All times are UTC+02:00




Post new topic  Reply to topic  [ 15 posts ]  Go to page 1 2 Next
Author Message
 Post subject: Stereo Tool 10.61 BETA
PostPosted: Sat Jul 19, 2025 11:30 am 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11443
Hi all,

Latest version - Beta 10.61-007 (July 31 2025)

KNOWN ISSUE: Mac version does not have a valid code signature and might not work if you don't have internet access. Our old certificate expired and the new one we got doesn't work, apparently (according to ChatGPT, so take this with a whole spoon of salt) due to some issue on Apple's servers. No Mac version will be released at this point.

Downloads
Windows: Pi (ARM Linux) 64 bit: Pi (ARM Linux) 32 bit (BETA002): macOS (BETA002): Changes/bugfixes
  • Fixed WatchCat ZLIB1__.DLL dependency
  • libStereoTool ignored plugin output level setting
  • Disable matrix mode for 1 channel, check if it's ok then
  • Made preringing protection tighter again (was loosened in previous beta). And that's now the default behavior.

PRIO TODO's (build system)
  • Maybe support AI translations - but how? (external developer) BUSY
  • Create single persistent Python version of SNMP monitoring tool for Pi SD card image BUSY
  • MicroMPX monitoring tool
  • R128 live - look up specs
  • Test SDI
  • WatchCat: Add ffmpeg to installer, and remove need for scripting for that. Also make that a free option.
  • MicroMPX: Add simple UDP packet forwarder. (what about RIST?)
  • MicroMPX: Allow telling MicroMPX which ports to combine for the same stream (v5? maybe unique stream id in rtp?)
  • MicroMPX V5
  • Add EDI support! Remarks Peter van Ak... AAC block size 960 (FDK code patch), EDI headers, later EDI metadata via input.
  • ShoutCast V2
  • Disable chkdsk in boot, see my reply email from 3/Apr/2025 14:43. See also https://chatgpt.com/c/6800ad26-a3c0-800 ... a34fea7465 to prevent this from happening.
  • NEW FM CLIPPING IDEA: Pre-analyse highs loss and compensate for it, to allow more snappy highs though on FM
  • Top-of-hour ID in MicroMPX and ST
  • EAS in MicroMPX and ST, ask Josh
  • HLS, via ffmpeg, also on Pi/Mac
  • JSON interface documentation, plus documentation of UECP mode
  • Auto Pilot
  • Preset descriptions. Separate categories. Filtering for specific formats, maybe with checkboxes per preset to set those.
  • Low bitrate AAC processing (Norway etc)
    PRIO BUGS:
  • Mac signing
  • VLC input can loop old audio: https://forums.stereotool.com/viewtopic ... 864#p78864
  • Keyboard input issue in 10.70: https://forums.stereotool.com/viewtopic.php?t=35234).
  • Check: Bug report that channel linking is broken in 10.70: https://forums.stereotool.com/viewtopic.php?t=35227

TODO's (build system)
  • Report that there's no output selection in libStereoTool, so you always get pre-emphasized out: https://support.thimeo.com/scp/tickets. ... 1681#reply
  • MicroMPX: Lukasz reports selecting system clock unavailable in Linux builds
  • Mac signing
    -
  • Check email delivery issue
  • Renamed .so files must be patched with patchelf --set-soname libStereoTool_noX11_arm64.so libStereoTool_noX11_arm64.so
  • Suppress "A piece of audio was missing" message from constantly popping up.
  • Show red error text (PSTATICTEXT with SetError(true)) above Input 1 + 2 settings when backup is running. See SetBackup call.
  • MicroMPX: Decoder hangs when sending multiple RIST streams to the same port number, and then stopping the one it's listening to.
  • MicroMPX: Add very simple file player with pilot/RDS. Page with backup player settings (MP3/WAV) + RDS settings. 1 extra tab in web interface. Maybe add uploading, support very low speeds. Play from USB (D).
  • MicroMPX: Make both output channels separately adjustable
  • MicroMPX: Add option to add phase inversion
  • Mathijs: Check Stokkemask stereo effect, can we reduce that? BUSY
  • Attempt to remove mallocs in webserver code: ServerWorker::tryWebsocketHandshake, excludeWarnings
  • Attempt to reduce mallocs in preset saving code
  • Web server: Close connection if it is not used for a long time, or increase max connections. Someone has connection issues with too many windows open: https://support.thimeo.com/scp/tickets.php?id=11202
    -
  • include/shout/shout.h need to be generated from shout.h.in somehow (can be a sed command if needed)
  • Check lib-untracked origins
  • Where do we put the build scripts (pscp, tar, plink)
    -
  • AES67: Hiccups for someone using HD out with Dante, waiting for feedback. Finish smoothing code and hope for the best...
  • 10.51: ABE version does not have ABE colors or login accounts - Wait for feedback, override default colors (and maybe saving them)
  • Clara HW test
  • Optimization: Exchange latency for quality for STB


TODO's (not related to build system)
Tom:
  • Multiband load preset can make number of visible bands go up/down constantly
  • Full screen mode is saved to presets! Should not.
  • iPhone 15 misses labels in Chrome
  • DcWave add delay, maybe make it work same as Wave, because now it's displayed unsynchronized with everything else.
  • Usability: Maybe redesign how RDS now playing text inclusion works?
  • Usability; Show streaming data in GUI (artist, title, song)
  • Multi-instance: Tray icon should not show instance 0, and ! if there's any error
  • Multi-instance: ! in overview if something is wrong
  • Tab navigation should show abbreviated texts
  • Don't center titles in narrow mode
  • ST-Enterprise: Copy presets to other instances
  • Compile ST to webassembly?
  • More different meter colors.
  • SCStreamingOutput is now a combination of VLC and builtin streaming, messy.
  • Split presets into built-in and user
  • LADSPA (or LV2) plugin? For ffmpeg, among others. Or inside ALSA(?).
  • Load presets via JSON API
  • Arrow in tiny info, maybe add sample rate? Maybe add optional element if there is space to bandpass display.
  • Maybe add option to select what to see in meter bar
  • MacOS: GUI doesn't move unless the window size is changed (VST2, VST3, stand alone not tested)

Hans:
  • STXtreme extra instance 8-character code registration doesn't know the file name where to store the license key.
  • Check if UECP RT A/B flags are handled correctly. Mail Matt Paton 18 jun 8:50AM
  • MicroMPX: Jeroen: op jullie pi image kun je de NIC niet kiezen, alleen default of loop
  • Mathijs' feature for new preset! Needed for YourSafe. Auto EQ after AGC would be nice to test
  • Export EXPORT.GUI list including parameter enum names, and include in libStereoTool zip file.
  • MicroMPX: Show warning if MicroMPX is enabled in ARM build at 44.1 kHz (doesn't work).
  • MicroMPX: Allow telling MicroMPX which ports to combine for the same stream
  • MicroMPX: Add PWARNINGs for errors that need a popup, for example entering wrong license key
  • Build: Test: Stereo Tool Pi CPU usage
  • Build: Add LInux (intel/ARM) VST3 builds
  • Investigate glitches when using Generic plugin in high CPU mode (could be on the caller's end)
  • New compressor: Add pre-compression clipper to reduce effect of loud highs that will be masked by clipping/limiting
  • Check if median levels are correct, may be displayed at wrong level
  • SCInstance 848 mutex can hang (Mac, with insane CPU load)

Overall:
  • URGENT: Socket error in libShout, followed by not reconnecting.
  • Disable chkdsk in boot, see my reply email from 3/Apr/2025 14:43. See also https://chatgpt.com/c/6800ad26-a3c0-800 ... a34fea7465 to prevent this from happening.
  • NEW FM CLIPPING IDEA: Pre-analyse highs loss and compensate for it, to allow more snappy highs though on FM
  • NEW COMPRESSOR IDEA: Maybe set a minimal speed, easier to use than speedup on drops. Nice curve into it. -> Probably not needed, issue was caused by Casper and that's fixed now.
  • Is some Auto EQ-like behavior in the new compressors useful to prevent too much sound adjustment? Or should we just make the bands wider? -> "Soft linking" fixed this. Maybe something is needed for R128?
  • Photos to travel site, and update it
  • Allow Auto EQ after AGC (or better: Let it output a sidechain with only weakening of levels) (Mathijs)
  • MultiVOC Jack version more than 2 channels
  • MicroMPX: On our SD card image, it doesn't show the NIC for ETH0 in the pulldown menu - why??? Works fine in other images.
  • Check if input affinities are set
  • AES67 Do we want to also support the system clock, including any offsets that are applied to it? Probably not?
  • File player for top-of-hour legal ID (for both MicroMPX and ST)
    -
  • UMPX TESTS F
  • MicroMPX v5: RTP
  • MicroMPX v5: RIST
  • MicroMPX v5: Improve quadrature encoding for very clean signals to avoid rounding errors in peak level detection.
  • MicroMPX v5: Add GPS timestamp to encoder library API so others can use our encoding
  • MicroMPX v5: Email joost 2
  • MicroMPX V5: Add partial keyframes for faster recovery without resending everything
  • MicroMPX v5: Re-check if 38.4 kHz is the ideal sample rate, since we never want audio above 16.4 kHz anyway!!!!!!!!!!!!!!. Can we use a frequency such that no LPF is needed (eg 33 kHz)?
  • MicroMPX v5: Check GPS calculation error (inverted atan, //TODO in GpsSyncedSoundcardOutput line 494)
  • MicroMPX v5: Maybe encode highest audio frequency
  • MicroMPX: Better way to tell it which ports contain the same signal (currently must use same port number)
  • Streaming input on Pi.
  • Directly after new build system: Put PhoneBooster in VST plugins
  • Directly after new build system: Add VST Linux support
  • Directly after new build system: Add AU support
  • Check reported "Socket error" in output stream. Apparently new in 10.50 but I don't know if that's true. https://support.thimeo.com/scp/tickets.php?id=9990
  • Streaming blocks if now playing file drive is unreachable (10.21): https://forums.stereotool.com/viewtopic.php?t=33962
  • Linux: "Full screen" mode - after closing the software opens without a GUI. Workaround added, real fix needed later.
  • Potential AAC streaming issue on Pi: https://forums.stereotool.com/viewtopic.php?p=77926
  • MicroMPX monitoring tool
  • Create C++ version of SNMP monitoring tool for Pi SD card image
  • K Black Pi RDS2 issue - fine on Windows, potentially ARM code error
  • Bug: HD output level is reduced when using BS412. Has been this way forever...
  • Add libVLC support to Linux/Mac builds (for OS parity)
  • Pi image: Check Pi memory patch
  • Check VST3 crash on Mac TOM
  • Check VST in Auditon on Mac TOM
  • MicroMPX multicast issue directly after startup
  • MicroMPX: Export the control signal that would control when the backup player becomes active to the JSON and library outputs so GPO's can be coupled to it.
  • Web interface KoekError display: Add last error time to web interface. But NOT for MicroMPX, because there the time is already part of the error message - or change that, but identical messages must still be logged separately in that case.
  • Output streaming via libVLC is broken; it does not detect when there's no connection, and does not detect connection loss in VLCGetBuffers. So it does not properly reconnect after an error.
    -
  • JSON interface documentation
  • Check Hard Limit behavior (Bojcha)
  • Add + button for streams, to make page less crowded
  • MicroMPX: When signal comes back, backup player stops and silence is played, can be 14 seconds if that's the buffer size.
  • Vorbis: Add sample rate; if we're running at 192 kHz that's also the stream sample rate, should be 48 in that case.
  • MicroMPX Add way to set IP address - but how??
  • Re-enable MPX demodulator settings in Stereo Tool (MPX Tool-like)
  • STXtreme 5 FM instances doesn't work if more than 2 are set to latency 128. Do more measurements and figure out what the limits are.
  • STXtreme: Implement smart glitch detection INTERNALLY so we don't need to run external tests all the time. Maybe we can already see it!
  • STXtreme multi-instance hiccups system Matt - why?
  • STXtreme: Audio seems fine even though buffer display shows that the buffer is very low.
  • MicroMPX: Add/modify JSON parameter that shows total used bitrate to include RIST, plus number of RIST clients.
    -
  • MicroMPX: Add way to set IP address of hardware
  • Multple streams: Support different metadata??
    -
  • MicroMPX decoder: RIST: Figure out way to get rid of RIST delay, or how to sync with non-RIST and how to specify RIST vs non-RIST delays - LATER
  • MicroMPX decoder: RIST: How to handle RIST delay in combination with GPS
  • Would it be possible to make an R128 measurement there, where I can set the measurement period for a sliding window between 1 and about 5 minutes?
  • MicroMPX: Reduce RIST overhead
  • Clipper + Hard Limit makes the spectrum dirty, mainly for pure sine waves - but why?
  • MicroMPX library: umpxDecoderDecryptAnalysePacket may try to decode a packet while internally a packet is already being decoded, causing threading issues.
  • MillingBroadcast's MicroMPX speed issue with encoder in Stereo Tool on Linux: https://forums.stereotool.com/viewtopic ... 454#p77454 - Waiting for feedback
  • Silence detector in ST with email/SMS?
  • Bojcha (via Discord personal chat with images): Clipper (measured HD out, no pre-emphasis) has HOLE between 5kHz and 13kHz about 0.4dB. Only way to fix it is to set Highs clipping from +1.51dB (default) to +0.95dB. He says it sounds better too... (less gaps due to highs, not less highs)
  • MicroMPX: Add GPS support for decoder without NMEA. That's how others do it - they just use the timestamp in the packet as the time of the closest 1PPS pulse. Will simplify the hardware!
  • MicroMPX: Silence detection in MPX is used to switch between studios! So reporting levels is really necessary.
  • Mathijs: Als je de AGC in Matrix mode gebruikt (trouwens voor MB hetzelfde denk ik), dan wil je eigenlijk channel linking gebruiken om te voorkomen dat L-R heel extreem bijgetrokken wordt tov L+R. Alleen: als dan de hoeveelheid stereo content flink toeneemt kan het zijn dat L-R de L+R mee omlaag trekt, en dat wil ik eigenlijk niet. Dat gebeurt sneller als je ook nog die "Stereo boost when AGC is below threshold" setting gebruikt. Zou fijn zijn als die channel link maar één kant op kan werken; dus dat L+R altijd vrij kan bewegen en dat L-R daar aan vast gelinkt hangt
  • Vinyl bass rumble removal
  • 32 bit CortexA53 old GCC version for GA
    -
  • Can we use USE_COMPACTED_STORAGE by default? Would save a lot of memory.
  • MicroMPX doesn't show EF mode error in Stereo Tool plugin version.
  • < 150 Hz mono to remove rumbly on vinyl.
  • Insert points: Via the sound card to outside and back? Or via VST's. To enable adding effects.
  • Vinyl click detection and removal.
  • Report that VST3 on Mac crashes on startup
  • Sound cards: Add VLC to input 2 for backup - but that requires a fix for when the stream drops out.
    -
  • Website: Write description of how to set up ST + MicroMPX, seems to be difficult for some people to find.
  • Add loading presets via JSON (nrk)
    -
  • Finish US pricing, update prices.
  • MicroMPX overview app
    -
  • When sound card/stream recovers after error, clarify current message which appears to indicate that it's not streaming at all.
  • Remove Stokkemask Clipper?
  • Remove Simple Clipper from STXtreme
  • Phase Delay isn't disabled at low latency settings, causing extra delay.
  • MicroMPX GPS: Extreme amounts of logging when GPS data doesn't come in due to bad antenna. Causes disk to fill up in a day, so the logging cleanup also doesn't help.
  • LiveWire 48 kHz FM output doesn't work with ASIO
  • Streaming: Add FLAC support
  • Test SDI
  • GUI: Sound card buffer slider behavior
  • Verify that inputs 1 and 2 have separate tilt corrections.
  • Analyzer
    -
  • Mac AU plugin should be easy to make now
  • Create small program to send UDP or TCP commands for UECP. Maybe even real UECP commands... Could be the start of a new product :) For now, ncat works (example: echo PS=Hello | ncat 127.0.0.1 4001), see https://nmap.org/download.html#windows
  • I/O: Add file input (and output?) so we can use files as backup material. Maybe both let it run all the time or make something like in MicroMPX that only plays when needed and pauses in between.
  • I/O: AES67
  • MicroMPX: Make readSlidingWindow smarter; only copy if necessary (lazy).
  • MicroMPX V5: Add L/R peak levels to encoder and decoder libraries.
  • MicroMPX V5: Simplify header structure
  • Add test tones to check phase and level issues for FM, like in old calibration.
  • Fix stepping of sound card buffer sizes (without control/shift should be 1, not .1 or 10).
  • Add RDS2 slide shows
  • Check report: Generic plugin crashes on mono input
  • GUI: Don't show 2nd channel for mono processing
    -
  • Delay issue in MPX display; not synchronized with the audio.
  • Overview screen password protection: How should that work?
  • Overview screen: Error should be shown here.
  • Mono bass for HD cannot be enabled without enabling the other "Reduce stereo" settings as well; needs to be split off just like for FM.
  • MicroMPX: Somehow detect and report number of incoming streams
  • Do something to draw graphs even without audio
  • Pragma messages: Check and remove where possible
  • Check BIMP meter bar; dynamics section MB on/off removes top bar
  • MicroMPX: Add error checking for IP addresses, don't open if the address isn't complete. (Maybe wait for ENTER? - Web interface).
  • Improve sound card error messages on STXtreme
  • MicroMPX: “umpxDecoderGetStreamFeedback()”: the signal lost counter in umpxDecoderStreamInfo seems to be stuck or sometimes not to count up when there is no input (with pilot generation active, internal sync turned off). I can’t rule out the problem is on our side, but it was working before and I think we did not change anything handling the umpxDecoderStreamInfo.
  • Streaming must read data from files from a separate thread, https://forums.stereotool.com/viewtopic.php?p=75718
  • Can we add an automatic gate for Blue Monday? Detecteer harde tikken die even hard zijn als input maar toch steeds harder worden en te hard zijn.
  • STAMP license error message is only visible in the native GUI (which doesn't exist); how do we show it in the correct GUI (the one in which the button was clicked)? Same question for overview page.
  • DHD: Remove logging, probably.
  • DHD: Add FM Professional to purchase options on the site.
  • Fix Declipper issues reported in https://forums.stereotool.com/viewtopic ... 4&start=30 and on Discord
  • Restart when sound card does not work is not working anymore
  • .st.gui_override does not survive firmware updates in SD card image; needs to look in root directory.
  • AES67: If a LiveWire port is selected under "Device" but nothing is detected there, that port number overwrites the active port field on preset loading. The pulldown itself cannot be changed if nothing is detected.
    -
  • Advanced Clipper: MAYBE make a filter that limits how much bass mids and highs protection can remove at the most. Can be done later.
  • AGC: Most likely we'll also need to change the release behavior compared to the compressors.
  • AGC: Bojcha: Some other stuf really needed, - startup level - return to platform - gate after agc (please) - some sort of compressor or some protection for bursts like on old AGC - why freeze is not fully freeze (there is delay)
  • MicroMPX: Run as service?
  • Loading a Stereo Tool preset on PiMP also sets things like whitelist, blocking access.
  • Make RDS/RDBS selectable or something, independent of 50/75 us pre-emhpasis.
    -
  • New compressor: Matrix mode: Limiters work on L+R and L-R, instead of L and R. That's probably bad.
  • MicroMPX: Allow RDS settings in library?
  • PEQ in MicroMPX FM output?
  • Sensitivity scrolling vs sliding
  • MicroMPX: Check pilot cleanness in decoder if pilot frequency is slightly off (19005 Hz).
  • Also, I notice when adjusting the buffer, there's no way to adjust by 1ms. There's 0.1ms (CTRL), 10ms (SHIFT), but it doesn't go by 1ms when no modifier is held. It would also be very handy to have the Synchronize with HD slider in between the HD and FM output cells instead of hidden in the More FM Output Settings.
  • Reset skin, and other things (basically anything that can be loaded?)
  • Pi: Make network settings changeable via web interface.
  • Startup delay on MicroMPX?
  • Update LiquidSoap links
  • mndxer1049: Better Natural Dynamics section with WB dynamic detection
  • mndxer1049: A very transparent, C-Level like new AGC design
  • mndxer1049: Auto Stereo Pilot
  • mndxer1049: Maybe: A button for putting all CPU settings at the maximum value and turning on all CPU intensive settings that only improve audio, but take quite a bit of CPU load. (ex. sliders like ‘Quality (increases CPU load)’ under DeClipper restoration, “Strictness (CPU)” and “Leif’s highs gap protection stages (CPU)’ under Clipper just to name a few of these ”CPU sliders” I’m referencing.)
  • Display bug in MPX display (weird spectrum glitches that occur occasionally)
  • Add streaming output to VST version?
  • In the audio input section I'd like to see options for speakers to be included without having the need to use virtual cables to route. I was thinking something similar to Rocket Broadcaster. There's input for mic and separate application source.
  • MicroMPX decoder: High frequency tilt correction is performed BEFORE data is collected for display, looks like overshoots.
  • Auto pilot
  • Redesign ND dynamics detection
  • Maybe add some frequency display to MPX display. https://forums.stereotool.com/viewtopic ... 16&t=33720
  • Add numeric displays to ITU and MPX levels in GUI https://forums.stereotool.com/viewtopic ... 16&t=33686
  • Insert/delete bands in multiband compressor, instead of (like now) just having to add/remove the last one when the number of bands changes.



    -
  • Non-phase linear: For example https://github.com/nxsEdson/Butterworth ... rworth.cpp
    -
  • STXtreme errors when reusing sound card
  • STXtreme set shared sample rate (instances can set different sample rates)
    -
  • Azimuth affects highs!!
    -
  • Ratio slider smarter steps (1.1, 1.2 etc) - MattL
  • Meters: add separate FM/HD output level displays.
  • New compressor and matrix mode?
  • Add 0.05 precision to parametric EQ and attack/release multipliers
  • RDS: Add settings for how long texts are shown or how fast to scroll for UECP inputs.
  • AES67 generic mode: Let people enter IP address directly.
    -
  • MicroMPX: Show temperature
  • MicroMPX: Show status per incoming port
  • Links: ID_GOTO must be a PLink that points somewhere.
  • Dynamic speeds: License detection only works after audio has been running, that's not ok.
  • When audio has run after a settings change, trigger a parsetChanged
  • Time offset on STXtreme, for RDS (also for ClockTime!)
  • New GUI: Support smaller monitors (smaller default size, especially for VST). Contact near05.
  • Export presets from web interface needs filtering, just like from native (so the code is there! Just reuse CopyPresetOm*).
  • Mac: VST2 sizes
  • Mac: gavinmcgrath: "Intel MacOS 11.7.3 in Wavelab 11.1: Blank screen." https://forums.stereotool.com/viewtopic ... 407#p73407
    -
  • Stand alone RDS decoder
    -
  • CPU usage of REFRESH_ENABLED is high, maybe use local array.
  • PDcWave delay not set
  • WatchCat: Keep track of when files have been processed and ignore them, but don't move them. (DropBox)
  • WatchCat: Can we include ffmpeg and lame?
  • Add latency wizard?
  • setFaultyValue for text boxes for wrong passwords for example - check with Tom how to use
  • Can we enable FM Calibration in "Ignore high frequencies" mode? Not easily
    -
  • Add check for running ST multiple times; make switchable if we want to allow it or ask if you really want to start if it already runs or something.
  • Matt: Pre-emphasize highs slider needs to get a clearer description! Something with avoid overshoots. Basically a sidechain EQ. "Add highs to sidechain".
  • Matt: UECP must be more easily accessible, too deep now.
  • Matt: FM+AM
  • New I/O framework: Add extra latency to plugin input so the input 2 to LQLL can be instantaneous.
  • Check Bojcha's good bass HPF settings (18200 Hz, ignore high frequencies, 48 kHz, 35 Hz HPF)


Quote:
Latest version - Beta 10.61-006 (July 30 2025)

KNOWN ISSUE: Mac version does not have a valid code signature and might not work if you don't have internet access. Our old certificate expired and the new one we got doesn't work, apparently (according to ChatGPT, so take this with a whole spoon of salt) due to some issue on Apple's servers. No Mac version will be released at this point.

Downloads
Windows: Pi (ARM Linux) 64 bit: Pi (ARM Linux) 32 bit (BETA002): macOS (BETA002): Changes/bugfixes
  • Processing "Legacy cleanup" caused limiter in MB to malfunction - checkbox added below "Processing overview" popup button.
  • Pre Amp wasn't disabed in Hear mode for early filters in chain (Declipper, Delossifier, Noise gate etc)
  • Casper mode: Added escape strength setting. Protects against "hanging" casper, which you might actually want for classical music, hence the setting.
  • Sudden Drop Protection: Added slider to make this wideband. Only activates if ALL bands agree that there's a drop.
  • Added visual color change between yellow and orange on input meter if it's playing backup input, to more clearly indicate that there's an issue.
  • Increased range for "Make dense audio louder", needed for some songs and doesn't appear to have any drawbacks.
Quote:
Latest version - Beta 10.61-004 (July 26 2025)

KNOWN ISSUE: Mac version does not have a valid code signature and might not work if you don't have internet access. Our old certificate expired and the new one we got doesn't work, apparently (according to ChatGPT, so take this with a whole spoon of salt) due to some issue on Apple's servers. No Mac version will be released at this point.

Downloads
Windows: Pi (ARM Linux) 64 bit: Pi (ARM Linux) 32 bit (BETA002): macOS (BETA002): Changes/bugfixes
  • Multiband: Improved horizontal band linking by removing inter-band dependencies, which caused noticeable wideband drops on certain loud sounds such as loud vocals in the previous 2 beta's. This is clearly better.
  • WatchCat: Removed libz1__.dll dependency that caused 2.0.17 to not start on some systems.
Quote:
Quote:


Top
   
PostPosted: Sat Jul 19, 2025 11:52 am 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11443
Multiband band linking ratio's
The new Adaptive multiband compressor mode supports infinite ratio operation without causing pumping or distortion. However, during real-world use, I noticed that something still sounded off: the audio was too aggressive and “in-your-face,” even though the compressors weren’t working hard.

At first, I thought the difference had to do with stereo widening. In infinite ratio matrix mode, the stereo difference (L–R) is also fully amplified, which can make the stereo image too wide. So, I added a ratio between L–R and L+R to tone that down (that's in 10.60). That helped solve some issues, but the main problem was still there.

In infinite ratio mode, each band is always pushed to the exact same output level, regardless of how much content is in that band. So if a certain frequency range is missing from the input (say it was cut off or filtered), that band won’t contain much audio — but the next band will try to compensate. And because of the infinite ratio, it’ll push that content all the way up, possibly boosting high-frequency remnants by as much as 6 dB. The result: a harsh, overly bright sound.

In lower ratio modes this still happens to some extent (which is necessary for consistency), but it's not annoying because the bands aren’t forced to the exact same level.

Previously, band linking had a hard limit. For example, if you have 6 bands, and the levels purely based on incoming audio would be -20, -20, -10, 0, 0, 0, and you link bands with 6 dB, you would get -20, -20, -14, -8, -2, 0. That helped to prevent harshness but didn’t solve the real issue.

Surprisingly, the real problem seems to be what happens before we hit that 6 dB linking limit. Basically, we always want to reduce the amount of adjustment somewhat. In the new behavior, instead of a hard linking threshold, each band now blends with the previous one using a soft ratio-based approach. This avoids abrupt shifts and gives a more natural feel. For example, using a 1:2 ratio, calculated on absolute (not dB) values, the example above would end up at around -20, -20, -16, -11, -7, -4.

This is actually better than traditional lower ratio behavior. With a real 1:2 ratio, we would get -10, -10, -5, 0, 0, 0. The real problem is band 3, which in this case would be pulled up (relatively) by 5 dB, vs 4 in our new design. But the bands above that would not be pulled up at all anymore, and in the new design, they are! So we still get much more consistency than with traditional lower ratio compression, while at the same time having better protection for the things that cause problems.

With real-world audio, what I've typically observed is that if one band is really loud, all the bands above it are reduced by about the same amount. Which is probably really what we want!

This description might be a bit unclear... I hope it's somewhat clear at least.

Note: Currently, the parameters are still in dB's. There's just an extra checkbox for "soft mode". The reason for this is that ratio's don't really make sense, and using percentages also doesn't feel very intuitive. So now we have a dB-scale like before, but the behavior gets a smoothing curve. If one band is down really deep and the next one isn't at all, then you reach the set dB value.

Stokkemask
The Stokkemask filter in some cases still influenced stereo too much. We have tried to optimize it further. PLEASE BE CAREFUL: THE NEW BEHAVIOR HAS NOT BEEN VERIFIED WITH AN EXTERNAL MEASUREMENT DEVICE.

If you go to Stokkemask - Experimental settings, you'll find 4 settings. 2 influence the processing, the other 2 have the same effect on the measurement used for the GUI. So you can change one of the two to see what effect is has on the other.

First, our measurement shape was slightly too wide. That caused us to touch the sides a bit more often than what should be needed, which caused our protection to kick in a bit too early. That's the 'Improved Gaussian formula" checkbox.

Second, the Stokkemask spec says that you can smooth the measured spikes using a 10 kHz lowpass filter. This one is tricky, because the official description also says that you have to sweep through the spectrum, and we're not doing that. So we're actually measuring in one spot. Long difficult story, what it comes down to: We can probably get away with ignoring very very brief overshoots. The VFW slider allows you to set up to 1 ms of averaging of overshoots. The 10 kHz value in the spec (with a very shallow slope) probably can be approximated somewhat with a setting of around 0.15 ms. That's not much, but it does allow us to ignore some short spikes.

Again, we still need to verify that this still complies to the standard.


Top
   
PostPosted: Sun Jul 20, 2025 10:54 am 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11443
I just received a report that the Stokkemask Gaussian fix causes exceptions on Intel CPU's with SSE4.2 support. Looking into that now.


Top
   
PostPosted: Sun Jul 20, 2025 7:33 pm 

Joined: Thu Jan 31, 2013 12:04 am
Posts: 117
Watchcat doesn''t work, also the 10.60.
I have a zlib1__.dll error
It doesn''tr find the dll file.


Top
   
PostPosted: Mon Jul 21, 2025 6:01 pm 

Joined: Wed Aug 18, 2021 7:24 pm
Posts: 6
10.61 Beta002 Standalone x64
ASIO does not initialize properly.

ASIO Device: Dante Via
Shows up in Dante Via for a split second and vanishes again.

Console Error: TODO: Dante is not yet supportedReceive failed!

--- Part above gets worked on in Discord ---

Also Dante AES67 Output needs Dante Domain Manager to handle SDP files.
DDM is subscription only and quite expensive.

Stereotool should announce itself to the network via SAP to make it work in Dante Controller.

Edit: 10.60 has the same issue.
10.51 did work.


Top
   
PostPosted: Sat Jul 26, 2025 7:58 am 

Joined: Tue Jun 26, 2012 10:18 am
Posts: 53
It says WatchCat 2.0.17 there on the download page but after installing with the "Windows WatchCat installer" the program still shows "WatchCat Version 2.0.16 (build 4bc7940) / build date Feb 24 2025"

Is this going to be updated to support Stereo Tool 10.60?


Top
   
PostPosted: Sat Jul 26, 2025 11:09 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11443
Quote:
It says WatchCat 2.0.17 there on the download page but after installing with the "Windows WatchCat installer" the program still shows "WatchCat Version 2.0.16 (build 4bc7940) / build date Feb 24 2025"

Is this going to be updated to support Stereo Tool 10.60?
I had to revert this build to 2.0.16 due to a bug that caused it to not start on many systems. I expect to have a fix in a few days (together with a 10.61 build - or more likely 10.70 since we've made some big audio improvements in the Adaptive Compressor, and some that are not yet properly tested in the Stokkemask filter in the final clipper).


Top
   
PostPosted: Sat Jul 26, 2025 11:18 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11443
BETA004 is posted. The WatchCat issue should be solved, and "soft linking" in the multiband compressor sounds better. I have updated all the Adaptive Bass Monster presets to use this new behavior. They sound much more pleasant (warmer, less in-your-face) than ever before now. I have been looking for a long time for what it was that I was missing in the sound, and I think this was it.


Top
   
PostPosted: Sun Jul 27, 2025 5:24 pm 

Joined: Thu Jan 31, 2013 12:04 am
Posts: 117
Still the same error


Attachments:
zlib1.jpg
zlib1.jpg [ 89.87 KiB | Viewed 394907 times ]
Top
   
PostPosted: Thu Jul 31, 2025 2:54 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11443
BETA007 is posted! If everything is ok (WatchCat ZLIB1 dependency should be gone now), this is a release candidate for the next version. We need to post an update for WatchCat, and for the Adaptive Compressor improvements. I'll put in part of the Stokkemask improvements (we have done some measurements now, and the new shape should be good. I'm not yet convinced about the VFW filter, and that also needs to be optimized further).


Top
   
Display posts from previous:  Sort by  
Post new topic  Reply to topic  [ 15 posts ]  Go to page 1 2 Next

All times are UTC+02:00


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Limited