Help - Search - Members - Calendar
Full Version: ABC/HR for Java, new version
Hydrogenaudio Forums > Hydrogenaudio Forum > Validated News
schnofler
Hello,

I have developed my Java version of ABC/HR a bit further during the last few days.

You can download it here.

EDIT: Also available at rarewares now (many thanks to Roberto):
http://rarewares.hydrogenaudio.org/files/a.../abchr-java.zip

The zip file contains an executable jar file and sources. You probably need at least JRE 1.3 to run it, I recommend installing the current JRE 1.4.2 if you run into any troubles.
Since I haven't tested the program on any other system than my own (Win XP, JRE 1.4.2), I would be grateful for any error or success reports.

New features in this version are:
  • Encryption
    In the test setup dialog you can create a key file and save the setup to an "encrypting config file". These config files contain a public key which is used to encrypt the results files. You can then use the key file to decrypt the results. This mechanism is not intended to be 100% safe (as was discussed on another thread before, there can be no such method).
  • Offset detection
    You can automatically detect offsets when setting up a test.
  • Quick ABX Test
    You can set up an ABX test with two wave files, without having to set up a whole ABC/HR config (similar to WinABX). The program will automatically correct the offset and volume of the files. Volume correction is not yet available for normal ABC/HR setups, mainly for two reasons: 1. I don't have full trust in my method yet (it works for synthetic tests, like two identical files, one of which was reduced in volume; for all other (normal) situations it comes reasonably close to what replaygain says). And 2. I'm too lazy to put more controls in the test setup dialog.
  • Greatly decreased latency
    It's near zero now, on my machine. That might be different on other systems, depending mainly on JRE version and OS (as I understand it, the Java Sound API is quite low-level, so implementations probably differ somewhat). The method I use to achieve this is a bit risky and might cause problems with synchronisation between the two threads of the program. I haven't found any myself, but I would be grateful for bug reports.
  • Position indicator in the time line during playback
  • "Bookmarks" in the time line
    Probably my favourite new feature smile.gif . By using ctrl-doubleclick you can set (or remove) a marker in the timeline. If you hold ctrl while changing the playback range, your cursor will snap to these markers.
  • Customizable labels for the rating scale
    Might be needed for near-transparency tests.
Have fun,

schnofler
ff123
Oh, this release is sweet. smile.gif

The position indicator and the latency reduction are welcome indeed.

On my computer (Win2K OS), there is a delay between when the button is pressed and when the switch occurs (but the interruption of sound is practically non-existent). There is still audible switching noise, though, which might be reduced with some sort of cross-fading.

Custom ratings labels -- cool!

ff123
Continuum
Cheers from the crowd! smile.gif

I have a problem though: after loading/creating a new test there are some windows redraw problems. The form is blank and visual objects only reappear after moving it outside the screen. I didn't have this problem with an older version (1.0a).

Java version:
Java™ 2 Runtime Environment, Standard Edition (build 1.4.1_02-b06)
Java HotSpot™ Client VM (build 1.4.1_02-b06, mixed mode)
schnofler
QUOTE
There is still audible switching noise, though, which might be reduced with some sort of cross-fading.

I suspect that the switching noise occurs because the playback of the next sample doesn't start exactly where the previous sample ended, which could happen because there is a slight delay between the moment when the starting point is calculated and when the actual switch happens (during this time the previous sample still plays). And even if the switch was sample-precise, a click sound would most likely be unavoidable without crossfading. I probably won't use crossfading anyway because it would be a pain to synchronize the threads with that (just imagine the user could switch again or the playback could end during crossfading ohmy.gif ).

QUOTE
I have a problem though: after loading/creating a new test there are some windows redraw problems. The form is blank and visual objects only reappear after moving it outside the screen.

Hm.. this doesn't happen on my machine, but I know the problem well enough. I had that same problem several times with other projects (Java "forgetting" to repaint a window after resizing or adding controls, even if it's explicitly told to do so). If you manually force it to repaint (e.g. by resizing the window a tiny bit), it usually does so instantly. I don't know of a good way to solve this problem (I do know a few "dirty" ways which usually include lots of trial and error). Ah well, maybe I'll find something on the web. Thanks for trying the program, anyway.
guruboolez
I'm pleased by the innovation of your ABX software. I tried to perform some blind test with it, but unfortunately, some annoying clicks killed my concentration. A click appeared each time I'm playing a file. I take a look on resource used by the software: 90%, on a notebook, Athlon XP 2000, on a WinXP SP1 system. Is it normal ? I'm using the lastest java application.
schnofler
QUOTE
I'm pleased by the innovation of your ABX software. I tried to perform some blind test with it, but unfortunately, some annoying clicks killed my concentration. A click appeared each time I'm playing a file. I take a look on resource used by the software: 90%, on a notebook, Athlon XP 2000, on a WinXP SP1 system. Is it normal ? I'm using the lastest java application.

First, thanks for testing. Unfortunately, I can't reproduce this clicking problem. Could you give me some more information on this? Like, does it happen with every file? Does it happen at the beginning of playback, everytime, regardless of where playback starts? I would really be grateful for any information. If I can nail the problem down, I will get a bugfix out as soon as possible.
The high resource usage is normal for this release. It shouldn't hurt system performance, though, because a lot of this is just "busy waiting", i.e. the program only eats up idle processor cycles. I already fixed this behaviour in my current "work-in-progress" version.
guruboolez
I had to try again to be more precise.
I have other things planed for the moment.

These clicks are maybe due to my 'soundcard' : AC chip. I should try on my desktop computer (I couldn't do it during the week).

Complete report before monday smile.gif
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2008 Invision Power Services, Inc.