Software help archive
Tracks and waveform display randomly go silent or do not load at all
Support
Tracks and waveform display randomly go silent or do not load at all
A read-only archive of old serato.com help threads.
Tracks and waveform display randomly go silent or do not load at all
Product
Scratch Live
Version
2.1.1
Hardware
Rane SL1
Computer
PC
OS
Platform
-
Arnie
9:29 PM - 29 November, 2010
Product: scratchlive
Version: 2.1.1
Hardware: scratchlive-sl1
Computer: pc
OS Version: win764
---
Hi,
I have tried looking for others having this problem to no avail. The problem I'm having is that while in a mix, tracks will stop playing randomly or do not load at all. With stop playing I mean that Serato registers the control record but does not play any audio or show the waveform.
The thing however is, and I have seen references to network issues before, that I want to play my music from my network. I have a big NAS exclusively for this stuff and made sure my network is reliable.
What I did is to assign a network share a drive letter and just start creating crates and adding tracks. This works just fine and I can do everything.
When playing music on this system (AMD X2 5200+, 2GB DDR2, Jetway NC91 mITX board), the load is negligible and temperatures are ok. I have figured out the correct USB port and buffer size to use without having drop-outs at all. Also, the system I use for playing is exclusively reserved for Serato and I make sure to have no other load on the network during my radio shows.
I am almost certain that this problem lies in the loading and caching of tracks and lowering the audio buffer makes the problem less worse but it is still definitely there. It is also noticeable when browsing crates. You can see it is having considerable trouble loading artwork if an artwork view is selected. When I turn off the artwork view, scrolling through the library is no problem at all.
Personally the feeling I'm having is that the whole browsing and loading files is not implemented efficiently enough to be working via the network. Honestly, even with current consumer hardware hardware or my setup specifically it should definitely be possible. The network is gbit, I have a competent HP switch and I have used Serato under much worse circumstances where its performance has been outstanding.
If someone can point out to me if the problem could actually be the network, and if it's solvable or going to be improved sometime soon, then I would be very much obliged. I'm hoping I can get this setup to work because it means a lot less administrative overhead for me to keep my music synchronized in two places.
Version: 2.1.1
Hardware: scratchlive-sl1
Computer: pc
OS Version: win764
---
Hi,
I have tried looking for others having this problem to no avail. The problem I'm having is that while in a mix, tracks will stop playing randomly or do not load at all. With stop playing I mean that Serato registers the control record but does not play any audio or show the waveform.
The thing however is, and I have seen references to network issues before, that I want to play my music from my network. I have a big NAS exclusively for this stuff and made sure my network is reliable.
What I did is to assign a network share a drive letter and just start creating crates and adding tracks. This works just fine and I can do everything.
When playing music on this system (AMD X2 5200+, 2GB DDR2, Jetway NC91 mITX board), the load is negligible and temperatures are ok. I have figured out the correct USB port and buffer size to use without having drop-outs at all. Also, the system I use for playing is exclusively reserved for Serato and I make sure to have no other load on the network during my radio shows.
I am almost certain that this problem lies in the loading and caching of tracks and lowering the audio buffer makes the problem less worse but it is still definitely there. It is also noticeable when browsing crates. You can see it is having considerable trouble loading artwork if an artwork view is selected. When I turn off the artwork view, scrolling through the library is no problem at all.
Personally the feeling I'm having is that the whole browsing and loading files is not implemented efficiently enough to be working via the network. Honestly, even with current consumer hardware hardware or my setup specifically it should definitely be possible. The network is gbit, I have a competent HP switch and I have used Serato under much worse circumstances where its performance has been outstanding.
If someone can point out to me if the problem could actually be the network, and if it's solvable or going to be improved sometime soon, then I would be very much obliged. I'm hoping I can get this setup to work because it means a lot less administrative overhead for me to keep my music synchronized in two places.
Arnie
11:25 AM - 1 December, 2010
I really hope someone can answer my questions. At the very least I would like to know if I have to start thinking about the alternatives or that there are possibilities to make this work. :)
Arnie
10:09 PM - 1 December, 2010
Ok so I had the same thing during a live show. I had a track that was loaded and was playing the intro. Then out of the blue the waveform went blank and it stopped playing, although the timecode signal tracking was working just fine. I really don't understand why this happens. The tune is loaded! :(
Samuel S
11:37 PM - 2 December, 2010
Hi Arnie,
There are a few reasons why we don't recommend playing tracks from a network drive. Using Scratch Live in this way is not something we would typically expect in a performance situation and therefore there hasn't been a large amount of testing done and there are problems that can occur. Firstly, network speeds can be unreliable for the data transfer rate that we require to stream live audio. When this can't keep up, Scratch Live will not be able to load the audio in time and it will drop out. This is a similar situation to using an external hard drive which lags or can't keep up which will do the same thing and drop both the waveform and audio as the data can't be loaded in time.
If you are trying to use this one thing that might help you out is to try raising the audio cache? The larger amount of buffered audio may be able to cover any slow transfer rates and data lags.
Please let me know if you have any further questions :)
Cheers,
Sam.
There are a few reasons why we don't recommend playing tracks from a network drive. Using Scratch Live in this way is not something we would typically expect in a performance situation and therefore there hasn't been a large amount of testing done and there are problems that can occur. Firstly, network speeds can be unreliable for the data transfer rate that we require to stream live audio. When this can't keep up, Scratch Live will not be able to load the audio in time and it will drop out. This is a similar situation to using an external hard drive which lags or can't keep up which will do the same thing and drop both the waveform and audio as the data can't be loaded in time.
If you are trying to use this one thing that might help you out is to try raising the audio cache? The larger amount of buffered audio may be able to cover any slow transfer rates and data lags.
Please let me know if you have any further questions :)
Cheers,
Sam.
Arnie
11:27 AM - 9 December, 2010
Thanks for your feedback Samuel. In the past week there have been some developments which I'd like to share.
Firstly, I owe you an apology because there have been hardware problems that I wasn't aware of yet which only surfaced last week during my live show. The harddrives on the fileserver were dying which caused some serious I/O issues. The problem worsened over the course of the past week so that explains some of the symptoms.
Right now I'm going to fix that and make sure my fileserver is in pristine condition when I attempt new tests. For now I copied my music locally so I can play without problems.
However, and this is me as a software developer talking, I think that the whole I/O process in Serato could maybe be improved. I understand that there are some serious challenges when it comes to buffering audio, especially if you can drop your needle anywhere in a track and need response almost instantaneously.
When looking at the average track size though and the amount of memory modern computers have available, it should be very easy to load tracks entirely in memory, even if they were wave. If you load up a file without decoding it, you can easily hold two playing tracks in memory, which would solve the problem of buffering and playback stopping like I had. Especially if your computer is dedicated to Serato, like mine, it should be able to use all the memory it wants, or maybe have the ability to configure some sort of watermark.
Because Serato is not very transparent like that I have no idea how buffers are implemented and of course this is probably a trade secret. I think that of all features though, this is probably the only thing that I think could be improved dramatically, because for the rest I think this is some great software. :)
Firstly, I owe you an apology because there have been hardware problems that I wasn't aware of yet which only surfaced last week during my live show. The harddrives on the fileserver were dying which caused some serious I/O issues. The problem worsened over the course of the past week so that explains some of the symptoms.
Right now I'm going to fix that and make sure my fileserver is in pristine condition when I attempt new tests. For now I copied my music locally so I can play without problems.
However, and this is me as a software developer talking, I think that the whole I/O process in Serato could maybe be improved. I understand that there are some serious challenges when it comes to buffering audio, especially if you can drop your needle anywhere in a track and need response almost instantaneously.
When looking at the average track size though and the amount of memory modern computers have available, it should be very easy to load tracks entirely in memory, even if they were wave. If you load up a file without decoding it, you can easily hold two playing tracks in memory, which would solve the problem of buffering and playback stopping like I had. Especially if your computer is dedicated to Serato, like mine, it should be able to use all the memory it wants, or maybe have the ability to configure some sort of watermark.
Because Serato is not very transparent like that I have no idea how buffers are implemented and of course this is probably a trade secret. I think that of all features though, this is probably the only thing that I think could be improved dramatically, because for the rest I think this is some great software. :)
Samuel S
10:44 PM - 13 December, 2010
Hey Arnie,
No worries :) In terms of loading two tracks entirely to memory, we use a large amount of RAM already to load library information and therefore don't load the entire two tracks to the audio cache. With most hard drives these days its not necessary to load that much audio to RAM as they are able to keep up with the data transfer rate required for live streaming audio. The reason that we don't recommend network drives is that in a performance situation it is not a common scenario and is not entirely reliable due to the limited amount of testing that has been performed on systems like this.
Cheers,
Sam.
No worries :) In terms of loading two tracks entirely to memory, we use a large amount of RAM already to load library information and therefore don't load the entire two tracks to the audio cache. With most hard drives these days its not necessary to load that much audio to RAM as they are able to keep up with the data transfer rate required for live streaming audio. The reason that we don't recommend network drives is that in a performance situation it is not a common scenario and is not entirely reliable due to the limited amount of testing that has been performed on systems like this.
Cheers,
Sam.
Arnie
10:46 PM - 13 December, 2010
Alright, well so be it then. I'll first make sure my fileserver is working properly again, and then I'll do some more testing. Who knows I won't have the issues any more.
One more thing: you recommend having more audio cache (60 seconds is the max, right) in this situation, right?
One more thing: you recommend having more audio cache (60 seconds is the max, right) in this situation, right?
Samuel S
10:51 PM - 13 December, 2010
In this situation I would recommend a larger audio cache to try and mask these drops in audio/waveform. Let me know how you get on? I'll close this thread off for now :)
Sam.
Sam.