and HP filters seem to be bypassed by default

00000000 000203   12: 'LP Bypass' (1.000000)

00000000 000203   15: 'HP Bypass' (1.000000)

 

Tilt is enabled by default:

00000000 000219   20: 'Tilt Bypass' (0.000000)

 

It seems this did the job:

c:\Users\bogi>mrswatson --plugin "TDR VOS SlickEQ GE.dll" --parameter 19,0.4 --parameter 27,1 --parameter 28,1 --parameter 29,1 --input "R:\RAMDISK\--02 - Vrchol sezony.wav" --output out.wav
- 00000000 000000 MrsWatson version 0.9.8 initialized, build 20150122
- 00000000 000000 Setting 2 channels
- 00000000 000000 Setting sample rate to 192000Hz
- 00000000 000000 Plugin 'TDR VOS SlickEQ GE.dll' is of type VST2.x
- 00000000 000000 Opening VST2.x plugin 'TDR VOS SlickEQ GE.dll'
- 00000000 000047 Set parameter 19 on plugin 'TDR VOS SlickEQ GE.dll' to 0.400000 (-0.8 dB)
- 00000000 000047 Set parameter 27 on plugin 'TDR VOS SlickEQ GE.dll' to 1.000000 (On)
- 00000000 000047 Set parameter 28 on plugin 'TDR VOS SlickEQ GE.dll' to 1.000000 (On)
- 00000000 000047 Set parameter 29 on plugin 'TDR VOS SlickEQ GE.dll' to 1.000000 (On)
- 00000000 000063 Starting processing input source
- 37329408 003750 Total processing time 3sec, approximate breakdown:
- 37329408 003750   MrsWatson Initialization: 74ms (2.0%)
- 37329408 003750   MrsWatson Input Source: 904ms (24.1%)
- 37329408 003750   MrsWatson Output Source: 1sec (38.0%)
- 37329408 003750   TDR VOS SlickEQ GE.dll Audio Processing: 1sec (34.6%)
- 37329408 003750   TDR VOS SlickEQ GE.dll MIDI Processing: 0ms (0.0%)
- 37329408 003766 Read 18664449 frames from R:\RAMDISK\--02 - Vrchol sezony.wav
- 37329408 003766 Wrote 37329408 frames to out.wav
- 37329408 003766 Shutting down
- 37329408 003766 Closing plugin 'TDR VOS SlickEQ GE.dll'
- 37329408 003766 Goodbye!

 

[Geoffrey Armstrong]

Thanks. I think it should be possible to create a preset within a VST GUI Host, such as Audacity (one that supports VST 2.x), save the preset and apply that in a simple Mrs Watson command.

 

I thought I detected some slight additional reverb with J River on Windows hosting TDR Slick EQ GE. I was playing a WAV version of a file. Then I remembered when I played the same album earlier today with Audirvana on Mac, I was playing a flac version. So I loaded the flac version to J River on Windows and the added reverb was gone. Sounded much closer to Audirvana on Mac. Obviously I kept the VST settings exactly the same on both platforms/players.

 

Perhaps flac fares better as the input file for processing by a VST plug-in as it's a smaller file size. Just guessing. It could also be a question of how well the host app handles latency as copy_of_a said in an earlier post. The OS may also be having an effect on the degree of latency.

 

If it manifests as slight added reverb it may be tolerable for many listeners. I also prefer without though.

[bogi]

It is possible to save a VST plugin preset into .FXP file. The MrsWatson manual on github contains:

 

Quote

 

To process a MIDI file through an instrument using an FXP preset to be loaded into the VST, you'd do something like this:

mrswatson --midi-file mysong.mid --output out.wav --plugin piano,soft.fxp

 

 

I found that the .fxp files can be easily exported in Audacity.

At first enable the plugin in Effect menu and open it from Effect menu.

Then set parameters you wish in GUI, see picture.

Then simply export .fxp from the menu icon in the left bottom corner of the plugin window.

 

 

The .fxp file is readable in a text editor but it is not a plain text format intended to be modified in editor.

 

Now I repeated the conversion using .fxp file:

c:\Users\bogi>mrswatson --plugin "TDR VOS SlickEQ GE",SlickEQ-GE.fxp --input "R:\RAMDISK\--02 - Vrchol sezony.wav" --output out.wav
- 00000000 000000 MrsWatson version 0.9.8 initialized, build 20150122
- 00000000 000000 Setting 2 channels
- 00000000 000000 Setting sample rate to 192000Hz
- 00000000 000016 Opening preset 'SlickEQ-GE.fxp' for plugin
- 00000000 000016 Plugin 'TDR VOS SlickEQ GE' is of type VST2.x
- 00000000 000016 Opening VST2.x plugin 'TDR VOS SlickEQ GE'
- 00000000 000047 Loaded preset 'Default' in plugin 'TDR VOS SlickEQ GE'
- 00000000 000094 Starting processing input source
- 37329408 003360 Total processing time 3sec, approximate breakdown:
- 37329408 003360   MrsWatson Initialization: 90ms (2.7%)
- 37329408 003375   MrsWatson Input Source: 910ms (27.1%)
- 37329408 003375   MrsWatson Output Source: 1sec (42.9%)
- 37329408 003375   TDR VOS SlickEQ GE Audio Processing: 872ms (26.0%)
- 37329408 003375   TDR VOS SlickEQ GE MIDI Processing: 0ms (0.0%)
- 37329408 003375 Read 18664449 frames from R:\RAMDISK\--02 - Vrchol sezony.wav
- 37329408 003375 Wrote 37329408 frames to out.wav
- 37329408 003375 Shutting down
- 37329408 003391 Closing plugin 'TDR VOS SlickEQ GE'
- 37329408 003391 Goodbye!

 

[Geoffrey Armstrong]

Thanks for the advice. I just tried something similar using Audacity on Mac; but ended up with an xml preset file. Will try again tomorrow to produce an fxp and apply it to a file with MrsWatson (love that name), then load the file to HQPlayer.

 

If I can achieve that, I will incorporate it into the front-end apps to HQPlayer I'm currently developing.

[Geoffrey Armstrong]

As you rightly advised Bogi, it is a simple matter to export a preset in fxp format and apply it to a file. I did that this morning and produced two versions of the same file as an experiment.

 

This worked and I was able to play the warm and cool versions through HQPlayer.

 

The file I used didn't necessarily need any Tilt correction. This was just a test. One point to note is that mrswatson will only accept a wav format file as input and will only output in that format. So to use this programmatically would involve making a wav copy of any file that was not originally in wav format, then converting the file produced by mrswaston back to whatever the desired format may be.

 

Here is the command line I used on Mac. The names of the input and output files as well as the plug-in name need to be surrounded by double quotation marks as they contain spaces.

 

mrswatson64 -i "01 Circlesong 1.wav" --output "01 Circlesong 1-warm.wav" --plugin "TDR VOS SlickEQ GE",warm.fxp

 

This could open up a whole Pandoras box, as of course mrswatson allows any vst 2.x plugins to be applied to a music file via the command line, and thus also programmatically.

 

Folks working in the recording industry will typically chain several plug-ins together. The needs of audiophiles are obviously far less demanding and we wouldn't want to go OTT with them.

 

Obviously I feel this Tilt control is very worthwhile and some of us may also want to apply VST plug-ins for room correction and some other forms of equalization.

 

In any case I think we should definitely support and encourage Teragon audio in their efforts with MrsWatson. They have some great resources for developers too.

 

Hopefully we will see support from them for VST 3.x plug-ins and Audio Units in the future, which they say is planned.

 

http://teragonaudio.com/MrsWatson.html

 

I would also like to see support for more input and output file formats.

 

[bogi]

MrsWatson can in one run process a set of plugins. The manual shows this example:

mrswatson --input mysong.pcm --output out.pcm --plugin plugin1;plugin2

With fxp files it could look like:

mrswatson --input mysong.pcm --output out.pcm --plugin plugin1,plugin1.fxp;plugin2,plugin2.fxp

Only on Windows: MrsWatson VST2.4 plugin can be used with Foobar2000 converter together not only with other VST2.4 plugins, but also with DSP plugins in Foobar2000 format. A Foobar2000 converter preset, once configured in GUI, can be used from command line. An example line from my script:

 

"%FOOBAR%" /runcmd-playlist="convert/%CONVPRESET%"

 

Then it is possible for example to take FLAC or any other format Foobar2000 understands (DTS, APE, ...) from your media library, convert it to a PCM format HQPlayer likes (for example WAV - then HQPlayer does not need to decompress and can use more resources for it's hard work) and apply DSP plugins of your wish, for example multichannel to stereo, room correction (or some type of headphone crossfeed, or stereo to binaural solution),  and if needed for example tilt correction. The resulting WAV can be placed in RAM disk.

 

That's the base of the Windows solution I am using. I am selecting tracks (for example a music album) in Foobar2000 and press a button in Taskbar toolbar. That button press invokes my script, which gets the selection information from Foobar2000, performs conversion into RAMDISK based on a given Converter preset, then creates playlist file form the RAMDISK WAVs corresponding to the original track selection in Foobar2000 and sends that playlist into HQPlayer. All with one button press only. I have one button which substitutes the content in RAMDISK with new content and other button to add tracks to the existing ones. Till now I used another set of 2 buttons to include equalization (using Graphics Equalizer plugin of Foobar2000) for too bright recordings. It seems I will substitute that rather with tilt processing.

 

Just an idea what a script can do. The advantage of my solution is that 2 player engines are not running at once and HQPlayer plays WAVs (the PCM format it can process most effortlessly) from RAMDISK. I'm using SoftPerfect RAM disk on Windows.

[Geoffrey Armstrong]

Although using MrsWatson this way is an intriguing idea. I've run into two problems with it.

 

1/ I'm getting some pops and crackles in the output files.

2/ The test file I worked with has an extra 8 seconds of padded silence added by the process

 

Have you tried listening to any files processed by Mrs Watson?

 

 

[Geoffrey Armstrong]

I should make clear, that although I said "This worked and I was able to play the warm and cool versions through HQPlayer" I only loaded the file and listened to the first few seconds to check it would play. When I listened to the whole file I heard the snap, crackle and pop at various places in the file. This occurs whether I use the 64bit version of mrswatson or the 32bit. It also occurs when I create the presets on Mac or Windows with Audacity and whether I run mrswatson on Mac or Windows to create the files.

 

Unless the problem is with the creation of the presets in Audacity I believe the problem must be with mrswatson. The plug-in works well within the other hosts I've tried it with.