similar to: AMR vs Speex on wireless networks.

Rethinking to BER in UMTS, it seems acceptable to use speex on it do you agree? ----- Messaggio originale ----- Da: marzullo maio <marzullo63@yahoo.it> A: Jean-Marc Valin <jean-marc.valin@usherbrooke.ca> Cc: speex-dev@xiph.org Inviato: Sabato 5 maggio 2007, 23:06:40 Oggetto: Re: [Speex-dev] AMR vs Speex on wireless networks. NB for CSD/HSCSD and WB for UMTS/HSDPA. The UMTS/HSDPA

NB for CSD/HSCSD and WB for UMTS/HSDPA. The UMTS/HSDPA services already provides an excellent turbo code correction. For what i know i can tell you more about corrupted packets. The BER ranges from 10^-5 to 10^-2 for CSD/HSCDS in transparent mode, and from 10^-9 to 10^-6 for UMTS/HSDPA. It depends on signal quality and if you're walking/standing or if you're in real mobility (for example

Hi, I've a problem; I've tryed to install wine 0.9.7 on a Mac Computer on Mac OS X System. I've downloaded source files from source-forge.net I have made configuration,and make dependecies and finally I did "make" !...but...at least appears a terrible message: #error unsupported cpu (minidump.c 172.2).... I've tryed this on two different computers: a IBM PowerPC G5

When i use VBR do i need to set VBR quality or speex quality? Is the speex quality referred to the overall quality and instead VBR quality only if i use VBR? Do these qualities override each other? Thanks. --------------------------------- L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail -------------- next part -------------- An HTML attachment was scrubbed...

Do i always need to tell the decoder that some packets has been lost? I noticed that if i discard packets without passing NULL and advising the decoder, sometimes my system hangs and i need to reset my PDA. Since the decoder is mantaining a state, i suppose i always need to tell it a packet has been discarded through a sequence number check. Or can i just pass to speex_decode_int whatever i

I would like to know if my vision of things is correct about frame structure. -------------------------------------------------------- Frame header -------------------------------------------------------- Subframe1 header|Subframe1 content| -------------------------------------------------------- Subframe2 header|Subframe2 content| --------------------------------------------------------

My % of packets lost vary usually in function of signal strenght. Since i cannot forecast the % of frames i will lost except with the criteria just said, could be a good approach to set the PLC tuning % dynamically during runtime basing it statistically on the number of ACKed frames? Does the % of frames lost include silence frames (6 bytes frames, i'm using VBR)? So do i need to consider

Does the PLC tuning affect only the quality in case of PL? Does it affect the quality also in good PL conditions? Thanks. --------------------------------- --------------------------------- L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail -------------- next part -------------- An HTML attachment was scrubbed... URL:

Ops i meant: --------------------------------------------------------------------------- Wideband bit|Mode ID|LSP|OL pitch|OL Exc gain| --------------------------------------------------------------------------- Fine pitch1| --------------------------------------------------------------------------- Pitch gain1| ---------------------------------------------------------------------------

I noticed my latency is gradually increasing in my audio loopback test application. I just hear what i say, encoding and decoding on the fly. I suppose 8000 Hz for the output device is not 8000 Hz for the input one, they differs since they could refer to different time generators (i suppose...). To solve the problem i rised the output sampling rate from 8000 to 8200 Hz, and now i'm not

I've not so much bandwidth to add channel coding but i've few bytes to spare for error detection. With a polynomial CRC-8 i've already more than 99.6 % to detect a burst error in a frame. So let's suppose the critical fields of Speex are: Wideband bit Mode ID LSP OL pitch OL pitch gain OL Exc gain So my plans are to calculate a CRC on these fields for each frame i send, then if i

Hello i have an Imac g4 flat panel 1ghz with geforce 4mx , ubuntu gutsy and a 2.6.22 kernel . I have tried nouveau, but i have some problem. I post also my log ... thx ___________________________________ L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail: http://it.docs.yahoo.com/nowyoucan.html -------------- next part -------------- A non-text attachment was

I decided to upgrade from openSuse 10.2 to openSuse 10.3. I installed xen things, but when I boot using xen option the system is blocked. It appears the green window, with little suse logo in the middle, and mouse cursor, but the system crashes. My computer is a Pentium 4 1,5 GHz, 512 MB of RAM. I attached "diff boot.msg boot.omsg" (as diff opensuse-normal-kernel opensuse-xen-kernel)

Can i use a separate instance of the decoder passing NULL to generate white noise or something that is similar to confort noise? I need to feed it to my output audio autput device for each frame i don't receive in time. Or should i just use the same decoder instance of voice and just pass NULL when i discard or loss frames? Is that the best way? How can i accomplish this purpose? Thanks.

You need to let the know packets are missing so it can generate something meaningful instead. Otherwise, you'll also have less packets to play, which means "something" has to be inserted (zeros, clicks, the content of the kernel memory, ...) instead. As for crashing, it shouldn't happen and I wouldn't be surprised if it had more to do with a soundcard underrun occuring from

hi, i'm used to manual partition and format disks before installing the O.S. but with centos 5 this results in an unbootable system. Usually, to manual partition the disks I use the installation disk in recovery mode.. The partition schema I choose was: /dev/sda1 /boot /dev/sda2 swap /dev/sda3 LVM PV inside the LVM VG I created the following LV / /var /home /tmp Then I formatted all

hello when i Tried to compile xf86-video-nouveau come this message: cc1: error: unrecognized command line option "-minline-all-stringops" make[2]: *** [nv_accel_common.lo] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 ___________________________________ L'email della prossima generazione? Puoi averla con la nuova Yahoo! Mail:

nouveau works with some little bug on my imac g4 geforce4mmx .. my dmesg [ 8086.085247] PCI: Unable to reserve mem region #2:8000000 at 98000000 for device 0000:00:10.0 [ 8086.085390] [drm] Detected an NV17 generation card (0x018900a3) [ 8086.085480] [drm] Initialized nouveau 0.0.10 20060213 on minor 0 [ 8086.085490] [drm] Used old pci detect: framebuffer loaded [ 8226.255226] [drm] OF bios

Ok, I ran my app with +ddraw,+d3d. I got a few warnings related to d3d here and there, ie: warn:ddraw:IDirectDrawImpl_SetCooperativeLevel (0x16d0f8) Unhandled flag DDSCL_ALLOWREBOOT, harmless warn:d3d:IWineD3DDeviceImpl_Clear Clearing depth and/or stencil without a depth stencil buffer attached, returning WINED3DERR_INVALIDCALL warn:d3d:IWineD3DDeviceImpl_CreateSurface Creating a surface with a

Hello, I'm starting to explore rsync to mirror some directories from a disk to another locally. If I understand it right, if I don't include any --delete* parameter, when updating the copy, if some files have been deleted from the source directory they're not deleted from the destination directory. That's exactly what I need, but I'd like those files be treated differently, I