Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Bebop0502]

I have read a cuouple of post on drop-outs to try find a solution. I had them more or less since quite a time, both in W7, W8 and W10 and in different versions of JRiver. Some days no issues at all but more normal is 1-2 drop outs on each album played.

I use a Packard lap top with AMD-processor an 8 GB internal memory with a USB connected to an external HD of 3TB filled up to about 50% (approx. 40,000 files, mainly FLAC but also some DSD files). I see no difference between file size (16/44.1 - 24/192 or DSD).

My DAC is an Aestethix Pandora Signature which is using WASAPI (exclusive mode) and is fed with a USB connector

I should like to have your opinion on why this is happening, like:
- If you think JRiver is causing this, how should the parameters be set to minimize this issue?
- If this issue is due to the laptop, what could then be the cause for this (AMD vs Intel...?)
- Same question if you think the DAC is the guilty one (my DAC is tube run; could this be something? Could it be Wasapi? Less risk with Asio?)
- Disturbences from the power net? Earth ground issues?
- Can it be caused by USB (non stable?)

It is quite frustrating when it happens, not least when I have visitors when I bragging about my system including JRiver. Even more frustrating is that I do not know where to start or what is the most likeley reason.

For the rest, I have no issues of whats so ever.

[JimH]

Is your USB drive USB 3.0?  If so, is it on a USB 3.0 port?

The "Weird Problems" link in my signature might give you some ideas.

Antivirus often causes problems.

DSD requires a good CPU.

Increase your buffer size in audio settings.

Power is not a likely cause.

[dtc]

It could be a lot of things, but it is possible that there is a delay to your USB DAC. I would suggest trying the Load Files at Start of Playback in the Audio Options. It puts the entire track in memory before playing. That pretty much eliminates disk and other timing issues during playback.

Playing DSD files does not take much CPU, but converting DSD to PCM on the fly does. So, if you are doing that, CPU may be an issue.

Is there anything else running when you notice the problems?

[Wasper]

First, try foobar as player.  You should have the same problem.  I would be really surprised if JRiver has something to do with your problem.

It tooks me months to figure out, but I had more or less the same problem as you. (http://yabb.jriver.com/interact/index.php/topic,107632.msg747141.html#msg747141) 

Be sure than your PC, amplifier and DAC are connected to the same power line.  Ground loops are definetely not good.

Also, I would suggest you to run with the less cable connected as possible.  (PC to DAC (USB), DAC to amplifier - remove any other input to the DAC if there is any).  Then keep only your DAC input to the amplifier.  If the problem is resolved, then add slowly your devices until you find the dirty one!

Even if I'm not an expert, I Hope it could helps you out and I know how frustrating it is...

[Bebop0502]

I really appreciate your support and it gave me some hints to start with. Some comments:

- I have separated power supply for my dac and laptop/hd. I will see what happens if I put on the same phase.
- I use Norton antivirus. I will disconnect both Norton and wifi to se if this changes anything.
- I have tried "load files" before but only together with "play from menory". This was worse. And running only from memory I get some sconds quiteness on each new tune. I thought load files had to run with play from memory. I will try only "load files"
- I run USB 3.0 cables throughout.
- I can't see more issues with DSD than PCM.

What is the meaning of a "good CPU? I guess you do not only thinking of size (I have 8 GB)?

Strange enough I have not had any drop outs today... I'm on my 3d album now. Yesterday was a pain in the ♧♢♡.

[dtc]

You should not use both play from memory options at once.  Disable Play from Memory and just use Load Files  to Memory. By selecting both you are basically using Play from Memory which is known to cause hiccups.

A good CPU  refers to speed, not the amount of memory. Under Help you can run a MC benchmark. You would like to see a JMARK  greater than 2500 as an example, although you can run with much less than that depending of what options you use.

Are you using any DSPs? IF you are upsampling or downsampling, you should use the SoX option in Audio Options.

Norton can be a hog.

What CPU load do you see during playback?

[Awesome Donkey]

Typically in my experience, an AMD CPU in a notebook/laptop performs pretty terribly compared to the Intel counterparts. I wouldn't be surprised if the system wasn't up-to-par with such operations, especially DSD. A CPU, especially AMD CPUs, can be pretty fast speed wise yet still suffer when dealing with single threaded and multithreaded applications.

The benchmark score will give us a better idea.

[Bebop0502]

You should not use both play from memory options at once.  Disable Play from Memory and just use Load Files  to Memory. By selecting both you are basically using Play from Memory which is known to cause hiccups.

//

What CPU load do you see during playback?

It seems that this seems to be the problem. I am now only using "Load files to memory" I played 4-5 albums yesterday and I have continued today with standard PCM up to 24/174 plus DSD64. I also changed "Buffering" from 100 to 250 ms. So far so good. I had no issues of what so ever. I do not use any DSP features.

Here are two print screens on CPU use. The first one (Chick Corea) is a standard 16/44.1 file and the other (Franz Schubert) is a DSD64 file. As far as I can see the usage is about the same around 17-18%.





Anyway, it works now and hopefully will continue to do so. ? Thanks to all of you. You are my heros of today...

[dtc]

Hope that was it. I think the program should make these options mutually exclusive.

[JimH]

We will make that change.

[dtc]

Thanks Jim!

[dtc]

One suggestion would be to also reword the two options. As they stand, there is no easy way to tell the difference. I would suggest options something like

Play from Memory
         Load compressed file to memory
         Load uncompressed file to memory

with the sub-options mutually exclusive.

That makes the options somewhat self explanatory.

[Matt]

Coming next build:
Changed: Memory playback has been consolidated to one option in Options > Audio > Memory playback.  The modes are: No memory playback, Load full file (not decoded) into memory, Load decoded file into memory.

[dtc]

Thanks Matt. I think that will help alleviate some confusion about the options. I really like the new option. It particular helps when loading large, hi rez files from a usb drive.   

[Bebop0502]

Play from Memory
         Load compressed file to memory
         Load uncompressed file to memory

I don't get this. What is it today? I have understood it as I load the file I want to play whatever size?

Do you mean that that, say a 16/44.1 will be compressed when loaded or is it that FLAC file will be loaded as FLAC or first be uncompressed and then loaded... or something else

[dtc]

There are 2 options. One decompresses a compressed file (like flac) and loads the decompressed data into memory. The other option loads the flac file into memory and decompresses it as it plays. The first option takes more memory. In that mode,  sometimes  MC does not allocate enough memory to hold the whole file, and part way through the track it reads more of the file, decompresses it and loads that into memory. That activity can cause hiccups. The second option loads the original file into memory and decompresses it as it plays. With that, there is much less chance of any hiccups. There is a constant small amount of CPU used to do the decompression on the fly.  In either mode, there is no compression added. The issue is simply when a compressed file (like flac) gets decompressed.

Currently if you check both, the decompress before loading option is used. That is what was causing your hiccups.