Stereo Tool

I suspect this was meant to be the link: http://pira.cz/rds/rtpclass.pdf

I made another few tests and "flat freq response" slider has no effect in digital mode, but in analog mode effect is very nice.

_________________
Urban Junglist

BETA 761-044

Windows 32 bit:
Windows stand alone: http://www.stereotool.com/download/ster ... 61-044.exe
Winamp DSP: http://www.stereotool.com/download/dsp_ ... 61-044.exe
VST 32bit: http://www.stereotool.com/download/vst_ ... 61-044.dll

Note #2: I have enabled the NAN checks again (which means that you might see error messages - if so please tell me - and the CPU load is a bit higher).

CHANGES:
- Improved Dequantizer by combining data from different block sizes.

How to use RT+ in RadioText
Put \+xx before and \- after an item
where xx is an RT+ type number (00-63).

Instead of a number you can also use the codes
WW or UR for an URL (type 29)
AR for Artist (type 04)
TI for Title (type 01)

So for example:
Now playing \+ArMadonna\- - \+TiLike a Virgin\-

You can use at most 2 codes. For now, codes are transmitted for ALL the texts, so this only works properly if you have a static RT text. I might change that later.


OLD CHANGES:
- Crash reported by Peter Kluyver - Not enough information. Occurred in 761-035, crash in ntdll.dll at 0x5b72E (exception 0xC0000005), not handled by Stereo Tool. No streaming via VLC used.
- Added second Multiband Compressor
- Added new Dequantizer filter to increase the bit depth of recordings.
- Added auto-restart behavior if buffers get too full or too empty.
- Added upto 10 seconds of audio delay for HD radio.
- Added option to show beta version number in title bar
- RDS: Added RT+ support.
- RDS: Fixed time zone 30 minute offset problem - viewtopic.php?f=6&t=5941&p=28209#p28209
- GUI: Improved display for bigger screens.
- GUI: Increased maximum waveform display from 3.2 to 5 seconds.
- GUI: Made line around Phase Delay in bottom display less bright.
- GUI: Bug fix: Composite clipper used but not registered warning wasn't shown.
- GUI: Save settings .ini when saving the settings to any file.
- GUI: Fixed delay of Delossifier in display.
- GUI: Fix 'Load default values' question for MB #bands change.
- Singleband Compressors: Don't allow Threshold values below 0.5 dB
- After-AGC Singleband Compressor: Added new Analog compressor mode with new limiter.
- Final Singleband Compressor: Added new Analog compressor mode with new limiter.
- Multiband Compressor: Added new Analog compressor mode with new limiter.
- Multiband Compressor: Band Link did not work in 3-band mode.
- Multiband Compressor: Fix frequency band shift (10%, since 7.51 or 7.60).
- Multiband Compressor: Fixed a filtering issue when using Density Drive > 1.00
- Improved CPU performance a bit, on multiple places (clipper, GUI).
- Small performance optimizations.
- Multipath Clipper: Made it stricter (use Smoothing mode)
- Bug fix: Saving FM sound card issue in VST: viewtopic.php?f=2&t=6027&p=29066#p29066
- ProppFrexx integration: Added RDS control via VST (canDo, vendorSpecific)

OLD CHANGES - NEW COMPRESSOR
- Make Shape correction values closer to 1 at lower Attack/Release time settings. That will also restore the sound for existing presets but keep behavior ok for AGC-like settings.
- Maybe add downward expander -> Not now I think... - no, should still be a separate step (so current noise gate before AGC is better).
- Made Auto option for auto-Shape sliders (Extreme Tweaker mode). Turn off to get old behavior.
- Check Auto Shape effect on audio (!) -> Result: It has some effect and it can sound slightly worse (fixable by reducing release speed, in the preset that I tested so far, which makes sense because without Shape there's something like release hold happening.).
- Freeze/Gate now also available for all bands combined.
- Moe as much code as possible out of templated functions to reduce executable size and compile time
- GUI: Decide which settings are visible in which mode
- Stop moving down way too far - new compressor, fast attack, slow release, mainly in RMS mode. It goes down way too far, then moves up slowly, then when it gets 'hit' again moves down way too far again. Reason: Attack doesn't stop after the loud sound stops. Need to talk to Joop again... --> EDIT: Solution is higher Attack and Release shape!!! More testing needed!
- Limiter crash on older CPU's fixed (previous beta)
- Optimized Limiter code for CPU usage
- Add Density parameter for each MB band
- Fixed another GUI bug
- Fixed Exception that several people reported when changing Gate Freeze settings (was caused by a GUI issue)
- Improved gate GUI display
- Improve limiter GUI display
- Optimize gating (turn off if not needed)
- Made 'limit before compress' again, switchable
- Add Gate for new compressor release (also GUI)
- Simplify interface (remove Oil Speed Adjustment sliders etc, move some things to Extreme Tweaker mode)
- Adjust limiter level for Band Mix
- Make Shape value small for high speeds - TO BE DONE: Test if the border value is chosen correctly (shape effect is deminished below 40 ms, reduced to '1' for 0 ms; not sure if 1 is low enough?)
- Fixed some more GUI crashes. They should all be fixed now...
- Check if audio is still the same as in older versions (!)
- Set MB2 default settings ok for NAB demoing (very fast)
- Fixed some (but not yet all!) GUI crashes
- GUI crashes from previous beta's are probably fixed (?)
- Limiter behavior visible in meters
- Fixed MB ITU mode.
- Fixed Median display in Analog mode
- Made final SB compressors switchable, parameters are also shared between analog and digital now.
- Added limiters to singleband and multiband Analog compressors. Warning: Meters are not yet visible in the GUI! Which makes it kinda difficult to setup (you can use Monitor to see the output of a single band and then adjust the limiter setting).
- I fixed something in the GUI but I had a situation where I couldn't start it earlier tonight so it's probably not completely fixed yet.
- Some settings from the 2nd MB overruled the 1st, causing all kinds of effects. That has been fixed now, and I verified that the audio of Digital mode in the 1st MB is (nearly) identical to that of BETA761-016 (nearly because in 016 the frequencies still were 10% off). New quickly put together preset (NICE2) attached.
- Optimized CPU load. CPU load now strongly depends on the settings, for the settings that I posted earlier it's now less than half that of Digital mode! On my system it went down (single core mode, CPU load on that core) from 12.3% to 2.5%! (Again, depending on settings). Digital uses 9.1%.
- Merged more parameters (Peak vs RMS setting) with old ("digital") compressor.
- Slower Release speed in Analog compared to Digital
- Removed Shape Enable settings.
- Fixed Knee in Analog mode
- Fixed Aggressiveness slider behavior for Analog (!), made it behave the same as Digital mode (acts only on release speeds).
- Merged more parameters (Attack Time, Release Time) with old ("digital") compressor. Adjusted numbers in Digital compressor to match the (correctly calculated) behavior in Analog mode.
- Merged parameters (threshold, knee, ratio, feedback, channel separation, band mix) with old ("digital") compressor, I noticed that I was switching back and forth all the time to copy settings, so this seems easier. Threshold is now also displayed in dB's.
Made Multiband version of the new compressor! Choose 'Analog' mode to switch to the new one. See attached preset for an example!
- Made new Traditional Peak mode which behaves more like normal Peak.
- Added Lookahead, removes huge spikes at start of attack for fast attack times and reduces IMD.
- Fixed Ratio behavior
- Fixed 'Density Drive' slider in Analog mode
- Fixed no audio with certain combinations of filters
- Added lowpass filter to control signal to smooth out bad effects and remove distortion
- Added ITU-1770 mode to reduce pumping and improve level control
- Added Channel separation
- Rewrote all code
- Fixed Ratio
- Channels are not processed separately (wasn't the case before, and that was a bug)
- Fixed attenuation display
- Attack and Release times must be set to half the numbers of the previous beta's. Oil change speed must be doubled.
- Restored old Shape behavior from version 009
- Replaced 'Nonlinear Release' by 'Exponential Release', which works better and only needs 1 parameter. Use this to go up FASTER if you're down DEEPER.
- Added Ratio
- Added Feedback mode
- Fixed out-of-control behavior for certain extreme settings
- Fixed some more weird behavior things if the difference between Attack and Release is REALLY huge (behavior was bumpy and caused clicking sounds).
- Known issue: both Shape On settings must be enabled. I will probably remove them in the next build anyway so I'm not going to fix it.
Fixed a lot of bugs, will type the details here later. Oil Temperature is now in % and didn't work properly before. PLEASE LOAD THE DEFAULT SETTINGS for the new compressor or it will sound horrible.
Changed Attack Time, Release Time, Attack Speed Threshold, Release Speed Threshold to understandable values. Made 1.0 the 'standard' behavior for Shape.
- Added smoothing to oil temperature behavior for both attack and release.
Added Oil temperature behavior which was missing from my automatic (hydraulic) door model. Will explain below...
- Added another Peak mode slider in "New Singleband". The old one (from yesterday) works mainly on release, but in some cases the volume didn't go down fast enough on a loud sound. The new one works mainly on attack (higher value makes attack continue AFTER a peak.
- Changed Peak mode behavior in "New Singleband".
- New Singleband: Split Speed into Attack and Release Speed
- New Singleband: Made Attack and Release Speed Threshold relative to Speed - so you don't need to adjust it to change the shape.

WAITING FOR FEEDBACK 7.70:
- Wait for ProppFrexx feedback
- Wait for Matt Aaron's feedback about HD synchronization

TO DO 7.70:
- Also finally make resampler for normal output to fix 2 output on different sound cards problem. Move Resample setting from input to output. Let's ignore using 2 inputs for now...
- Look into EOS -> Not now.
- Run VTune and find something to optimize -> resampler
- Fix warnings
- Disable NAN checking
- Run Valgrind or something similar
- Optimize resampling (pop can be placed in loop, check can be removed)
- Make lookup table for sinc!
- Add presets
- Make BS412-like graph for levels?
- Change version number
- Release

TO DO 7.71:
- EQ with 'standard' sliders.

TO DO WITH LEIF:
- Clipper issues Leif / Wolf Stahl (probably the same thing)
- Stokkemask: try if I can move the data a few samples to make 2nd smoothing step unnecessary.

BETA 761-045

Windows 32 bit:
Windows stand alone: http://www.stereotool.com/download/ster ... 61-045.exe

Bug: I just found that I'm writing some logging to a file, that might cause crashes on some systems. Will be fixed in the next beta, probably tomorrow.

Note yet thoroughly tested, but this version contains a sample rate adjustment for Normal Output. For all of you who are sending the FM output and the Streaming output to a different sound card and are stuggling with buffers getting too full or too empty!!!


Hm, my todo list appears to be empty

Is this posible to Fm output ? I mean some monitor to show us actually sample rate ( let`s say 176120 hz) and we can adjust it according our PC (another line below it to show local PC sample rate) . We know that every PC has difference in clock . And if this is implemented, can we turn off the "Synchronize to output" option ?

The difference in sample rate is so small that I don't know what's more reliable, the sample rate that the sound card uses or what Windows shows. Also, resampling will still be needed because I have no way of guarranteeing that the speeds won't vary over time (in fact that's what the resampling fixes), and I also cannot set the sound card to another sample rate - so I would have to resample anyway.

For FM output I can't do this because it would change the pilot and RDS frequency, that's why I need to adjust the FM output to match the input, and then the normal output to match the FM output.

Yep , fully understand .
I just forgot for the Pilot and RDS will be shifted in frequency.You are right .

Todo list isn't empty anymore. Will do some performance optimizations in the next day (among others for the resampler that's used for synchronization).

To quote my 2 year old daughter "CITED!"

(She can't say excited)

_________________
I just use ST with liquidsoap to enhance a streaming radio station I like...

When I was reading about RT+ my understanding is that you can use one of the other modes 11a or 12a for example to send your format text, and then 3a to define what text is what.
I also know on RT that you have to alertnate between A and B in order to have the decoder cleanly switch between the 2. I do say now that i've implemented this method i do like it better on your end. I don't have a RT+ encoder besides ST, im trying to acquire one for this.

Right now from what i understand your sending your RT then using 3A to define where in the RT data we already send is the title/artist. That's brilliant.
Is it possible to allow the variables to be loaded from a text file.

Simian can output a file with raw text formatted to be decoded by ST.
So i can dynamically do this: Now playing \+ArMadonna\- - \+TiLike a Virgin\- in the output file.
It then reads as rt.txt I'd then like to be able to pull that into stereo tool with the command \+Ti\r"{driveletter}:\rt.txt"\- for example. Loading the RT+ from a file would be extremely helpful in interfacing with Simian. Or at least if you can make it optional.

Edit: I think my being sick is clouding my reading skills. But when i tried implementing the \+ \- from text files I ran into it broadcasting that over the RT. I'll try again with the latest release.