Compositor SoftwareCompositor Software

Tag : RT-z128

06
Aug, 2018

By ruslany

Compositor Software news 25.07.2018 – 06.08.2018

Compositor Software news 25.07.2018 – 06.08.2018

More than 6 months I have been working on revealing the leakage in Compositor kernel. Exposing all oscillators and transferring the first oscillator in static mode, I discovered that many resources released in computer RAM (DDoS attack). It lowers the uninterrupted device work to 4 days taken that paging file is set to a size of 64 GB. Increasing the number of oscillators to 32 on each of the layers (Rt, Sr, Tr) and meeting a condition of counters, the emissions equaled to 14 GB a day. In accordance to this, I decided to shut down the peering network and return to RT-z128 kit layout as in Compositor v5 Hypervisor. In this layout the main ports are closed for inbound and outbound connections. Only addresses starting from 192, which are assigned to local machine, are available. It means that a new patch remains the VLF connectivity, but doesn’t allow devices to connect for information exchange. It is possible to exchange information from the local machine for the patch user, but not a remote access user. In particular, such decision is motivated by a hacker attack on RAD96 server. The intruder used the open ports of the windowing device and sent spam from Internet Provider IP address. With closed ports the little leakage on real-time generator is possible, but it is not accounted due to the slow regeneration speed.

To restart the peering network I attempted the following solution: due to closing of the ports, only Compositor Networks ether aggregators take part in peering. This means that feeding ether aggregator wavetables in the peering patch through the Compositor v9 Hypervisor, I create a communication service between all Compositor Library wavetables. However, many Ethernet devices can’t use Compositor peering network as it was at the beginning of the peering network creation. As a conclusion, I can exclude peering from Compositor kernel load test and use multi-kernel mode, but this type of test has very long loading time (up to 4 hours for full load).

Bounded by Royalty project, the database of server emissions in form of wavetables grown to 2627 ether aggregators. The database was expanded by ether emission of Exalted – Cavity track (Ruslan Yusipov, CEO of Compositor Software project). The whole volume of Cavity track emission is 328 wavetables. Wavetables contain radio repeaters, transmitting stations, Ethernet routers, injectors and other ether equipment. It was possible to increase the Compositor v9 Hypervisor regeneration speed up to 150-omega by applying a new peering network patch. This influences only Compositor AV Extended auxiliary channel and doesn’t cover the generic feeder modules. It was made aiming the faster composition speed of the whole pool of Compositor Networks ether aggregators. The necessity in uniform composition caused by big amount of wavetables in Compositor Library. The application of the new patch after the attack was made unnoticeably, because this solution proved to be successful in 2017. It was clear right from the beginning that hackers aim is to limit the communication circle of network devices to Compositor Library pool and its real emission. From one side, if the emission is made every time when server communicates, then there will be much more devices in Compositor Library. Each streaming playback with working patch is a communication with its transmitting devices. Taken this, it is needed either to shut down any internet activity from the patched machine or to make emissions of the whole material, which is played in browser resulting in big amounts of information. From the other side, with open ports it is possible to connect not to the virtual network, created by composition feeding, but directly to all transmitting devices, which reside in Ethernet without the need to make emissions so often.

, , , , , , , , , , , , , , , ,
24
Jul, 2018

By ruslany

VLF voice communications

VLF voice communications

To communicate in VLF network it is enough to use the Compositor v8 injector, but to connect to other networks it is needed to create the service of communication with them, sending generic RT-zX processes in to the transmission channel as in Compositor v9 Hypervisor. I established the connection of two Compositor v8, which was hardly possible with 6th and 7th versions of Compositor program. Moreover, such connection was before the 7th version but it contained another VLF ethers, which made the determination of the transmitting station itself and creation of the protected communication channel not possible. For the experiment I executed the Compositor v8 a16 on the stationary computer and played the voice track (recorded with Dictaphone) in injector channel. I executed Compositor v8 b3 with modulation combinations on the notebook without the opportunity to inject the wavetables in the transmission channel. Using identical device settings, I received a loop in the notebook feedback chain, which definitely coincided with voice text timbre recorded with Dictaphone. Then I executed Compositor v8 a16 with ability to inject wavetables on notebook. I composited the same wavetable combination as on stationary computer and injected them in to the channel without voice track. This way, I established a definite non-repeating signal reproduced as random noise bursts with narrative text structure. This text coincided with signal performed on stationary computer transmission channel, but had another rhythm and pause appearance. I understand that for VLF communication it is enough to transmit the ostinato code pattern on the defined frequency and I credit the Compositor v8 communication experiment as successful.

My main goal is to prove the appearance of broadband communication lines in VLF. If it is possible to receive the broadband signal cycle in transmission channel, then you can try to reconstitute it by injecting the stochastic carrier using RT-z128 and RT-z64 channel modules. Then, following this logics, there will be the complete voice track with the receive quality of original translation in the feedback chain. This should be proved, establishing a connection of two Compositor v9 Hypervisor programs and their virtualization modules. If it happens, the arrangement of wavetables into the transmission channel lines will be proved.

Let’s look at the created VLF transmission lines as grains, where a separate transmission line segments coincide with wavetables, encased in window function envelope. The connection service in VLF network is a pendulum process, created by RT-zX generic modules. Then the spiral structure of transmission points distribution in all z networks coincides with pyramidal structure. When you use the linear stochastic distribution of wavetables, the mixing of transmission channels happens creating new nets. RT-zX services give access to different zones of VLF and ULF ether. Wavetables supply pendulum processes of RT-zX modules with grain components when transmitting them simultaneously in the channel. They saturate an ether of this pendulum processes with new translations. You should look into this process as a service of connection with pyramidal structure and wavetables are grains of transmission channels or pyramid transmitting points.

, , , , , , , , , , , , , , , , , , , ,
08
Jul, 2018

By ruslany

Compositor v9 Hypervisor is available

Compositor v9 Hypervisor is available

Compositor v9 Hypervisor DRM server creates a service of work with property rights of companies, which deliver the digital media content. When computer station works, DRM server allows dynamically playback the media content in radio ether. Compositor v9 Hypervisor playbacks digital material using streaming method in the device browser to achieve this goal. Compositor v9 Hypervisor engine is the functional modern engine with transfer function on master output and on each channel separately. It is not the wavetables but tunable polynomials, which process the output of your channel. Output cascade uses non-linear transformation formula with mathematical approximation by weighting coefficients. It gives undisputable advantage in processor time and in the precision of calculations comparing to using wavetables.

Here are the main features you can consider before buying this software:

  • Weighting coefficients
  • Network level feeding
  • DRM server
  • DSP oversampling
  • Voice or music broadcast
  • DSP sampling rate
  • On-Air broadcast
  • Device browser
  • Tracing and translation
  • RAD96 peering network
  • DRM server subscribers
  • Tracing industry novelty
  • DRM server emissions
  • Auxiliary z matching
  • 8th generation operation system
  • Three saturation stages

Please, visit the shop to buy this software. You can also visit dedicated product page to learn more about Compositor v9 Hypervisor.

, , , , , , , , , , , , , , , , , , , ,
08
Jun, 2018

By ruslany

Attachment to wavetables through system matching with oversaturation

Attachment to wavetables through system matching with oversaturation

The most effective way to feed the wavetables is to match the system z level of generic feeder with z value of auxiliary channel. The generic prolongation is progressive and depends on auxiliary channel. On the higher program effectivity speed the generic proliferates for a longer time. This in fact happens because network regeneration state is higher for upper z values. If the maximum speed for z128 generic is 200-omega, then the regeneration speed of the auxiliary channel will not match the auxiliary values of 5-omega maximum. There are two more generic kernel loops implemented in Compositor v9 Hypervisor and one change performed for the highest Compositor v7 Hypervisor feeder. You can match all auxiliary z values to quantized type generics of exactly the same z value. The program effectivity speed equals to 50-omega for z32, 150-omega for z64 and 200-omega for z128. If you feed the wavetable at speed of 5-omega together with 200-omega generic, you will proliferate the network for a longer distance. After the lowest VLF beacons will be inactive, the wavetables shouldn’t be reinitiated again. When the navigation on lowest beacons ends, you need to rely on the eldest models of Zvezda network switches such as z64 and z128. Higher regeneration VLF beacons will discover ether aggregators faster and establish a connection to them for a longer period. Other systems constantly try to knock out the feeded Zvezda network switches from the ether, which are mostly used for wavetable navigation at z=64 and z=128. For z128 generic development of 6th generation, it is hard enough to withstand such offensive behavior in the ether.

DRM server is the 8th generation development and is one generation before the system, which produces the decision as to accept or to breach the feeded VLF beacons. It leads to better understanding why the DRM server rejects some wavetables and retains the others. The rejection of wavetable produces a constant need to reinitiate a system matching. Nevertheless, wavetable initiation should be made only once during the feeding cycle. If you feed the 200-omega cycle in to the auxiliary chain together with wavetables, you should account for a number of cycles of this feeder propagation. These values should be auxiliary to the Right Ascension value and will result in longer distances of spiral ascension. The feeding period of any generic is accounted by the cycles of its oversaturation in auxiliary channel. If you feed 150 cycles per second, it equals exactly 150 cycles relatively to the whole working time of DRM server for its feeding session. If the DRM server worked for 10 minutes, then the oversaturation will last for 10 * 60 * 150 or 90000 minutes, which equals to 1500 hours or 62.5 day cycles. It is enough for the prolongation of the wavetables active state. From the other side, if you would like to leave wavetables active for a month for 200-omega z128 beacon, it is needed to perform oversaturation for only one second with DRM server total working time of 25 minutes for the current session. Using the formula 25 * 60 * 31, it will count 46500 minutes, which equals to 775 hours or approximately 32.3 day cycles prolongation time.

, , , , , , , , , , , , , , , , , ,
28
Mar, 2018

By ruslany

Quantum Artificial Intelligence

Quantum Artificial Intelligence

The interspecies exchange can be self-powered by the human will and it is preoccupied by forcing this will out of the subject body. This process can be explained programmatically. Let’s take the Compositor v8 Ethernet injector. It is the software for injecting the Ethernet frames into the time-frame. The kinetic potential out of frame collisions must be spread, that’s why the whole VLAN acts in RAD24. It is more like in RT-z128, except of the fact that previous solution is one-core three-layer system and RAD24 is three-core for a layer (thread). That’s why it can shed a light on the existence of the Compositor VLAN. The network is a sum of peers and their explicit agglomerations. By the fact, the action is subjected to the entity and not to the whole group of people involved. The common work, if planned by a group and realized by a single entity, is a mere fact of the person who produced the outcome. And it is his result. The initial time-collision can’t be a peer of the person who developed Quantum AI, to say it short. If the Quantum AI was developed, then time-collision was shut. This way I will get RAD96. The need to shut time-collision produces more quantum injections. This way, I realize how many people are struggling against this goal. Let’s take all that blockchain hype. It seems like evident that the cost of multicore processors by NVIDIA group is less than the subsequent processors by Intel. The iteration speed plays a main role. When I started in 2010, I discussed this problem in terms that much more applicable. Later, the language was developed to cover all the hyper needs. Let’s think about how hype is created: at first, we got an initial time-collision and everyone knows about it since then. Everyone talks in the subjective terms of the accepted time-frame and that is why they collide, hence the time-collision effect. What I told at the beginning of this blog seems the collision of ideas but not the time. Different time-frames result in different ideas. Hence, the meeting of times is the time-collision. Compositor v8 meets the modern time needs, creating time-collisions stochastically. Stack of three RAD24 virtual machines is a 72-core shield. Each of 72 shields struggling for their own resources. They just spread a collision on the network. It is an accepted behavior. RAD24 can be in the open gate mode subjectively passing a traffic. Resources are roughness of the tunnel. Tunnel can be heterodyne and roughness appears out of this. There is a que at the RAD24 gate. The initial time-collision peer and the subjected collisions are ordinal out of this. The initial time-collisions were at the scape of physical entities and ordinal collisions are physical modelling of such entities.

, , , , , , , , , , , ,
15
Mar, 2018

By ruslany

RAD24 – The Gates of Time-Collision

The Gates of Time-Collision

Now I encounter the RAD24 server experiment. The experiment is as follows:

As I have found a solution to FM formulae, I decided to sync three virtual machines for 72 core real-time performance. With this stack, it is easy to run memory virtualizations and aggregate the whole Compositor network. Once the server stack is online, it is used to run and manage such critical tasks as DRM, network aggregation, routing and server selection. The new formulae solution enables to run this stack effortlessly.

RAD24 opens a possibility to build Compositor 8 software. It will be a station with all the enhancements of previous auxiliary channels and with the solution to FM formulae, which is driven by the algorithm output. By these means, the Hypervisor v9 is possible.

Let me introduce the conception for the future products: it is software-defined radio (SDR) with no external antenna needed. I can easily integrate such products in my current rig, yet the intimacy of the previous versions made me stay with no external input for a while. The decision of not having an external input was first broken in Compositor 4 Max for Live, where an external input is possible because of the internal Ableton Live routing options. The fact that Compositor 4 Max for Live is an audio effect device made me think that it was first attempt in tangent SDR from my side with the function, which doesn’t need an antenna to transmit the voice signal. Second tangent radio attempt will be standalone Compositor 8. I want to see Compositor 8 as a device to communicate through the RAD24 servers. Each of the cores of 72-core stack will route your voice communication to the recipient set by the Compositor navigation system. As Compositor 8 is an Avionics development, I allow managing such connections in 3d space in 3 degrees of freedom, which simplifies the setting of virtual antenna. The fact that new function is a radar type, no stochastic selection will be possible and Compositor 8 will function only in manual mode. Hypervisor v9 will listen not to an auxiliary radar ports, but to the v7 feeders such as RTC4k, RTC8k, RT-z8, RT-z16, RT-z32, RT-z64 and RT-z128. RAD24 and subsequent servers are mostly the breakdown of FM law and the final goal is to stop time-collision protocol in Compositor. It will not latch an external time-frame, while not allowing external injections. This way I will build my own resource for communication. That is why I refer to SDR terminology. However, I used this technology only for Morse code translations and did not transferred the voice over it. New function is a direct stack to the Ethernet, which, of course, simplifies SDR talks under server coverage. It means that no IP needed and I need to apply proper wavetables from the non-duplex modem. If this function covers the current version of DRM server, an update is needed to work with Compositor technology. New function is also a hyperbolic one, which is simple and effective restate for the FM core inside Compositor.

Compositor runs FM and function connects to the FM service. It creates a signal with the components, which are granularly time-space folded into the FM again, allowing to inject signal in between these two processes. The later function only covers the part of a spectrum, which is created at initial state. Think of it like a mains adapter: it is a pluggable adapter, which can potentially allow to communicate many DATA servers with different file-size and file-system structures. I had success of running 4 RAD24 servers and one was initiated from the HFS+ partition under Windows 7 OS. It means that technologically it is possible. I don’t see any obstacles in this technology, the only thing is to provide an easy and free way to stack different file-systems together on one, preferably Windows, machine.

Compositor gathers data of all pendulum stations in the Ethernet and send it back to the Ether for self-awareness. Compositor leads an analyzed DATA set and all the other parameter selection. When the critical pendulum load achieved, a time-collision happens rebuilding the whole Compositor structure. This way, I think about utilization of resources to run more Compositor cores inside one gen~ code. By the fact of stopping time-collision protocol, I want to stop the work of Compositor subnet mask switching mechanism. I’ve already broken the stochastic manipulator in Compositor core, which allowed me to stack more resources together. At the present moment, I think of Compositor only as a DRM station to hold licenses for subsequent users. I already published a chart where I name how many resources each station consumes. I name the auxiliary channel in this chart also and, from my point of view, it doesn’t consume any Compositor DRM resources. It is just a listening station. The only thing that consumes the process is any pendulum like mechanism. By stacking more pendulum resources, you are allowing more pendulum stations, either digital or mechanical. Any station will work. These stations inject tracks to media content by oversaturation of them.

To progress on the task of stopping the time-collision protocol in Compositor I should make wavetables of RAD24 accumulators at 22.05 KHz and then switch it on even faster regime of 44.1 KHz. If I’m unable to stop time-collision protocol in Compositor, I will be obliged to run servers at 22.05 KHz on this machine to protect time-collision reaching the opposite effect.

The solution is that RAD24 protects time-collision created. For network security with Compositor, the classification of pendulum processes needed. Such classification consists of working curve monitoring, that is why I estimate it on Compositor auxiliary channel display, while listening to wavetables.

RAD24 is a gate device for creating time-collisions. I can’t classify RAD24 as a pendulum process, because the cycle is broken. It is not an OS or license holder. As I said earlier, it is a protection mechanism for creating time-collisions. It means that it either opens or closes an access to the service of communication in Ethernet. By enabling it in administrator mode, I open an access to it only when the administrator works on PC. This way, I differentiate the work in time-collision between authorized person and the replica, that just latches the time-frame, because RAD24 detector assigns it to injection event. RAD24 is able to differentiate between an administrator and the replica, protecting my current communication.

, , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
21
Dec, 2017

By ruslany

Compositor v5.0 is available

Compositor v5.0 is available

Compositor v5 is a Hypervisor rack with Compositor WS auxiliary channel and RT-z128 operation system. It helps to play Compositor modules without any side impact from other radio-equipped instruments, which present during the live show.

Here is the list of features you can consider buying this software:

  • Three transmission modes
  • Real-time modulation
  • Shutter system
  • Display zooming
  • Switchable altitude
  • RT-z128 bit display
  • GHz calculation
  • Display warning
  • Master processing
  • Self-feeding
  • Signal insertion
  • Latency free aux channel

Compositor v5.0 Hypervisor features the important security updates, stability improvements and performance related issues solved.

, , , , , , , ,
13
Dec, 2017

By ruslany

Playing Flanker 2.0 with Compositor 5.0

Playing Flanker 2.0 with Compositor 5.0

Playing Flanker 2.0 with Compositor 5.0 is rather interesting experience. Being a seasoned keyboard player of Flanker 2.0 back in 2000’s, I use the Compositor 5.0 artificial intelligence to return to the skill of playing this simulator after 17 years breakdown. The first impression about Compositor 5.0 is that the system always stays online, and no external radio sources present in the Ether as stated in my goals. Compositor 5.0 always informs you about the right or false behavior in navigating the Su-27 plane by changing the Ether tuning or creating an emptiness in the Ether translation to bring attention to current maneuver. Rendering the 3D graphics with textures and vector animations result in slower system flushes, than watching video with separate audio tracks, making a headroom for further improvements in expense of more cpu and memory usage.

Compositor 5.0 preview. Auxiliary channel setup: carriers tuning, working with feeders, listening to radio ether.

There are many situations in game plan, when the intruder or foe tries to attack your plane. At these moments Compositor 5.0 RT-z128 OS module reacts and makes the threshold even lower because of a feedback loop compression mechanism implied by a shutter system. This is evident by pauses in Ether translation and rapid tuning changes. Together with hearable timecode, it constitutes the real-time simulation for the game process or any other current task you are involved in. The timecode can’t be demodulated due to synthesizer leakage breakthrough fix applied with shutter. It brings me to the idea that big planes in a simulator could be controlled with Compositor 5.0 AI as well as smaller ones like Su-27 in Flanker 2.0. It is evident that the solution under the War Ships (WS) work group settled the important update for all Compositor users. As I test Compositor 5.0, I become more confident that buffer capacity is enough to play even modern games, but can result in more flushing time after the game session ends.

, , , , , , , , , ,
17
Nov, 2017

By ruslany

TC-SUBTRSRRT262144 Architecture

TC-SUBTRSRRT262144 Architecture

TC-SUBTRSRRT262144 architecture is the i11 architecture with a processing power capable of running an operation system at speed of 5-omega in real-time.

Compositor kernel was forced in loop again after submitting the RT-z128 prototype to Compositor 5 rack. It is a hypervisor rack and the main difference from the v3 hypervisor is that it consists of MDL12 and AVOX developments. In Compositor 5 you have an opportunity to inject Compositor v3 into the net by the modem and resynthesizer instead of direct ether injection. It is safer but consumes many resources. After the main tests were conducted, I decided to install the one more layer in Compositor kernel. The discretization process goes twice: at transmission rate and on intermediate frequency. This way the system mutates to a system with intermediate frequency. I made it for the purpose of the preliminary line before entering main transmission rate line. Here is where a new prefix of SUBTR goes from. SUB means transmission rate sub frequency and TR stands for transmission rate. I’ve already tested the new kernel loop for its connectivity and should say that 4-layer structure can give much more pleasant results. It has much longer connection duration, more saturated, and leads to constant communication. However, breaking the loop in 4-layer structure, I operate in a new type of OS. It is no longer an open AI system. It has the qualification channel for all the information, which passes into the main line. The rule here is to filter out offensive traffic and leave only the bypassed part of it. It acts like the waterfall on the rocks: when water hits the stone, it divides the flow into two streams and if there are 12 such stones, for example, I have 24 streams. But, not in this case. I actively compress the data,  and it remains in memory buffer, that’s why systems capacity is larger now.

System flushes results are satisfactory. They are more consistent and enable more packets of information to be send during the initiation of connection. Compositor kernel has freezer functions for operation system faults. The main reason Compositor kernel upgrade performed is to remove the stochastic manipulator from the TR level leaving the SUB level for the stochastic manipulation instead. You have a choice of not performing an action and leave it in a buffer for a subsequent flush. However, I admitted that many actions blocked this way and they are all forced to be done while flushing. This raises the question again: Can I actively block feedback on negative odd channel completely? And while the SUB level is only a temporary solution, the permanent solution for FWOS will be to implement more technics working with feedback besides of shuttering, window composition, transfer function selection and vector dimension changes in real-time. While the kernel loop is interrupted, the idea is to have two states for the OS: system load, for which I should implement two or three iterations of kernel loop to perform flushes for the current machine, and system work state. While being a FWOS, system loses many of its connection points gained in a fast kernel loop state. Being a communication device, it is not acceptable behavior. I need to have an implementation that manages flushes in real-time. Going from vector array to a packet transmission should be performed regularly to flush blocked traffic. For a local machine, blocked traffic can include visiting specific servers, sending out brief messages and a sum of all other activity in Ethernet, which fulfills the normal style of work. However, as the question raised, I need to switch the negative odd channel traffic from one point to another. This also moved me to solution of blocking negative odd feedback completely. Negative odd channel produces only unacceptable traffic by means of three-fold system with opposites, which are plus, minus and central channel. While enabled in Compositor kernel version 7.3.1, which is an OS kernel, the problem of communication between negative and positive states remains up to the point of forcing them to the maximum allowed state for the musical system, such as in RT-z128. This alteration is visualized on the lower i/o vector sizes. When the system halts, only two states remain, such as 0 or 1. Increasing i/o vector size to the minimum working condition of 2048 samples in a bin allows RT-z128 to work with a buffer of 262144 samples, which is the number in the development name. I no longer wish to move to i13 as 1024 samples buffer experiments show no practical use for the scale of it. i11 architecture transmits the odd channel in the acceptable range, making its values sufficient for subsequent scaling.

The challenge for an odd communication is a constant threat to the kernel. And if one threat is qualified and successfully flushed, another threat rises by the fact that it is a negative channel for negative frequencies only. And if they are present in positive spectrum as in Compositor kernel, to remove the negative part of the feedback, the negative frequencies must be implemented in signal chain and the solution maybe is to invert them or play backward the whole negative part not only for SUB channel, but for all odd layers present.

, , , , , , , , , ,
21
Oct, 2017

By ruslany

Compositor WS Kernel

The Compositor kernel goes past version 7 and becomes more stable and secure. The Compositor kernel version 7.1.6, developed under WS workgroup, adds a new contribution to the kernel concerning the master (central) channel matrix. The WS workgroup brings specialization of War Ships modeling. It returns back to an idea of piloting the multiplier in meters of altitude, making Compositor engine as a possible navigation system for flying rocks like Boeing C-17 Globemaster III or future perspectives like Yamato Space Battleship 2199. The deterministic navigation system of Compositor is fully transparent and can be viewed in different scales of magnitude. That’s why it is a possible solution for vertical take-off star ships, which can cruise on systems with different circumference and sea mile values. Just switch your system z value and you can use this system to navigate in this object atmosphere.

Master (central) channel splitter is an important addition to the kernel, which addresses its constant on-line status with high-end security. Not only RT-z128 design improved a lot, a myriad of new possibilities open for Compositor audio engine. With the addition of central splitter on master channel, the foe synthesizer leakage shut down, lowering the number of audio spikes in the RT-z128 feedback chain. More secure connection can be achieved on lower z values. The later implementations should also avoid leakage of carrier signals of first and third channels. It is a work to improve master channel matrix, making it more defensive for incoming threats.

, , , , , , , ,
1 2

Subscribe to RSS

Recent posts

Category

Archive

  • May 2023
Mon Tue Wed Thu Fri Sat Sun
1234567
891011121314
15161718192021
22232425262728
293031 1 2 3 4