Yes there is! I have to say that I do very much agree with Soundcheck and Gordon! This is exatly what I experience in my setup (Avantgarde Duo, 300B SET 100kg monoblocks, S&B TVC, Lenovo R500, USB DAC). If you try to use same setup on say 2.6.29.5-rt21 vs 2.6.29.5 vanilla (no changes but that), there are sound differences on my high resoulution system even though the CPU only does 5% of activity or so. If you have a "low-mid-hifi" setup, I guess that this difference is not so easy to catch. Lowering the latency does make quite a difference in my system, I can't really explain it mathematically like "quod erat demonstrandum", but lowering latency does improve sound. Bass gets tighter and deeper, timber is longer and 3D effect increases...
IMHO: Keep an open mind, that never hurts...
Cheers
IMHO: Keep an open mind, that never hurts...
Cheers
@TBM: Where have you been? THX for (finally) joining this (my) defense-battle aganst the
the other fellows over here.
It seems that you're pretty much up2date with your kernel.
I just looked up certain performance benchmarks. 2.6.29.5-rt seems to be a real
great improvement over older patches.
BTW I am getting away from ZEN. I'll try the mainline kernel plus rt-patch.
Please let us know your setup. (OS+kernel+config) ( You can also drop me a mail)
I'd like to bring the kernel related Wiki section up2date.
Cheers
the other fellows over here.
It seems that you're pretty much up2date with your kernel.
I just looked up certain performance benchmarks. 2.6.29.5-rt seems to be a real
great improvement over older patches.
BTW I am getting away from ZEN. I'll try the mainline kernel plus rt-patch.
Please let us know your setup. (OS+kernel+config) ( You can also drop me a mail)
I'd like to bring the kernel related Wiki section up2date.
Cheers
Hi there.
If somebody is using ecasound, checkout my Wiki update. I did a slight change to the git compile options.
ecasound introduced liboil to improve its performance by using processor relevant features such as SSE.
Check it out. I think you'll like the result. Let me know what you experience.
Enjoy.
If somebody is using ecasound, checkout my Wiki update. I did a slight change to the git compile options.
ecasound introduced liboil to improve its performance by using processor relevant features such as SSE.
Check it out. I think you'll like the result.
Let me know what you experience.Enjoy.
One more for today.
I updated the Kernel Wiki. ZEN is dead, at least for now from my perspective.
I described how to build a customized mainline 2.6.29.5 kernel patched with the rt21 patch set provided by Thomas Gleixner..
I think it is worth a try. The guys improved the patch-set a lot lately.
Enjoy.
I updated the Kernel Wiki. ZEN is dead, at least for now from my perspective.
I described how to build a customized mainline 2.6.29.5 kernel patched with the rt21 patch set provided by Thomas Gleixner..
I think it is worth a try. The guys improved the patch-set a lot lately.
Enjoy.
Another one:
I added another paragraph to the Kernel Wiki. I describe how you can install a 2.6.30 Ubuntu kernel on a system like Jaunty.
This way you'll have access to a much newer Alsa. Further you can use other new features. They have done a lot with filesystems such as BTRFS. Of course there are much more interesting changes. You need to look up the kernel changelog.
The upgrade is a pretty trivial exercise. Everybody can try it.
Enjoy.
I added another paragraph to the Kernel Wiki. I describe how you can install a 2.6.30 Ubuntu kernel on a system like Jaunty.
This way you'll have access to a much newer Alsa. Further you can use other new features. They have done a lot with filesystems such as BTRFS. Of course there are much more interesting changes. You need to look up the kernel changelog.
The upgrade is a pretty trivial exercise. Everybody can try it.
Enjoy.
RT
Yeah, l moved from ZEN to RT some time ago, RT is superior.
I am a bit on/off the web.
There are often a "religious" discussion regarding audio, and there are a lot of (interesting) meanings, but unfortunately there are a lot of people not having good enough equipment to even hear difference on original CD and a CDR copy or a new cable in their system, and those guys are often the most hard headed in their opinions. If the spec is better it does not automatically mean that it sounds better...
Keep an open mind, there are a lot of "strange things" that influences sound...
Keep it up!
T
Yeah, l moved from ZEN to RT some time ago, RT is superior.
I am a bit on/off the web.
There are often a "religious" discussion regarding audio, and there are a lot of (interesting) meanings, but unfortunately there are a lot of people not having good enough equipment to even hear difference on original CD and a CDR copy or a new cable in their system, and those guys are often the most hard headed in their opinions. If the spec is better it does not automatically mean that it sounds better...
Keep an open mind, there are a lot of "strange things" that influences sound...
Keep it up!
T
Hi Soundcheck,
Is there any reason why any of the following packages is not included in the list of MPD compilation requirements?
git
libao-dev
libsamplerate0-dev
Also, in the ecasound pages: 'git-clone' is no longer available on 9.04. It has been replaced by 'git clone'. And 'sudo autogen-vc.sh' did not work, but 'sudo ./autogen-vc.sh' did. Very minor details.
All in all: Setting things up was easy, thanks to your wiki.
Is there any reason why any of the following packages is not included in the list of MPD compilation requirements?
git
libao-dev
libsamplerate0-dev
Also, in the ecasound pages: 'git-clone' is no longer available on 9.04. It has been replaced by 'git clone'. And 'sudo autogen-vc.sh' did not work, but 'sudo ./autogen-vc.sh' did. Very minor details.
All in all: Setting things up was easy, thanks to your wiki.
Hello Soundcheck,
I was about to try to install the 2.6.30 kernel as indicated in your WIKI, but I saw something strange: the second and the third dpkg command are identical. Is this a typo because the third command has no effect as far as I can see, or is another command required here?
Kind regards,
Eddie
I was about to try to install the 2.6.30 kernel as indicated in your WIKI, but I saw something strange: the second and the third dpkg command are identical. Is this a typo because the third command has no effect as far as I can see, or is another command required here?
Kind regards,
Eddie
Question for Soundcheck - usb dac
Hi Soundcheck,
I read your Linux Audio Wiki, great work.
I am not using Ubuntu, but Mandriva Linux currently. In your Alsa wiki, you mention a modification to the usbaudio.c file. I do have a USB DAC running.
I am rather new to Linux so the question might sound a bit odd. If I just go ahead to modify my usbaudio.c file per your suggestion but do not run any upgrade scripts, will this bring any benefits already to my USB DAC reproduction? Or some re-compile work is needed to make this change effective?
Thanks
Hi Soundcheck,
I read your Linux Audio Wiki, great work.
I am not using Ubuntu, but Mandriva Linux currently. In your Alsa wiki, you mention a modification to the usbaudio.c file. I do have a USB DAC running.
I am rather new to Linux so the question might sound a bit odd. If I just go ahead to modify my usbaudio.c file per your suggestion but do not run any upgrade scripts, will this bring any benefits already to my USB DAC reproduction? Or some re-compile work is needed to make this change effective?
Thanks
audible differences
Hello all,
I recently installed the 2.6.30 kernel but I cannot run it in the generic mode because of a monitor problem, a rather common problem with Ubuntu 9.04. However, the recovery mode works and that is the one that I use for music playback with Ecasound. I compared the new kernel with the best that I have had until now, 2.6.28-rt3. What I did is to play one well-known CD using the new kernel, listen very carefully and immediately after that play the same CD with the old kernel. The differences were clear. With the old rt-kernel more details are heard and the soundstage is more realistic. Not a big difference but clearly distinguishable.
I have no idea of the technical differences between the two. I think that we face here the same problem as with loudspeakers, amplifiers, cables etc. The measurements that are performed do not explain the audible differences. Another complication is that all contributors to this forum have very different playback equipment. If I discuss changes with e.g. Soundcheck we practically always have the same conclusion. This is for a part due to the fact that we have almost the same equipment and we have the same listening attitude.
I started with HiFi 40 (!) years ago and this discussion did not change. Meanwhile I enjoy listening to music and I hope that somebody builds a decent piece of hardware for audio playback because I do not like the messing around with software as I did the last years.
Kind regards,
Eddie
Hello all,
I recently installed the 2.6.30 kernel but I cannot run it in the generic mode because of a monitor problem, a rather common problem with Ubuntu 9.04. However, the recovery mode works and that is the one that I use for music playback with Ecasound. I compared the new kernel with the best that I have had until now, 2.6.28-rt3. What I did is to play one well-known CD using the new kernel, listen very carefully and immediately after that play the same CD with the old kernel. The differences were clear. With the old rt-kernel more details are heard and the soundstage is more realistic. Not a big difference but clearly distinguishable.
I have no idea of the technical differences between the two. I think that we face here the same problem as with loudspeakers, amplifiers, cables etc. The measurements that are performed do not explain the audible differences. Another complication is that all contributors to this forum have very different playback equipment. If I discuss changes with e.g. Soundcheck we practically always have the same conclusion. This is for a part due to the fact that we have almost the same equipment and we have the same listening attitude.
I started with HiFi 40 (!) years ago and this discussion did not change. Meanwhile I enjoy listening to music and I hope that somebody builds a decent piece of hardware for audio playback because I do not like the messing around with software as I did the last years.
Kind regards,
Eddie
Hi folks.
I havn't been around for a while. I am still busy with my battery driven FITPC2, TINY CORE Linux and a proper home network setup.
1. THX for bringing up the little issues in the WIKI. I fixed them.
2. The newest usbdriver has been changed quite a lot. I havn't really looked into it.
3. EDDIE: Try to get the 2.6.29-rt kernel to work. It is IMO pretty well done.
4. Lots of stuff can impact the sound inside a PC. Inside activities are causing a physical
change on the output data stream. Timing problems belong to that group of physical effects.
EMI/RFI. The PC causes pretty heavy power spikes. Heat, vibration, you name it add
more stuff.
Nobody has ever properly explained what's causing the issues. I guess there are too many of them.
However, these issues do exist (clearly audible on high resolution systems).
Example: Flac decoding is causing 20% extra processor load. This will have an impact
on the physical conditions and timing. E.g. even Gordon Rankin (Wavelength audio)
meanwhile agrees to this in general. ( Remember: Over years he and other specialists
were claiming there is nothing because it can not be measured. Obviously there is a
huge need to get the right measurements in place.
There are tons of other problems out there. Computer Audio Asylum is a nice source for
getting a bit more insight.
Another anecdote: Over years "people" were claiming iTunes (on OSX) outputs bit
perfect data. Many of the so called specialists measured it.
People saying it sounds like c... got shot.
Recently a 1500$ application called Amarra (SonicStudio) was launched.
Guess what happened: They measured something like a 1.5db attenuation and a 10db
extra noise supplied by iTunes.
Folks we are not there yet. The whole thing becomes reality if one of the big shots will
officially prove it. At that stage everybody will jump on the train. I'll bet you.
I havn't been around for a while. I am still busy with my battery driven FITPC2, TINY CORE Linux and a proper home network setup.
1. THX for bringing up the little issues in the WIKI. I fixed them.
2. The newest usbdriver has been changed quite a lot. I havn't really looked into it.
3. EDDIE: Try to get the 2.6.29-rt kernel to work. It is IMO pretty well done.
4. Lots of stuff can impact the sound inside a PC. Inside activities are causing a physical
change on the output data stream. Timing problems belong to that group of physical effects.
EMI/RFI. The PC causes pretty heavy power spikes. Heat, vibration, you name it add
more stuff.
Nobody has ever properly explained what's causing the issues. I guess there are too many of them.
However, these issues do exist (clearly audible on high resolution systems).
Example: Flac decoding is causing 20% extra processor load. This will have an impact
on the physical conditions and timing. E.g. even Gordon Rankin (Wavelength audio)
meanwhile agrees to this in general. ( Remember: Over years he and other specialists
were claiming there is nothing because it can not be measured. Obviously there is a
huge need to get the right measurements in place.
There are tons of other problems out there. Computer Audio Asylum is a nice source for
getting a bit more insight.
Another anecdote: Over years "people" were claiming iTunes (on OSX) outputs bit
perfect data. Many of the so called specialists measured it.
People saying it sounds like c... got shot.
Recently a 1500$ application called Amarra (SonicStudio) was launched.
Guess what happened: They measured something like a 1.5db attenuation and a 10db
extra noise supplied by iTunes.
Folks we are not there yet. The whole thing becomes reality if one of the big shots will
officially prove it. At that stage everybody will jump on the train. I'll bet you.
32 or 64 bit OS
Hello Soundcheck,
Thanks for the suggestion. I will try the 2.6.29-rt on my new computer hopefully within a week or two.
Two additional questions: my new PC will have a 64 bit processor and 4 GB RAM. I understand that to fully use this capacity I should use the 64-bit version of Ubuntu. Did anyone of you compare ecasound and sox with 32 and 64 bit operating systems? These results seem to differ widely for different applications.
Are the real-time kernels available in 64 bit or only in 32 bit?
Kind regards,
Eddie
Hello Soundcheck,
Thanks for the suggestion. I will try the 2.6.29-rt on my new computer hopefully within a week or two.
Two additional questions: my new PC will have a 64 bit processor and 4 GB RAM. I understand that to fully use this capacity I should use the 64-bit version of Ubuntu. Did anyone of you compare ecasound and sox with 32 and 64 bit operating systems? These results seem to differ widely for different applications.
Are the real-time kernels available in 64 bit or only in 32 bit?
Kind regards,
Eddie
Today I was listening to a well-recorded audio CD. VLC with default configuration (only switched to the correct alsa device) produced cracks occasionally, female voice was getting distorted in highs at times. Mplayer with precaching of a few MB sounded clean and correct.
I could not find any active plugin of effect in VLC responsible for the distortion. However, I am positive the SW did not perform in bit-perfect mode. This experience makes me even more believe that those notable differences in various SW players sound reported in this thread are in fact caused by some DSP or even bugs and the players are not bit-perfect. Such as the mentioned lower volume of iTunes - apparently nobody checked for bit-perfectness otherwise the issue would have been disclosed right away.
I could not find any active plugin of effect in VLC responsible for the distortion. However, I am positive the SW did not perform in bit-perfect mode. This experience makes me even more believe that those notable differences in various SW players sound reported in this thread are in fact caused by some DSP or even bugs and the players are not bit-perfect. Such as the mentioned lower volume of iTunes - apparently nobody checked for bit-perfectness otherwise the issue would have been disclosed right away.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.