New software mods for the Squeezebox Touch...they work.

[Dynobot]

Also, I've learned from another thread discussing your mods (a different forum)...

I also read on that other forum, after someone posted to Soundcheck about him trying these mods and/or putting them in his next release of mods that Soundcheck actually wants credit for these mods. Apparently he says he thought of them in 2007 but I guess he forgot about them until now since he did not use them any of his Squeezebox mods until now....oh well.

At any rate, I will keep them on my site and help anyone who wants to use them.

[Gazjam]

Both you guys are doing a great job...keep it up!
Maybe great minds think alike?

Heres a thing to consider...

When I upgraded the firmware on my Touch to the latest "nightly" build (7.6.1 - r32869) the "optimum" SPDIF value changed?
I was sitting at 51 as anything higher sounded a bit "etched" and this value gave the best balance between clarity and musical flow.

*seriously, Im not making this stuff up!

Upgraded the firmware, factory reset, Soundcheck mods then yours..and the sound wasn't as clear as before.
I had found that the higher the SPDIF value the "clearer" the top end seemed to be..so I set it to the original recommended value of 59.

Sound is better that it was, more of what made it better than without the mods.
Some voodoo going on here I think.
Software like Amarra etc (insert good/bad comment here..) seems to rely on similar tweaks to system processes, CPU priority etc, so I for one reckon theres definately a good thing going here.

[Krisbee]

I neither own nor have even heard a Touch, but it's interesting to see someone apply the ideas of tuning thread and interrupt priorities in a real-time Linux environment to this popular kit.

While not directly applicable to the Touch, here's one web ref. that explains a few concents and shows how IRQ priorities can be juggled/tuned for low-latency audio.

http://subversion.ffado.org/wiki/IrqPriorities

For a more general overview of real-time Linux in the audio context there is:

http://apps.linuxaudio.org/wiki/real_time_info

And there's always the ALSA low-latency how-to:

http://www.alsa-project.org/main/ind..._latency_howto

Unfortunately like a lot of ALSA documentation the last ref is not very up to date.

[WAD62]

Hi all,

Thanks for trying my SB Touch mods.

FYI I update the site from time to time. I included some information about "Tuning" the mod to suit your system.

Enjoy!

Hi Ed, are you aware of any similar mods for the SB receiver or SB3?

[Dynobot]

Both you guys are doing a great job...keep it up!
Maybe great minds think alike?

Heres a thing to consider...

When I upgraded the firmware on my Touch to the latest "nightly" build (7.6.1 - r32869) the "optimum" SPDIF value changed?
I was sitting at 51 as anything higher sounded a bit "etched" and this value gave the best balance between clarity and musical flow.

*seriously, Im not making this stuff up!

Upgraded the firmware, factory reset, Soundcheck mods then yours..and the sound wasn't as clear as before.
I had found that the higher the SPDIF value the "clearer" the top end seemed to be..so I set it to the original recommended value of 59.

Sound is better that it was, more of what made it better than without the mods.
Some voodoo going on here I think.
Software like Amarra etc (insert good/bad comment here..) seems to rely on similar tweaks to system processes, CPU priority etc, so I for one reckon theres definately a good thing going here.

Maybe not voodoo, I started using the mods with the latest release, which is why I used 59.

BTW, the owner of these mods is FFADO like I referenced in my site. They came up with them years ago and actually published this information which I gave credit too on my site.

http://subversion.ffado.org/wiki/IrqPriorities

I know Soundcheck from many years ago on Audioasylum. It was his pride speaking...

If you notice on my site I reference other sources too like Texas Instruments in ref to the usb tweaks I use ie very short cables etc.

[Dynobot]

I neither own nor have even heard a Touch, but it's interesting to see someone apply the ideas of tuning thread and interrupt priorities in a real-time Linux environment to this popular kit.

While not directly applicable to the Touch, here's one web ref. that explains a few concents and shows how IRQ priorities can be juggled/tuned for low-latency audio.

http://subversion.ffado.org/wiki/IrqPriorities

For a more general overview of real-time Linux in the audio context there is:

http://apps.linuxaudio.org/wiki/real_time_info

And there's always the ALSA low-latency how-to:

http://www.alsa-project.org/main/ind..._latency_howto

Unfortunately like a lot of ALSA documentation the last ref is not very up to date.

Yes Krisbee those are the true sources of information concerning real time linux and alsa. I made references to the very sources on my site.

Thanks for posting that

[Dynobot]

Hi Ed, are you aware of any similar mods for the SB receiver or SB3?

No only products that use Linux

[Krisbee]

Dynobot,

Now I've had a chance to read more of the information on your excellent website I'd say it was well worth a visit for anyone interested in computer audio.

One thing that really struck me was how you had discovered by accident that cutting the 5volt power connection on a USB cable gave such good results that you sold both your ADuM isolator and AUDIO-GD USB to Coax converter. It's an idea that SOTM incoporated into their very expensive tX-USB: Low noise USB audio PCI card with switchable BUS power. Of course this may not work on a USB DAC that needs the 5volt connection, but must be worth a try on a cheap USB cable.

To stray off topic a little further, in your comparison of Windows v. Mac etc. you express a perference for the Squeezebox Touch, but only if the Squeezebox server software riuns on a separate machine. But unless I missed it you don't say which DAC etc you were listening to, or perhps this choice was independent of the rest of your audio chain..

Although Im a dedicated Linux user I still have not built a separate "audio computer", yet I lean towards an Atom powered device or using an embedded sytesm like one of the ALIX m/boards combined with VoaygeMPD linux. Talking of which have you ever seen this website?

http://kubotayo.web.fc2.com/voyagempd.html

It includes some patches for MPD itself to make further use of real time capabilities. I've made no use of these myslef yet so can't say if they are effective. But they are accompanied by a detailed description.

One last question, I wonder if you had ever tried the switch from ALSA to OSS4? Sadly, my DAC's USB connection does not function although recognised in OSS4 , but my Envy24HT card works OK.

[Welder]

I think this is great despite not being a Touch user.
(I am trying/have tried similar with MPD with the assistance of my more Linux knowledgeable friend)
I’ve believed for some time now that experimentation in both hardware and software setup can make a considerable difference to file based audio.

There have been two main problems imo with regard to this kind of system optimization for audio
a) overcoming the “bits are bits” mindset prevalent in computer based audio reproduction
b) the majority of audio enthusiasts just don’t have the computer/software skills to experiment with such modifications.

It’s exciting to read that people are actually hearing a difference in how an OS is set up.
It’s interesting that the more traditional audio enthusiasts are quite happy to report that changing this capacitor or that valve/transistor makes an audible difference but, often still insists that computers were designed to multitask and prefer to concentrate on everything after the computer rather than the source itself.

Like Chris, I found your preference for tx USB style powering solution interesting and something I’m considering for my server build as mentioned in my Music Server thread.
It is rather expensive atm though.
I’ve had good results with the ADUM chip connected to a Linear power supply but not all USB Dacs seem happy with this arrangement.

As mentioned somewhere, some IRQ lines are on the MOB and I wondered if you had experimented with the removal of these lines? Regrettably I’ve scrapped a few boards now attempting to do just this.

There are software solutions available for thread prioritisation and core mapping. Have you tried any?

Imo it’s great that people such as yourself are attempting to bridge the chasm that exists between the computer/coding/scripting and audiophile sides of quality audio reproduction.

[magiccarpetride]

Hi magic,
The best way to insure you do not have extra spaces is to go to the end of the file and backspace until you get to the last character. same goes for each line, backspace to get rid of spaces.

Thanks so much for your selfless help, Dyno. I've applied your mods last night, and wow, wow, wow! And WOW!

What the hell is going on in there? How is it possible that commenting out two lines of code and adding 4 lines of code can result in such staggering improvement? I confess I have not the foggiest what's going on, but man it works. Does it ever work!

I also have to confess that I was extremely skeptical that this mod could make even the slightest difference. After going through the flurry of mods and tweaks (2011 is the tweak year for me), I was convinced that I've reached the point of diminishing returns. The only way I could see things sounding any better was if I were to do some serious upgrades to my electrical wiring, which would require a five figure investment on my part.

Then along came your four lines of code, and wham, bam, a la chazam!

Thank you, thank you, thank you!