maxlorenz said:
If you wanna play with DAC's you may consider -ecdesigns-' "ultimate NOS DACs", which come with a different aproach to USB: shift register reclocker on BCK. He explains better than me
Hi Max.
I've been in contact with ecdesigns. I had plannned to purchase the receiver/reclocker only. I planned to run it I2S into my DDDAC towers. Then I read some posts questioning the effect of reclocking. That's why I put this project on hold for the time being.
And now - the details presented by the DAC1 made me suspicious. These details are most probably not caused by a lack of jitter effects only. RADIAN mentioned that earlier in context of his ESI Juli experience.
The question is here - Is there a DAC out there which combines the great sound of a NONOS DAC and the level of detail of a dac such as a the DAC1?
Cheers
\Klaus
Hi Klaus,
Well, that's why a dare to recommend DI-DACs
I own OS DAC: a highly tweaked (by myself, regrettably ) M-Audio superDAC (AK4393 DAC chip) which is very detailed and has deep bass, etc. But, it is kind of flat, not in the spatial sense but in the textures, is it difficult to explain...like you said, on the long run DDDAC with the excellent instrumental textures and colorful midrange is more satisfactory (emotional satisfaction is the key), even if you know that there are details that you don't get...
With DI-DACs it is different, stange at first, detailed but "organic" meaning that the detail is not detached from the rest of the sounds but forming part of the whole experience. The best part is the sense of structure that one gets. I say that is like listening to a piece with your preferred conductor instead with an average one (Furtwangler, in my case). Is gestalt the correct term?
EC describes the sound better tan myself and makes a bold statement (which I second) that it sounds as if the DAC was extracting all the musical info there is...
I wish I had the money to try the DI8M...
Well, that's why a dare to recommend DI-DACs
I own OS DAC: a highly tweaked (by myself, regrettably
) M-Audio superDAC (AK4393 DAC chip) which is very detailed and has deep bass, etc. But, it is kind of flat, not in the spatial sense but in the textures, is it difficult to explain...like you said, on the long run DDDAC with the excellent instrumental textures and colorful midrange is more satisfactory (emotional satisfaction is the key), even if you know that there are details that you don't get...With DI-DACs it is different, stange at first, detailed but "organic" meaning that the detail is not detached from the rest of the sounds but forming part of the whole experience. The best part is the sense of structure that one gets. I say that is like listening to a piece with your preferred conductor instead with an average one (Furtwangler, in my case). Is gestalt the correct term?
EC describes the sound better tan myself and makes a bold statement (which I second) that it sounds as if the DAC was extracting all the musical info there is...
I wish I had the money to try the DI8M...
Hi folks.
I'd like to share an experience I made a week ago.
We had our annual DIY-fair here in Germany.
Guess what. By coincidence I met Dr. Uli Brueggemann - remember -
the guy who wrote Acourate - the tool I recommended for generating
your filters for brutefir. ( He also seemed to be involved in Tact and Lyngdorf developments)
We had a nice chat about Linux, convolution, his setup and so forth. He gave a very interesting "lecture" to the crowd for almost an hour to explain his setup and the theory behind it.
( Quite some information you'll find in the whitepaper on his homepage)
Some words about his setup:
A fanless Linux-PC, RME soundcard for digital out ( AES) right into
3 TacT PCM/PWM amps. No DA conversion in between.
His speakers were active 3 way speakers and some Lyngdorf subwoofers in the room corner.
The room was not really what you'd call accoustically well prepared.
He is running brutefir in 64bit mode, calculating the crossover and the room correction.
He put together a perfectly time, phase , frequency and room alligned setup.
The sound was extremely dynamic and clean. I never had such
a clean stage in front of me. Breathtaking!
After this experience it was clear to me that I'd need a full active
system with a full digital AMP. No DA conversion in between.
I am just looking up the options around the TI TAS 5010/5100.
I told him of course about this thread and the WIki. Because he
was telling me that he still regards the PC as a "Black Box".
I'd call his statement kind of understated. He is e.g. also running an rt-kernel.
However. During the last week he followed the recommended tweaks
and we've been in contact about it.
Of course I was curious what would happen once he was finished.
Guess what: According to him --- NOTHING!
I didn't really expected this.
So. What could be the reason for this:
1. His chain is rather immune against incoming jitter.
2. There is a stage further downstream causing so much trouble that the PC generated jitter is not affecting the final result.
3. .......?????
Anyhow- The audible result has been extremely clean. That's why
I would vote for the number 1 option. But you never know.
I know that the RME cards do show the improvments done further
upstream. The secret must reside inside the Tacts.
I think it's really time to look after a full digital multichannel project
now. That's the way to go!
Perhaps some of you guys have an idea where to start with this approach? Perhaps something like this, http://www.diyaudio.com/forums/showthread.php?s=&threadid=91148&perpage=25&pagenumber=1
even though I regard this a bit too complex, since the whole filtering would be done with brutefir.
Cheers
\Klaus
I'd like to share an experience I made a week ago.
We had our annual DIY-fair here in Germany.
Guess what. By coincidence I met Dr. Uli Brueggemann - remember -
the guy who wrote Acourate - the tool I recommended for generating
your filters for brutefir. ( He also seemed to be involved in Tact and Lyngdorf developments)
We had a nice chat about Linux, convolution, his setup and so forth. He gave a very interesting "lecture" to the crowd for almost an hour to explain his setup and the theory behind it.
( Quite some information you'll find in the whitepaper on his homepage)
Some words about his setup:
A fanless Linux-PC, RME soundcard for digital out ( AES) right into
3 TacT PCM/PWM amps. No DA conversion in between.
His speakers were active 3 way speakers and some Lyngdorf subwoofers in the room corner.
The room was not really what you'd call accoustically well prepared.
He is running brutefir in 64bit mode, calculating the crossover and the room correction.
He put together a perfectly time, phase , frequency and room alligned setup.
The sound was extremely dynamic and clean. I never had such
a clean stage in front of me. Breathtaking!
After this experience it was clear to me that I'd need a full active
system with a full digital AMP. No DA conversion in between.
I am just looking up the options around the TI TAS 5010/5100.
I told him of course about this thread and the WIki. Because he
was telling me that he still regards the PC as a "Black Box".
I'd call his statement kind of understated. He is e.g. also running an rt-kernel.
However. During the last week he followed the recommended tweaks
and we've been in contact about it.
Of course I was curious what would happen once he was finished.
Guess what: According to him --- NOTHING!
I didn't really expected this.
So. What could be the reason for this:1. His chain is rather immune against incoming jitter.
2. There is a stage further downstream causing so much trouble that the PC generated jitter is not affecting the final result.
3. .......?????
Anyhow- The audible result has been extremely clean. That's why
I would vote for the number 1 option.
But you never know. I know that the RME cards do show the improvments done further
upstream. The secret must reside inside the Tacts.
I think it's really time to look after a full digital multichannel project
now. That's the way to go!
Perhaps some of you guys have an idea where to start with this approach? Perhaps something like this, http://www.diyaudio.com/forums/showthread.php?s=&threadid=91148&perpage=25&pagenumber=1
even though I regard this a bit too complex, since the whole filtering would be done with brutefir.
Cheers
\Klaus
Always interesting to push limits, and to "see" a better setup than one self have. Strange that there was no sound impact on our tweaked setup... But when he is using a RT kernel, thats perhaps the biggest improvement of them all. Something in the chain could also be smart and buffering/reclocking the data stream?
Was he using USB interface?
I am running 2.6.24rc2 with kmaikaze1 and it sounds better and better!
Cheers,
Tom
Was he using USB interface?
I am running 2.6.24rc2 with kmaikaze1 and it sounds better and better!
Cheers,
Tom
Hi Soundcheck,
Good to hear about a fine all digital multiamp system. In my case the hardest part is finding/building a HiEnd analog active crossover. I would love to avoid that step.
Anyway, I am starting learning and playing with line commands...so far I only managed to create an Xterm Today I read on Ubuntu's site that this could be potentially dangerous as one creates a terminal with permanent admin privileges, so a cracker could jump-in...sudo is safer in that view. What do you think?
Does the Xterm last forever?
Errr...I can't find it anymore...how can I revert the mod?
Sorry to bother
M
Good to hear about a fine all digital multiamp system. In my case the hardest part is finding/building a HiEnd analog active crossover. I would love to avoid that step.
Anyway, I am starting learning and playing with line commands...so far I only managed to create an Xterm
Today I read on Ubuntu's site that this could be potentially dangerous as one creates a terminal with permanent admin privileges, so a cracker could jump-in...sudo is safer in that view. What do you think?Does the Xterm last forever?
Errr...I can't find it anymore...how can I revert the mod?
Sorry to bother
M
maxlorenz said:
Me too!
I am looking forward to a FIR filter based crossover. maxlorenz said:
That's correct. It is "potenially" dangerous. ----- So what?
Under Linux the security risk is much lower compared to Windows.
It's up to you to "sudo" everything or to make your life easier.
Perhaps I should point that out in the Wiki.
Cheers
\Klaus
upgrades
Hello all,
Now it is time for an extended update from my side. First a very brief history: I started playing CDs with a laptop and a Twindac two years ago, using Foobar. I did the regular update with a RAM-disk and an ASIO plug-in and I was content because the sound was better than that coming from my very expensive Theta's.
But, early this year I made the step to Suse Linux and XMMS again using the RAMdisk. Setting the realtime option of XMMS and using 'nice -20' was again an improvement. In this period I came across the messages of Soundcheck on 'real' realtime. So in June I installed Ubuntu 7.04 with a realtime kernel. At this time also 'brutus' came along and with lots of help from Soundcheck I managed to get it installed. No doubt that 'brutus' gave better music than XMMS.
The following upgrade came soon: converting the wav's to 48 kHz. I was sceptical beforehand on how this could improve the sound, but after hearing one CD I was convinced. The trick seems to be that bunches of 44.1 kHz samples cannot be send in a regular way. After converting more than 300 CDs offline Soundcheck came with another conversion program, the Secret Rabbit Convertor. And, as more people concluded, it easily beats the results of Shribath, the first program.
With SRC I began to get problems with my laptop. I could not get the batch version working. First it turned out that one of my memory modules gave errors, but replacing it helped only partly. A second problem was that I had two USB harddisks of the same brand. Ubuntu called them LACIE_ and LACIE__ but the order of the underscores was random. A futher problem of the laptop was that its harddisk and fan were too noisy for music late at night.
So the radical solution was the purchase of a new PC, a desktop with a 500 GB internal harddisk, 2 GB fast memory, two processors at 2 GHz and quiet fans. The price of this PC was 700 Euro, quite nice I would say, considering that it has an Asus P5KC motherboard with a PCI-express bus and amongst many other options it has a complete fan control system build in. So I could easily set the fans to the lowest speed, but with temperature control. (Easily means that it took me three evenings to get to this point).
I installed Ubuntu 7.04 on the PC and tested the SRC batch program. With all data on the internal harddisk it ran without any problem and more than two times faster per CD than the laptop. I made batches of 50 CDs that took 16 hours.
After that, on september 14, I started to install the 2.6.21 realtime kernel just as I had done two months earlier on the laptop. This kernel kept me struggeling for several weeks with all kinds of unexpected error messages. In the meantime version 2.6.22-ck1 had come across so I tried to switch but again lots of errors. I heard rumours that this last kernel was for Ubuntu 7.10, but this was denied on this forum.
I now know a little bit of what was causing these problems. The first is that Con Kolivas stopped with the development of his realtime patch. Further kernel 2.6.22 appears to have had insufficiently tested parts, by now it does not exist any more. News came that in 2.6.23 everything would be cured, especially the realtime part. This would combine the best parts of the works of Con Kolivas and Ingo Molnar.
So, I waited for this one and meanwhile installed Ubuntu 7.10 after completely cleaning the internal harddisk, using the old USB-disks as backup for the WAV's. The news came, from soundcheck about a realtime patch called kamikaze, apparently the best ever. And he promised to upgrade the Wiki accordingly.
About two weeks ago I installed 2.6.23 with kamikaze5 and it went easily without any problems. Last step was to install a new version of the brutus scripts, again made by Soundcheck. Yesterday evening (!) was the first time that I enjoyed a full evening of listening to music without interruptions or whatever due to software problems. The final result is that my my system is sounding better than ever. High frequencies are more natural, attacks of drum sounds are faster, voices sound very good, the sound stage is wide and deep, …
So after almost one year I am very happy with Linux and brutus, thanks to Soundcheck again. It has cost a lot of time and occasionally frustration, but I learned a lot about computers meanwhile (I hope). I considered a fanless PC several months ago but the price difference with the PC that I now have allows me to upgrade my loadspeakers with Bastanis widerange units and Airforce tweeters. I have a feeling that this is a better investment in improving my sound.
Next upgrade: a digital amplifier? I heard a Lyngdorf system recently with two tiny monitors, two corner woofers and a digital processor. Quite nice considering the size, but by no means high-end sound.
Kind regards,
Eddie
Hello all,
Now it is time for an extended update from my side. First a very brief history: I started playing CDs with a laptop and a Twindac two years ago, using Foobar. I did the regular update with a RAM-disk and an ASIO plug-in and I was content because the sound was better than that coming from my very expensive Theta's.
But, early this year I made the step to Suse Linux and XMMS again using the RAMdisk. Setting the realtime option of XMMS and using 'nice -20' was again an improvement. In this period I came across the messages of Soundcheck on 'real' realtime. So in June I installed Ubuntu 7.04 with a realtime kernel. At this time also 'brutus' came along and with lots of help from Soundcheck I managed to get it installed. No doubt that 'brutus' gave better music than XMMS.
The following upgrade came soon: converting the wav's to 48 kHz. I was sceptical beforehand on how this could improve the sound, but after hearing one CD I was convinced. The trick seems to be that bunches of 44.1 kHz samples cannot be send in a regular way. After converting more than 300 CDs offline Soundcheck came with another conversion program, the Secret Rabbit Convertor. And, as more people concluded, it easily beats the results of Shribath, the first program.
With SRC I began to get problems with my laptop. I could not get the batch version working. First it turned out that one of my memory modules gave errors, but replacing it helped only partly. A second problem was that I had two USB harddisks of the same brand. Ubuntu called them LACIE_ and LACIE__ but the order of the underscores was random. A futher problem of the laptop was that its harddisk and fan were too noisy for music late at night.
So the radical solution was the purchase of a new PC, a desktop with a 500 GB internal harddisk, 2 GB fast memory, two processors at 2 GHz and quiet fans. The price of this PC was 700 Euro, quite nice I would say, considering that it has an Asus P5KC motherboard with a PCI-express bus and amongst many other options it has a complete fan control system build in. So I could easily set the fans to the lowest speed, but with temperature control. (Easily means that it took me three evenings to get to this point).
I installed Ubuntu 7.04 on the PC and tested the SRC batch program. With all data on the internal harddisk it ran without any problem and more than two times faster per CD than the laptop. I made batches of 50 CDs that took 16 hours.
After that, on september 14, I started to install the 2.6.21 realtime kernel just as I had done two months earlier on the laptop. This kernel kept me struggeling for several weeks with all kinds of unexpected error messages. In the meantime version 2.6.22-ck1 had come across so I tried to switch but again lots of errors. I heard rumours that this last kernel was for Ubuntu 7.10, but this was denied on this forum.
I now know a little bit of what was causing these problems. The first is that Con Kolivas stopped with the development of his realtime patch. Further kernel 2.6.22 appears to have had insufficiently tested parts, by now it does not exist any more. News came that in 2.6.23 everything would be cured, especially the realtime part. This would combine the best parts of the works of Con Kolivas and Ingo Molnar.
So, I waited for this one and meanwhile installed Ubuntu 7.10 after completely cleaning the internal harddisk, using the old USB-disks as backup for the WAV's. The news came, from soundcheck about a realtime patch called kamikaze, apparently the best ever. And he promised to upgrade the Wiki accordingly.
About two weeks ago I installed 2.6.23 with kamikaze5 and it went easily without any problems. Last step was to install a new version of the brutus scripts, again made by Soundcheck. Yesterday evening (!) was the first time that I enjoyed a full evening of listening to music without interruptions or whatever due to software problems. The final result is that my my system is sounding better than ever. High frequencies are more natural, attacks of drum sounds are faster, voices sound very good, the sound stage is wide and deep, …
So after almost one year I am very happy with Linux and brutus, thanks to Soundcheck again. It has cost a lot of time and occasionally frustration, but I learned a lot about computers meanwhile (I hope). I considered a fanless PC several months ago but the price difference with the PC that I now have allows me to upgrade my loadspeakers with Bastanis widerange units and Airforce tweeters. I have a feeling that this is a better investment in improving my sound.
Next upgrade: a digital amplifier? I heard a Lyngdorf system recently with two tiny monitors, two corner woofers and a digital processor. Quite nice considering the size, but by no means high-end sound.
Kind regards,
Eddie
Hi folks.
Some "good" Kernel News:
Miguel Boton, the guy in charge for the Kamikaze kernel patches has renamed his patchset to "ZEN"-patches.
He also introduced a new way of managing and fetching the source data.
Very convenient - I'd say.
Now you're able to keep track on the progress these guys are doing.
Miguel is also providing "stable" snapshots of the kernel incl.
patches.
The current release is 2.6.24-rc3-zen1.
The realtime tunables are btw set to highest performance now.
I doubt that you'll get a better Kernel then a ZEN Kernel.
I'll update the WIKI accordingly.
\Klaus
Some "good" Kernel News:
Miguel Boton, the guy in charge for the Kamikaze kernel patches has renamed his patchset to "ZEN"-patches.
He also introduced a new way of managing and fetching the source data.
Very convenient - I'd say.
Now you're able to keep track on the progress these guys are doing.
Miguel is also providing "stable" snapshots of the kernel incl.
patches.
The current release is 2.6.24-rc3-zen1.
The realtime tunables are btw set to highest performance now.
I doubt that you'll get a better Kernel then a ZEN Kernel.
I'll update the WIKI accordingly.
\Klaus
soundcheck said:Hi folks.
I'd like to share an experience I made a week ago.
We had our annual DIY-fair here in Germany.
Guess what. By coincidence I met Dr. Uli Brueggemann - remember -
the guy who wrote Acourate - the tool I recommended for generating
your filters for brutefir. ( He also seemed to be involved in Tact and Lyngdorf developments)
We had a nice chat about Linux, convolution, his setup and so forth. He gave a very interesting "lecture" to the crowd for almost an hour to explain his setup and the theory behind it.
( Quite some information you'll find in the whitepaper on his homepage)
Some words about his setup:
A fanless Linux-PC, RME soundcard for digital out ( AES) right into
3 TacT PCM/PWM amps. No DA conversion in between.
His speakers were active 3 way speakers and some Lyngdorf subwoofers in the room corner.
The room was not really what you'd call accoustically well prepared.
He is running brutefir in 64bit mode, calculating the crossover and the room correction.
He put together a perfectly time, phase , frequency and room alligned setup.
The sound was extremely dynamic and clean. I never had such
a clean stage in front of me. Breathtaking!
After this experience it was clear to me that I'd need a full active
system with a full digital AMP. No DA conversion in between.
I am just looking up the options around the TI TAS 5010/5100.
I told him of course about this thread and the WIki. Because he
was telling me that he still regards the PC as a "Black Box".
I'd call his statement kind of understated. He is e.g. also running an rt-kernel.
However. During the last week he followed the recommended tweaks
and we've been in contact about it.
Of course I was curious what would happen once he was finished.
Guess what: According to him --- NOTHING!
I didn't really expected this.
1. His chain is rather immune against incoming jitter.
2. There is a stage further downstream causing so much trouble that the PC generated jitter is not affecting the final result.
3. .......?????
Anyhow- The audible result has been extremely clean. That's why
I would vote for the number 1 option.
I know that the RME cards do show the improvments done further
upstream. The secret must reside inside the Tacts.
I think it's really time to look after a full digital multichannel project
now. That's the way to go!
Perhaps some of you guys have an idea where to start with this approach? Perhaps something like this, http://www.diyaudio.com/forums/showthread.php?s=&threadid=91148&perpage=25&pagenumber=1
even though I regard this a bit too complex, since the whole filtering would be done with brutefir.
Cheers
\Klaus
If you can get a HDMI driver up an running under linux there would be plenty of inexpensive multichannels amps to choose from, even fully digital amps based on the TI chipset, e.g. the Panasonic SA-XR700 (7 channels of pure digital power)
cph2000 said:
Interesting approach. I just looked up the HDMI-Audio support
of Linux. Seems to be in the very beginning of being implemented. Though some graphic adapters are supposed to be supported already (since 2.6.22- ALSA 1.0.14).
If I could address every channel separately via HDMI, as it is done with a multichannel card, HDMI might be a very interesting approach in combination with a Panasonic amp as you mentioned.
However, AFAIK HDMI would still be a serial interface,
where the receiver is supposed to recover the signal. I wouldn't
expect any advantages over USB and/or Firewire on the clocking/jitter side.
\Klaus
HDMI 1.0 and up allows transportation of multi channel LPCM (no compression), i.e. 8channel/192 kHz/24-bit audio capability !
(you can not do this with USB 1.0 due to bandwidth restrictions)
The signal is thus NOT decoded at the receiver.
The receiver approach has the advantage of simple remote controled volume regulation at the amp (no digital attenuation need on the source signal in the PC) or diy multchannel attenuate designs needed. You get many channels of amplification for next to nothing.
(you can not do this with USB 1.0 due to bandwidth restrictions)
The signal is thus NOT decoded at the receiver.
The receiver approach has the advantage of simple remote controled volume regulation at the amp (no digital attenuation need on the source signal in the PC) or diy multchannel attenuate designs needed. You get many channels of amplification for next to nothing.
Has anyone released an IIR crossover for Linux?
I am using IIR style filters/RME sound cards under windows now.
I'd like to try Linux for fun, now, and was wondering if their was anything out there.
For now, I don't want to try Brutefir as I'm satisfied with the latency, low CPU usage, and sound quality of IIR filters.
(Got to justify my recent embedded/realtime Linux book acquisition)
I am using IIR style filters/RME sound cards under windows now.
I'd like to try Linux for fun, now, and was wondering if their was anything out there.
For now, I don't want to try Brutefir as I'm satisfied with the latency, low CPU usage, and sound quality of IIR filters.
(Got to justify my recent embedded/realtime Linux book acquisition)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.